Home

User's Manual

1. Odeon Room Acoustics SOFTWARE ODEON Room Acoustics Software Version 13 Full User s Manual Written by Claus Lynge Christensen and George Koutsouris Published in 2015 Odeon A S Scion DTU Diplomvej bulding 381 DK 2800 Kgs Lyngby Denmark Tel 45 8870 8845 Fax 45 8870 8090 www odeon dk Introduction Thank you for using ODEON the Room Acoustic Software package that helps you make complete studies of room acoustics for any space through simulations and measurements We hope you will enjoy working with ODEON and that you will find it extremely useful in the acoustical design of new spaces and improvement of existing ones ODEON has a very user friendly interface that allows setting up a project very fast and elegant presentation of results This manual is intended to serve as an introduction on modelling room geometries in ODEON software on the facilities for measurement and simulations and the applied calculation principles behind It will not cover in depth all facilities included in the ODEON software explanations of displays calculation parameters results etc are available as context sensitive help from within the ODEON applications shortcut F1 It is recommended to use the online help to learn about the specific features available from the different displays the interpretation of results calculation parameters etc The contents of this manual are as follows Chapter 1 covers installation of the pr
2. i 4 Vi VEAN VA N ay AN X se NY SAA N IN rN AAS ff Odeon 1985 2005 CAHRISMA project rooms Hagia Irene par is a model of a Byzantine church in Istanbul which like the examples above also includes measured room acoustical parameters The church was modelled as part of the CAHRISMA project Conservation of the Acoustical Heritage by the Revival and Identification of the Sinan s Mosques Acoustics Hagia Irene has an approximate volume of 39 000 m3 and RT of 4 3 sec at 1000 Hz You can find similar rooms in 21 the CAHRISMA project subfolder These rooms need to be calculated Open any of the rooms access the Joblist and press ALT A to run all jobs ERATO project rooms Open the Jerash_Present_Empty par ancient theatre and observe the bounding box around the room The bounding box is not part of the actual geometry but it is necessary in such a case where the room is not watertight by itself there is an open ceiling The box encloses the geometry and allows ODEON to perform calculations All sides of this dummy box are assigned 100 absorption in order to successfully model the open air You can find more models from ancient theatres in the ERATO project subfolder Noise control rooms The example Studstrup Power Plant par found in the main Rooms folder is a model of a turbine h
3. SPL dB N O OO PO CERRY a WI Pa 0 05 0 0 05 0 1 0 15 0 2 0 25 0 3 0 35 0 4 0 45 0 5 0 55 0 6 0 65 0 7 0 75 0 8 0 85 0 9 0 95 Odeon 1985 2013 Licensed to Odeon A S time seconds Impulse response with prominent spikes at regular distances that indicate the presence of flutter echoes in the field between the buildings The frequency is 4000 Hz C Odeon12Combined Measurements IR_BetweenBuildings2 wav Raw decay curve at 250Hz 60 R T ONNaRON 58 i Jv E Measured 56 PAETE Were Noise floor 54 F a nsceies Onset time Truncation time SPL dB a 0 2 0 1 0 0 1 2 0 3 0 4 0 5 0 6 0 7 0 8 0 9 1 1 1 1 2 Odeon 1985 2013 Licensed to Odeon A S time seconds At 250 Hz the echoes are less distinct but the reverberation time is longer 124 12 Genetic Material Optimizer Combined and Auditorium editions only One of the most important new features of ODEON 13 Auditorium and Combined is a state of the art optimization tool that allows you to refine the materials in a model in order to match the real acoustic conditions in an existing room The tool is based on a research presented in Forum Acusticum 2014 Krakow Poland Christensen C L 2014 A genetic algorithm is utilized in the optimization tool The optimization process can also be called material calibration or material tuning Optimizing the absorption coefficients of materials inside an acoustical room model to match with measured data i
4. e Open the simulated multi point response results for a selected job Click the Measured versus Simulated tabsheet Choose Open Edit measured data file eaba File Toolbar Measured versus Simulated Options Tools Window Help Nt BD S S PSS eS Select next frequency band Up m 3D Source Select previous frequency band Down Show table for frequencies Next Parameter Previous Parameter Alt F Right Left Next receiver R ulated mode Previous receiver Shift R fer curves 1 Energy parameter curves 2 Statistics Energy parameters Measured versus Simulated Show table f i Alt R Ow table Tor receivers at 1 000 Hz Receiver 1 Toggle hide graphs T Open Edit measured data file Ctrl M Reload measured data Ctrl R Export ASCI text file Ctri A Copy to clipboard Ctrl C Q m OQ coe 0 8 0 6 0 4 JobList Bi 02 Active sourg 7i P1 S01 0 P3 No de e ee R amp e r 3 amp 2288888 EE Solo No da Receiver Frequency Hertz 2 No de e Now you can enter the measured data You need to follow the format shown below RECEIVER 1 T 30 z 2 09 2 04 1 92 2 01 1 97 1 84 edt 2 11 2 2 2 09 2 09 2 02 1 89 D 50 0 1927 0 4857 0 4753 0 4519 0 5582 0 5768 C 80 a 2 36 1 14 0 76 1 10 2 59 3 15 s Ts 162 29 112 26 107 36 112 23 90 32 85 17 5 SPL 10 49 10 84 10 54 9 69 8 56 8 82 5 LF 80 0 0062 0 0026 0 0067 0 0247 0 0458 0 0781 0 0273 STI 0 7 SPL
5. modelling the columns for ColYCnt I NumColY for ColXCnt 1 NumColX MReset MTranslate ColXCnt L NumColX 1 w 2 ColY Cnt W NumColY 1 0 NumbOffSet Auto Layer Columns 0 502 0 502 0 000 Box I ColumnW ColumnW h n columns in the room olive colour NumbOffSet Auto MTranslate 0 0 1 2 Layer Table plates 0 000 0 502 1 000 bluish colour Box 1 3 3 0 1 tb tables end end HHH Symmetric modelling Symmetric rooms can be modelled taking advantage of the ODEON convention for symmetric models This allows generation of symmetric or semi symmetric rooms with symmetry around the XZ plane Y 0 symmetric modelling is always carried out in the main coordinate system it does not take into account manipulations carried out using UCS MTranslate etc Modelling a surface symmetric around the main axis e g a reflector above the stage can be done using symmetric points Modelling left and right walls at the same time can be done using a symmetric double surface Symmetric points Surfaces symmetric around the XZ plane Y 0 can be made using symmetric points If defining the point Pt 2 1 0 1 0 1 0 in the geometry file using the point 2 in a surface definition of the geometry file will refer to the auto generated point Pt 2 1 0 1 0 1 0 173 Thus the following surface definition Surf 1000 Symmetric surface 1 2 2 1 will model a surface symmetric around the XZ plane Y 0 e g an end wall or a reflector If the surfa
6. AUDIENCE AUDIENCE 6 50 X4 065 0000 Noms AUDIENCE AUDIENCE 6 50 3000 06 0 000 Noms AUDIENCE AUDIENCE 0 55 30004 0650 0 000 Noma AUDIENCE AUDIENCE 3 57 99 Water surface in swimming pool Ref 14 30004 0 650 0 000 Noma AUDIENCE AUDIENCE __ 100 Rough concrete Ref 15 30004 0 650 0 000 Noma AUDIENCE AUDIENCE 5 25 __101 Smooth urpanted concrete Ref 15 Mat 2000 Eine seats in the rear part level 127 and above ovre partet 102 Smooth concete painted or gazed Ref 15 103 Porous concrete blocks without surface firish 400 800 ko n3 Ref B1 13 63Hz 125Hz 250Hz 500Hz 1000Hz 2000Hz 4000Hz 8000Hz alw Class eel SS T osoo 0 50000 0 66000 0 80000 0 88000 0 8309 0 70000 0 70000 0 80000 B 63Hz 125Wz 250z SOOHz 1000Hz 2000Hz 4000Hz 8000Hz olw Class o BB JobList Surround mode Speaker layout 7 0 Conwolve Surround RIR with Signal Se ix Comvolved wave results Into one wave tle gt j Enabled Signal Sab Path Signal fle rime Chae CANE TEBE Reener I TORBEN Rod lev Hax out TT Corvolutions in micno 1 3 l 1 v Orchestra Mozart MOZART sopr 21 _ _ 3 4700 Average 2 1towadsp46 Soprano 0 00 5 08 _no REI vtec Lev Delay in coc SEEN SIESTE e 2 v Orchestra Morart MOZART Wii 21 _ 34700 Average D 3 1towadsP46 1stvioin 0 00 32 68 1 0 00 0 00 Soprano MOZART ME 3 Vi Orchestra Mozart MOZART v2 21 _ 34700 Average 4 itowadsP4 2nd vioin 0 00 12 04 a 2 0 00 0 00 1st vioin wo
7. T g 5 2 X y 2 3 8 9 xy z 8 470 10 596 3 770 10 y 2z 11 xyz 12 xy 2z 13 y 2 10 amp y 2 8 470 10 596 3 280 8 470 0 976 3 280 8 470 0 976 0 770 8 470 10 596 0 770 8 470 10 596 3 280 Remember Surface properties assigned in ODEON such as materials surface sources reflectors and grid surfaces are preserved when the model is re modeled in SU and re exported for ODEON Since version 11 the Sketchup file skp file is automatically copied with an ODEON project when using the File gt Copy files menu from within ODEON 2 4 Importing DXF and 3DS files If using models from AutoCAD it may be an advantage to use the 3ds file format for export from AutoCAD and Import into ODEON rather than the axt format This procedure will ensure that all relevant data from the CAD drawing are exported in a way which can be used directly by ODEON 27 The support for the DXF file format Drawing eXchange Format allows import of CAD models exported from modelling programs such as CAD package http www autodsys com www intelliCAD com SketchUp http sketchup com AutoCAD www autodeskcom iT O O 3DStudioMax http www discreet com Rhinoceros www rhino3d com o Z o Microstation http www bentley com en x US Products MicroStation There may also be other programs around capable of creating geometry data which can be used with ODEON The DXF import engine in ODEON suppor
8. The Odeon Auditorium and Combined excluding the demo version installs with a number of anechoic recordings that may be used with the Odeon software without restrictions except for the orchestra recordings For the orchestra recordings in the WaveSignals Orchestra folder special restrictions apply Customer acknowledges that the anechoic recordings are the sole property of TKK Helsinki University of Technology Customer acknowledges the right to use the recordings for internal use only and the right to make one copy for archival and backup purposes Customer agrees not to modify or remove and not to tamper with or disable any automated display performance or execution of any credits references or notices without the prior written consent of TKK Helsinki University of Technology Customer agrees not to copy modify adapt combine publish rent lease sell distribute the recordings or sublicense or otherwise transfer the sublicense If the anechoic recordings are used in the Customer s work in such a way that an ordinary reasonable user familiar with the recordings would be likely to recognize it Customer agrees to include the following notice in the resulting work Derived from anechoic symphony orchestra recordings by Jukka Patynen and Tapio Lokki Helsinki University of Technology and licensed from Odeon A S GENERAL This Agreement constitutes the entire agreement between you and Odeon A S concerning the ODEON product If a
9. end loop A Counter surface is divided on two lines and must follow the syntax CountSurf lt First Surface Number gt lt NumberOfSurfaces gt lt Optional name gt lt ListOfPointNumbers gt lt FirstSurfaceNumber gt A unique number from 1 to 2 147 483647 for identification of the surface Using the same number but with negative sign defines the surface and its counterpart mirrored in the XZ plane Y 0 A CountSurf will take up several surfaces numbers which must all be unique lt NumberOfSurfaces gt 166 The number of surfaces to be created by the CountSurf call lt Optional name gt Optional user defined name for easy identification of the surface e g Beam lt ListOfPointNumbers gt Each surface may be bounded by between 3 and 50 corners which all lie in a plane Corner numbers refer to the corners which must have been defined e g using the Pt or CountPt statement before using the surface statement The order of listing must be as obtained by travelling around the surface s edge in either direction The list of corners must be on the same line A room may contain up to 10000 surfaces Example CountSurf 1000 5 Beam in ceiling 1000 1100 1200 1300 will produce five surface the first one containing the numbers given in the ListOfPointNumbers the next surface with 1 added to all the corners in the list etc Of course all the points referred to need to be defined typically this is done using a CountPt def
10. io i aa e a E a 0 00 0 00 e om 0 00 tlevation Sttz fixed 125 Hz Fined 250 tte Freed 500 fixed 1000 Hz Freed 2000 Hz fixed 4000 Hz Fixed 8000 Hz Freed a o E a w 0 00 a Bi o m a 0 00 0 00 E o w a w 0 00 aa o E E a w 0 00 in a a 0 00 im i a ow a 7 ar o E a a 0 00 o co EE ee o m a a ow 0 00 a o E a 0 00 a 00 a v o E a C 0 00 i a P a o co T R o E a a 0 00 a o m a 2 00 a o E a o 0 00 E o 0 00 a o E a 2 00 a 0 00 Copy Display opbans Lelt gt Fight Top gt Bottom Front gt Back Revolve Let Leiria Pee band Albendeunnit Bards io depla in crectviy roser Selected bard y RightoLett Bollom gt Top Back Front Top gt Lett en Copy selected frequency bend 4 en dmonic tangs 20001 e Apply calibration Freqercy 63 126 250 wit 1000 2000 4000 S000 Hz as Total Power Spectrum ooo 000 10 000 0 00 is a00 100 000 2 dB arOrredssenstiv AsdSPL 7 00 menet Create file 45 Enter the dB values sound pressure level for the horizontal and vertical plots at the selected frequency band in the corresponding tables The angular resolution is 10 degrees If data are not entered for all angles e g if the data are not available ODEON will do interpolation between the angles entered For angles between the polar plots ODEON will perform elliptical interpolation Equalization If the source to be used in ODEON has a fixed frequency depe
11. 0 1 Power 10 SPL_250 AW_250 0 1 Power 10 SPL_500 AW_500 0 1 Power 10 SPL_1000 AW_ 1000 0 1 Power 10 SPL_2000 AW_ 2000 0 1 Power 10 SPL_4000 AW 4000 0 1 Power 10 SPL_8000 AW_ 8000 0 1 m gt 133 Open the Optimization Tool Click the Genetic material optimizer button 9 in the Main Toolbar or press the shortcut SHIFT CTRL Q The interface should look like in the following picture List of jobs List of genetic parameters Editable list of materials Original and best error displays of Genes ep gt 10 Activ 4 Sa eer po oa SS ae i Best Fitting Last error decrease 7 2No desapton RE SS a IB Orig Fitness i T e 7 M Best Fitness 5 No description Evolution Method smEjitist No description s 7 ro desarpban Inividuals per material 2 X 3 R A te ye pan Crossover Probability 93 9 3 10 No description Inversion _Probabiity 19 9 S 11 No desoript z z 12 No description Mutation Probability 15 n lt E 22 0 9 63 125 250 500 1000 2000 4000 8000 prere Ra baal 63 125 250 500 1000 2000 4000 8000 Frequency Herz Material J 63Hz 4 125Hz 250Hz SOOHZ 4 1000Hz 2000Hz 4000Hz 8000 Hz Absorption Coefficient material 51 Sackwal in sud 21 frst guess 50 0 50000 0 80000 0 50000 0 50000 S0000 0 50000 0 50000 0 50000 0 7 W Initial 9 CEOE eh Sh Ia GI BI D S e D Me Aea EN B L OON DAN DAL IO 1A I n w Optimized 10007 Sold wooden door Babran 1973 0 140
12. 10300 0 07900 0 08400 0 07900 0 07200 0 09700 0 07900 0 11400 187 57 so s 1945 S High Limit 3 10003 Double gazing 2 3 mm gass 20 0 07200 0 06800 0 06200 0 05900 0 02600 0 01800 0 02800 0 04300 6 11 s 2 mm gap Knstensen 1984 eS ret 40 absorbent 0 42900 03W00 0 40000 0 42100 0 52200 0 3220 0 25500 0 22400 iss 5 LLCO CP TT 5 3068 Phywood paneling 1am thick 0 14400 0 238500 0 18900 0 15200 0 07000 0 0540 0 11300 0 05100 6 60 5D 63 125 250 500 2000 8000 6 1991 s i sas Sa Sa eee pemn a a SRR yee NET an Frequency Hz v O Error as a function of generations z E0 iT s ANAN S BHT TT rt itz 7 z j 5 N f IN A A A m Hy fi ttf y A AN TY aval A f j TI f V T N MY WAINI AA AA 4 V 4 N y f V jg INT AT 1A a j i f A j o V VUIA AMAIAN a fi I4 N 9 2 AHAHA Ye ee p gt AAA f ES f lt l MARAVI VV LIV NVUUV WW TTT a y SA hy AJ A J Fi 0 14 0 21 1 6 1 13 1 20 2 5 2 12 2 20 3 5 3 12 3 19 4 4 4 11 4 18 5 3 5 9 5 16 6 1 6 7 6 14 6 21 7 6 7 13 7 21 8 5 8 12 Generation Individual 0 2 0 7 Testing generation 8 Tracing rays 910125 of 1100000 Individust 18 Remaining time horm s 00 00 22 Now have a look again at Multi point response for jobs No 1 and 2 in the Job list SE The deviation between measurements and simulations for most of the parameters has been reduced significantly Double click on the display to get a table with the diff
13. 2009 standard should be followed To obtain good estimates of reverberation time the minimum source receiver distance should be used in order to avoid strong influence from the direct sound The minimum source receiver distance according to ISO 3382 1 is mn ep m where V is the volume of the room m3 C is the speed of sound m s T is an estimate of the expected reverberation time s Thus for a typical concert hall a source receiver distance less than 10 metres should be avoided in order to get good predictions measurements of the reverberation time Minimum distance from the receiver to the closest surface If a receiver is placed very close to a surface then results will be sensitive to the actual position of the secondary sources generated by ODEON s late ray method If such a secondary source happens to be very close to the receiver e g 1 to 10 centimetres this may produce a spurious spike on the decay curve resulting in unreliable predictions of the reverberation time indeed if the distance is zero then in principle a contribution being infinitely large would be generated To avoid this problem it is recommended that distances to surfaces are kept greater than say 0 3 to 0 5 metres 95 Anyway for measurements it is for other reasons recommended to keep distances greater than a quarter of a wavelength i e 1 3 metres at 63 Hz A distance of 1 metre is required by ISO 3382 96 10 Line array sourc
14. 6 is used 154 Getting the voltage and energy right Values are in tension V The values should be reduced by a factor which compensates for the NumberOfSubBands Ns used if Ns sub bands are used then the Re and Img attributes of each node should be multiplied by a factor If there were 6 sub nodes in each band and Re 1 15534266201273234 and Img 0 0 for each of them then the equivalent voltage for one sub band representing the total power in that band would be 1 15534266201273234xSqrt 6 2 83 V 2 83 V is the standard voltage for which sensitivity at 1 metre is given for loudspeakers so with this setting it will reproduce the sensitivity values at a distance of 1 metre se the Dipol_Domain_Frequency xml sample 2 83 V may seem like an arbitrary value but it is generally agreed upon as standard voltage this is the voltage at which an 8Q speaker deliver 1W The reason why the industry has moved from 1W to 2 83V Loudspeakers rarely have a frequency linear impedance of 8Q whereas an amplifier is actually fairly capable of supplying constant voltage Position node Attributes X 1 Y 2 Z 0 324 Position of Source of transducer origin given in metres Position defaults to 1 1 1 Orientation node Attributes Azimuth 27 180 around vertical axis counter clockwise is positive Elevation 10 90 up is positive and down is negative Rotation 10 180 counter clockwise rotation around
15. Cone lt Number gt lt NumberOfSurfaces gt lt Radius gt lt RevAngle gt lt Height gt lt optional name gt lt Number gt A unique number from 1 to 2 147 483647 for identification of the first point and surface in the Cone Using the same number but with negative sign defines the surface and its mirrored counterpart in the XZ pane Y 0 A Cone will take up several point and surface numbers which must all be unique lt Radius gt Radius of the Cone must always be greater than zero lt Revangle gt 177 Revangle must be within the range 360 and different from zero If RevAngle is 180 a half cone is generated if its 360 a full cone is generated Positive revolution angles are defined counter clockwise lt Height gt The height must be different from zero If the height is less than zero the orientation of the cone is inverted Height is oriented in the Z direction on the Figure The Cone example shown was generated with the following code HH const N 16 const R 15 const H 10 Cone I N R270 H Cone shaped ceiling HHH Hint The cone can be made elliptical using the MScale statement The Dome statement The Dome statement generates a full dome half hemisphere covering the full 90 vertical angle In most cases the Dome2 statement is probably better suited The syntax for Dome is Dome lt Number gt lt NumberOfSurfaces gt lt Radius gt lt RevAngle gt lt optional name gt lt Number gt A
16. Corner numbers refer to the corners which must have been defined e g using the Pt or CountPt statements before using the surface statement The order of listing must be as obtained by travelling around the surface s edge in either direction The list of corner point numbers must be on the same line A room may contain up to 10000 surfaces Example 1 surface made from point 1 2 3 4 Surf 100 floor 1 2 3 4 Example 2 surface made from point 1 2 10 11 12 13 14 4 5 Surf 200 Ceiling 1 2 10 gt l4 4 5 If there is a need to programmatically build a list of points this can be done using the PList and ResetPList statements Building lists of points using PList and ResetPList The PList and ResetPList statements are used in special cases together with the Surf statement Twelve lists are predefined namely PList0 to PList9 and PlistA and PListB which are handled 162 automatically by ODEON when modelling af Box and other Solids The PList statements allow to programmatically construct a list of points e g a list like 100 110 120 130 140 150 160 170 180 190 200 this can be done using a for end construct in the following way adding a point number at a time for MyCounter 0 10 PList0 100 MyCounter 10 end It is also possible to add a number of points to a point list e g another PList to a PList In the following example PList is assigned the points 100 110 120 130 140 150 160 170 180 190 200 10 11 12 13 15 Plist Pli
17. EAC ga em A Ree E ates eR ter cee UT E Wr NE ESOS Ne ee RESALE ee er eR aA 4 00 Surface editor 3 00 3 00 Surface Description Layer oF O G Blue laye Y d g aay No description 1Bluelayer 5 00 G r 2 00 a i h 2 00 ea No description 1 Blue laye 5 00 g 5 a a a E a a a a A e gi N oh A E a E asia a No description 1Bluelayer 5 00 lt gt 6 Pez No description 1 Blue layer 5 00 1 00 1 00 Bea No description 1Bluelayer 2 50 oe S 0 00 0 00 r p i Mirror Vert mirror at 0 000 Metres Vetoa Horizontal Horiz Minor at 0 000 Metres 1 00 1 00 a x Z Extrusion Surface E 0 000 x 7 2 a00 0 000 Ei 0 00 12 00 14 00 je jeo jioa j 2 00 j140 pecai e eee X i 4 16000 6 000 Ed X 6 25 metres Z 9 00 metres Point input a Mouse Drag E Scroll RMB Zoom Alt LMB Select Point Ctrl LMB Move Point Ctrl Alt LMB Move Surface Shifts LMB New point LMB Start the program a shortcut to the program is found at the Windows Menu Start Programs ODEON OQDEONExtrusionModeler Initial settings Before starting modelling geometry select the drawing plane X s 6 Z anene Y or Z which is best suited for the geometry to be modelled Also Cor select properties for grid and snap spacing 5 a 2 Fr Using a drawing of the floor plan as the layout It is possible to load a 2D background drawing section plan as basis for the geometry in various image formats png jpg gif
18. M Rindel J H Christensen C L amp Gade A C 2010 Multi Channel Orchestral Anechoic Recordings for Auralizations Proceedings of ISRA 2010 29 31 August Melbourne Australia Zang M Tan K C amp Er M H 1998 Three Dimensional Sound Synthesis Based on Head Related Transfer Functions Journal of Audio Engineering Society vol 46 10 836 844 140 Appendix A Vocabulary The techniques of auralisation make use of many of technologies and a lot of technical terms and abbreviates are commonly used in the literature Here is a short vocabulary to some of the most used expressions the vocabulary is not a complete description of the individual words the context under which the words are used are many and the subjects are rather complex Anechoic recording Anechoic recordings are recordings of sound sources made without any reflections from the surroundings contributing to the recordings A common problem with anechoic recordings are that they may often include to many high frequency components because they are usually near field recordings and because they are recorded on axis where these components usually dominate When using such recordings with auralisation systems this may often result in unrealistic sharp s sounds especially in case of long reverberation times Anechoic recordings are usually recorded in an anechoic room but semi anechoic recordings may also be acceptable for use with auralisation systems
19. ie mas emf wmf and to use this drawing as the basis for the drawing To load a background drawing e Use the File Load backgound drawing menu to locate your image file e Double click on your desired origo in the drawing as requested could coincide with an intersection between a horizontal and a vertical module line e Double click at a point having some horizontal distance from the origo as requested could be another vertical module line 34 Enter the distance between the two points e g distance between vertical module lines e If the drawing is very dark this may be disturbing use the File Make background drawing lighter to lighten the drawing This can be repeated The drawing has now been scaled and is fixed to the drawing canvas when scrolling the drawing canvas or zooming the drawing will adjust appropriately If not satisfied with your scaling of the drawing repeat the process above E Left click the mouse at the positions where the points in the surface are desired if no points and Drawing an extrusion surface lines are generated then you need to bring the current surface in edit mode using the Insert or Esc shortcut or by double clicking a point in the surface Once all points in the surface have been defined finish the current extrusion surface by starting a new one using the ctri A shortcut or pressing the Insert or Esc shortcut To assign a drawing depth x Y or z and an extru
20. please visit www odeon dk dongle update Generating remote request file A remote request file has the extension req To generate a Remote request file follow these steps 1 Attach your ODEON dongle to the USB port on the PC 2 Run the ODEON program 3 Check what your current license includes using the Tools License information menu entry 4 Generate your license request file using the Tools Generate Remote Update Request file menu entry this will display the shown dialog 5 Email the generated request file e g User2005_Dongle102009 req to sales odeon dk Update option Update to most resent version of Odeon Update user name must be supplied separately Requested upgrade None Time license Requested update type No update to time license Update time limitation Remove time limitation Requested hours Once Odeon A S has received your request file an invoice will be processed and sent to you When the invoice has been paid an encrypted license file will be e mailed to you and this file has to be downloaded into your dongle for a license upgrade see below Updating the dongle downloading license file cif to dongle Extract the license update file from e mail to your hard disk e g to your desktop Attach your ODEON dongle to the USB port on the PC Run the ODEON program Use the Tools Download license update to dongle menu entry PR A a Select the file using the Select license
21. then you may start checking if the geometry is meaningful and without errors This may involve e Viewing the room in a 3DView e Viewing the room in the 3DOpenGL display e Analysing the geometry for unacceptable surface warps in the 3DGeometry Debugger e Analysing the geometry for unacceptable surface overlap in the 3DGeometry Debugger e Checking for missing surfaces in the room forming holes in the geometry The Unique edge s function available from the 3Dview may help you shortcut e Testing water tightness of the room tracing rays in the 3D Investigate Rays Or 3D Billiard window E Viewing the room in a 3DView The 3Dview displaying your room once loaded into ODEON has a large number of facilities which can be useful when creating and verifying geometries for ODEON 5 3 3D View Interior Exterior mode lo amp amp Corners in surface no 6010 409 xyz 26 100 14 300 139 700 56 xy z 29 700 15 000 139 700 55 xyz 30 300 12 800 139 700 42 xy z 37 100 14 600 139 700 44 xy 2z 37 300 12 650 139 700 46 xyz 37 100 12 850 139 700 46 xyz 37 300 6 650 139 700 50 xyz 37 100 6 650 139 700 52 xyz 37 400 0 000 139 700 150 xyz 37 100 6 650 139 700 148 xyz 37 300 6 650 139 700 146 xyz 37 100 12 650 139 700 144 xyz 37 300 12 850 139 700 142 xyz 37 100 14 600 139 700 155 xy 2 30 300 12 600 139 700 156 xyz 29 700 15
22. thus if you wish that certain surfaces are not glued together e g if upper and lower part of a wall should be assigned different materials either draw the surfaces on different layers in the CAD program preferable or turn off the glue surface s option may lead to an excessive number of surfaces and may not work with poly faces Max point margin If points in the DxF file are within this margin the points will be considered equal Allowing a certain amount of point margin will allow the Glue function to perform better if the coordinates in the model are not exact However if you have modelled both sides of a surface e g outside and inside surface of a balcony front edge the Max point margin should be smaller than the distance between these surfaces otherwise the points on either side of the surface will be considered the same with disastrous results Max warp ODEON will split four point surfaces into 2 three point surfaces if the surface s warp exceeds this value The glue option on the other hand will try to glue surfaces as long as this does not lead to a surface exceeding the max warp Editing the imported geometry It may be necessary or at least desirable to make changes to geometry after it has been imported The 3Dview display which displays the geometry once it has been imported is a useful tool for this purpose please see the context sensitive help available in this display from within ODEON for further details shortcu
23. twice first time through the directivity pattern i l which includes the overall frequency response second time through the recorded source signal which inherently include that response Calculating Adding acoustical inverse parameters I I I I I I I I For calculation of acoustical parameters it is spectrum 4 desirable that the true frequency content is I included in the directivity pattern in order for l example to correctly estimate SPL or STI ecu en j o Calculating BRIR gt However for auralisation the directivity 43 pattern should be equalised with the inverse spectrum of that recorded at the font axis of the natural source signal i e the wave file with human voice recorded with a microphone at the front axis ODEON version 8 5 and later can manage to create correct estimates of parameters from natural sources and at the same time create correct auralisation where the overall spectrum is only included once But it is necessary to use a source marked natural ODEON is installed with some directivity patterns which have the word NATURAL attached to their names e g BB93_Normal_Natural So8 When natural directivity patterns are selected from within Point Source Editor a green natural label is displayed next to the equalization entry fields If having existing directivity patterns of natural sources which are not marked natural this can be done using the Tool gt Directivity patterns Mar
24. uniformly and if using a value of 1 then all reflections are handled using Lambert or Lambert Oblique as was the case in ODEON 8 0 and earlier 91 9 Achieving good results The following section discusses how to obtain good results and indeed what is a good result It is not a straight answer as to how the best result is obtained merely a discussion that may provide some ideas as to what can be done in order to obtain reliable results in a program such as ODEON The desirable precision subjective limen Before discussing how to achieve good results it is a good idea to outline just what a good result is The subjective limen or just noticeable difference jnd on room acoustical parameters should give a good suggestion as to the desirable precision If the error between the real measured with some precision and the simulated room acoustical parameter is less the one subjective limen then there is no perceivable difference and the result is really as good as it can be so it would be senseless to look for more precise results In many cases it will be difficult or even impossible to obtain results at this precision and a poorer one will probably also be satisfactory for most purposes Parameter Definition Subj limen ISO 3382 1 2009 and IEC 60268 16 2003 for STI Reverberation time derived from 5 to 35 dB of the decay curve EDT s Early decay time derived from 0 to 10 dB of the decay curve Ds0 Deutlichkei
25. where a recorded signal for auralisation may be associated with the directivity pattern In order to select the right source for auralisation purpose please read chapter 3 Head Related Transfer Functions and digital filtering To create binaural simulations a set of HRTF s see Appendix A Vocabulary is needed The HRTF s are different from subject to subject and in principle you may measure your own ones and import those into ODEON using the Tools gt Create filtered HRTF menu entry in the ODEON program Measuring HRTF s is however a complicated task so you will probably be using the supplied ones If you should be interested in creating own sets of HRTF s for ODEON additional information can be found in the help available from within the ODEON program 56 The imported HRTF s to use for auralisation are pre filtered into octave bands in order to reduce calculation time The octave band filter parameters for the selected filter bank can be seen on the filter bank name at the Auralisation Setup menu The filter parameters are M The M value is taken into account if the Apply enhancement option is checked If the file name contains the word Mddd where ddd is a floating point number then localization enhancement was applied to the HRTF s Zang Tan amp Er 1998 This means that frequency dips and notches in the individual HRTF s have been exaggerated in order to improve the directional cues in the HRTF s The M factor determines ho
26. 0 01600 0 02500 0 02900 0 03400 24 4 SO OS to F Current obran 1973 5s 03 Low Limit 2 3004 Wooden floor on joests Ingersiev 0 22600 0 15200 0 14800 0 23000 0 11100 0 07000 0 08100 0 08900 187 57 S0 B 025 3 1948 S 02 High Limit 3 10007 Sold wooden door Bobran 1973 0 12200 0 07500 0 10200 0 09900 0 08600 0 15600 0 14500 0 13700 23 78 E 2 0 15 4 30 30 absorbent 0 32300 0 19900 0 27900 0 3600 0 22000 0 34000 0 31500 0 25000 74 a Ba TAES a 5 14307 a Aud21 Theatre 0 03200 0 03900 0 05500 0 0470 0 09600 0 0920 0 12000 0 05400 100 79 0 63 125 250 500 2000 8000 i e ae lt a SS a Re e 7 Frequency Hz n a C 63Hz 125 Hz 250 Hz w 500 Hz rs 1000 Hz D C 2000 Hz a C 4000 Hz E J 3 3 8000 Hz 0 2 0 3 0 4 0 5 0 6 0 7 0 8 0 9 0 11 0 13 0 15 0 17 0 19 0 21 1 1 1 2 1 3 1 4 1 5 1 6 1 7 1 8 1 9 1 11 1 13 1 15 Generation Individual Testing generation 1 Tracing rays 96927 of 132000 Individual 16 Remaining time h en s 00 00 05 Investigation of Simulation Parameters A tool for investigating how the number of rays and the transition order affect the agreement between simulations and measurements The tool calculates the average error in JND as a function of the number of rays Different curves are displayed for different transmission orders helping the user to determine the optimal number of rays and transmission order for the simulations 12 oe Calculating average error between simulated and m
27. 00 7 00 Reverberatio Vv 2 v S 0 00 2 50 v nd E_Omni50 0 00 50 00 T 30 vi 2 vi s E 0 00 2 50 v amp E_Omni80 0 00 80 00 5 Curvature_C Formula 1 a E 10 00 10 00 v m 6 Ts Centre time 0 v ms 0 00 200 00 W amp F J 7i SPI SPI 4 dA 10 00 1000 y i B Energy receptor to edit Type specific data for reverberation time EDT re eae e8y i ears acai Start 0 0 dB Display XI parameter ISO 3382 2 B 2 oe a Soo 100 a Figure8yC Figure 8zl Room acoustic wide band parameters Number Type Visible Decimals Y Origin Unit Manual Min Grid Max Grid Measured Description a SPL_A Formula Vv 1 E dB 10 00 10 00 Fj Aweighted SPL 63 8000 Hz E SPL_Lin Formula v 1 dB E 10 00 10 00 v Linear SPL 63 8000 Hz SPL_C Formula M 1 dB E 10 00 10 00 C weighted SPL 63 8000 Hz SPL_A_Direct Formula M 1 dB al 10 00 10 00 v A weighted direct sound presure level 63 8000 Hz STI STI v 2 W E 0 00 1 00 v Speech Transmision Index specfically STI Jnd ref J Acoust Soc Am STl_Female STl_Femal v 2 W 0 00 1 00 V Speech Transmision Index specfically STI_Female Jnd ref J Acoust STI_Male STI_Male v 2 v 0 00 1 00 v Speech Transmision Index specfically STl_Male Jnd ref J Acoust Soc RASTI RASTI T 2 v m 0 00 1 00 W Speech Transmision Index specfically RASTI Jnd ref J Acoust Soc Wide band parameter formula 1 SPL_A 10 Log10 Power 10 SPL_63 AW_63 0 1 Power 10 SPL_125 AW_125
28. 000 139 700 6409 xyz 26 100 14 300 139 700 ONAN RE i 5 gt z Corner closest to mouse pointer 1301 x y z 0 000 6 500 137 700 8007 ceiling part 3 8008 ceiling part 4 8009 ceiling part 5 8010 CEILING PART 6 Viewing corners and coordinates in a selected surface using the N shortcut and highlighting corner and displaying the coordinates of the corner closest to the mouse pointer using the M shortcut The perspective option shortcut P allows you to turn off the perspective of the room to get an isometric display of the room This may prove valuable when investigating warped surfaces The Unique Edges option in the 3DView display shows edges which only occur on one surface Such an edge is free it might be the edge of a free hanging reflector but it also could be the result of an error whereby two surfaces which should join along an edge do not 40 Example Modelling a box shaped room consisting of 6 surfaces but forgetting to define the 6t surface in the geometry file This room will have a hole where the 6 surface is missing The unique edge s option will show where the missing surface should have been 2 7 Combining geometries It is possible to combine geometries imported from an external CAD program with geometry modelled in the Extrusion modeller or modelled in the parametric modelling format of ODEON A geometry imported from a CAD program or generated in the Extrusion mo
29. 59 To obtain a reasonable correct level a first approach is to adjust the auralisation output against levels of some kind of sound in the room in which you are e g if you are simulating voice try to compare the level of the playback with the level of somebody speaking in your room This method should make it possible to make a rough adjustment and it s certainly better than none A more precise method is to use the calculated SPLa as a reference if it is calculated at an absolute level e Present the auralisation signal over a loudspeaker in the room in which you are sitting e Measure the sound pressure level in the room at the position where you will be sitting when listening to the auralisation and adjust the output level of the loudspeaker amplifier until the measured Laeq corresponds to the calculated level At this point you have a physical reference level which can be used for calibration of you auralisation playback level e Change between playing your auralisation sample over headphones and over the loudspeaker while adjusting the level of the auralisation playback until you are satisfied that the levels are the same This method is somewhat inspired by the old Barkhausen method for measuring loudness level in Phon and should at least in principle allow perfect calibration of the level the resulting level being within one subjective limen Headphones The binaural auralisation results created in ODEON are binaural signals wh
30. A 12 2 RECEIVER 1 T 30 3 2 20 1 95 1 99 2 01 1 76 1 81 2 Example of input file Measured results for Receiver 1 and 2 are to be imported For receiver 2 only the Ta parameter is available In most cases parameters are only available for the 125 to 4000 Hz bands For these a dash should be assigned to the 63 and 8000 Hz bands 126 Remember Click F1 while working in the Measured versus Simulated tabsheet to get the corresponding section in the help file which includes instructions on how to merge measured results to simulations Manual Optimization A manual optimization calibration of a room model against a real one is based on trial and error Usually two complementary acoustic parameters have to be used Complementary parameters are those ones which vary in different directions A common pair is T30 with Cso In principle as Tso increases Cso degreases and vice versa In addition Cso is a rather sensitive parameter that varies highly with position With an initial guess of materials a first simulation has to be carried out If measured results have been imported properly into the multi point response the Measured versus Simulated tabsheet will show the degree of the agreement If the agreement is not satisfying some materials need to be changed When small changes are needed for a material there is no need to change or define a new material in the Global or the Room material library You can just modify the absorption coefficient
31. Editor It is described how to view the files from within the Array Source Editor in the Testing if XML files section below SPL Lin 500 Hz 16 Having loaded one of the files into the Array Source Editor the Near field balloon Far field balloon and a 3D_ Direct display can be studied Note that these directivity patterns are strongly frequency dependent and have pronounced symmetry properties At high frequencies their patterns break down and show strong fluctuations see figure E4 and viewing an Octopol in the Near field balloon display may not prove interesting een Renee nt Uren t because it will never radiate any energy in its symmetry planes which is what is displayed in this window horizontal or vertical as selected on the other hand the The Quadropol directivity pattern viewed in the 3D_Direct display at 500 Hz Because the distances between the transducers are 1 metre the Quadropol pattern has broken down at 500 Hz Quadropol in the following figure gives a meaningful display tef Left 5 7 2 8 11 3 F 19 8 28 4 36 9 metes Rake 10 00 Odeon 1985 2009 Licensed to Odeon A S 10 00 metres front 10 00 metres front RAR l 10 00 Odeon 1985 2009 Licensed to Odeon A S The Quadropol sample viewed in the Near filed balloon tab in the Array source editor To the left at 63 Hz to the right at 4 kHz Testing if XML files can be imported correctly into
32. Number gt A unique number from 1 to 2 147 483647 for identification of the first point and surface in the Cylinder Using the same number but with negative sign defines the cylinder and its mirrored counterpart in the XZ pane Y 0 A Cylinder will take up several point and surface numbers which must all be unique lt NumberOfSurfaces gt For a full cylindrical room with a revolution angle of 360 around 16 to 24 surfaces are recommended For columns a number between 6 to 8 is recommended lt Radisus gt Radius of the cylinder must always be greater than zero lt Revangle gt Revangle must be within the range 360 and different from zero If RevAngle is 180 a half cylinder is generated if its 360 a full cylinder is generated Positive revolution angles are defined counter clockwise lt Height gt If the height is less than zero the orientation of the cylinder is inverted If height equals is zero one circular surface is generated Insertion point The insertion point of the cylinder is always the centre of the floor bottom surface Connection points The foot points in Cylinder are stored in PlistA The top points in Cylinder are stored in PListB The example shown was generated with the following code CylinderStatement Par HH const N 16 const R 15 const H 10 Cylinder 1000 N R 270 H TB Cylindrical room HHH Hint The cylinder can be made elliptical using the MScale statement The Cylinder2 st
33. ODEON Coordinate snapping When entering new points clicking the mouse coordinates of points will be truncated to the nearest snap e g the snap is 0 25 and X of the clicked position was 0 15 then the X coordinate of the point will be 0 25 if the Snap to grid option is enabled It is possible to fine adjust the coordinates in the Point editor lower right corner of the Extrusion modeller If Snap to existing points is enabled the X Y or Z coordinates of already defined points are also snap able Ortho snap Enable Ortho snap if the surface to be drawn have many edge angles of 90 or 270 degrees The option can be turned on and off while drawing a surface Lock equal points When this option is enabled it becomes possible to move points in multiple surfaces if their X Y or Y Z or X Z coordinates are equal This is useful for multi level rooms Snap to existing coordinates Makes it easy to draw new point making use of horizontal or vertical coordinates in existing points even when the do not match current snap size Lock H and V snap If this option is checked then Horizontal snap size will automatically be updated when the vertical snap size is changed and vice versa Relative or absolute extrusions Use the ctri H shortcut to toggle between relative or absolute extrusions in the Surface editor table When extrusions are displayed in relative measures an extrusion may be defined as Z 10 and dZ 5 telling that the extrusion starts at
34. ODEON Irrelevant drawing entities which are not supported Many CAD drawings are in fact 2D paper drawings rather than 3D models Such drawings do not contain sufficient information to create a 3D surface model and are ignored in the import process Examples of drawing entities which are ignored are circles dimensioning lines texts etc 2D drawing data may coexist peacefully in a drawing containing useful 3D data the 2D data are as stated simply ignored It is possible to convert a few 2D entities into model data useful for ODEON provided that it is done from within the CAD program se the 212D entities paragraph BLOCK s are not supported ODEON can not import entities which were inserted into a drawing as BLOCK s Any BLOCK in a drawing which contains relevant 3D surface data must be exploded using the EXPLODE command 28 before exported to the DXF file ODEON will notify the user if the DXF file imported did indeed contain BLOCK s 3D surface entities supported by ODEON e 3DFACE e Poly meshes MESH WEDGE PYRAMID BOX CONE CYLINDER SPHERE DISH DOME TORUS EDGESURF RULESURF and any other entities based on poly meshes e Poly faces the PFACE entity and any entity based on poly faces 22D entities supported by ODEON LINE POLYLINE CIRCLE ODEON can import LINE POLYLINE ARC and CIRCLE entities so called 22D entities when the elevation height is set to a value different from zero using the ELEV command in
35. Oblique Lambert The Oblique Lambert method allows frequency dependent scattering to be included in late reflections of point response calculations this option is recommended 90 Reflection Based scatter The Reflection based scattering method automatically takes into account scattering occurring due to geometrical properties such as surface size path lengths and angle of incidence The use of the method is recommended unless that part of scattering has already been included in the scattering coefficients assigned to the room s surfaces Interior margin Typical geometrical offsets in the boundary of the room default is 10 centimetres Surfaces which are closer to the boundary surfaces of the room than the distance specified by the Interior margin will also be considered boundary surfaces this means that surfaces such as doors and windows which may be modelled as being slightly on the inside of the boundary walls will still be considered as boundary surfaces Interior surfaces are displayed in a green colour teal in the 3DView whereas boundary surfaces are black so a change to the Interior margin will be reflected in this display when the Room Setup dialog is closed This measure tells ODEON that effective scattering provided by boundary room surfaces should be restricted below a frequency derived from this measure see chapter 6 5 for details To get an idea of our suggestions to this value please look into the geometries supplied wi
36. Remember The Play impulse response file tool plays back the whole file including the noise floor or any other content The Auralisation tool uses only the part between the onset and truncation times for convolving with the signal Therefore manual changing of onset and truncation times will affect auralisation 11 7 Examples of Impulse Responses Warnings A reliable measurement result relies on a healthy impulse response Although it is difficult to give a general definition on what a healthy impulse response is here is some guidelines that may help in obtaining acceptable results 1 Sufficient signal noise ratio at least in the octave bands of interest Tip If your equipment does not allow sufficient increase of gain for obtaining a good signal to noise ratio set the sweep length as long as possible eg 60 sec Every doubling of the sweep length results to a suppression of the noise floor by 3 dB As an example increasing a sweep signal from 2 sec to 64 sec will result to a 15 dB gain in the signal to noise ratio 2 A noise floor as flat as possible at the end of the response if the noise floor is included in the impulse response is a good indication that the impulse response is not polluted by impulsive noise 3 No presence of quantification artefacts due to poor recording resolution input level at recording of the impulse response should be high enough 4 Presence of spikes hills in the impulse response could indicate prese
37. a flutter echo can be removed changing the angle of a surface by a degree or two 2 3 Importing a model from SketchUp There are a number of choices available for modelling rooms for use with ODEON No matter the choice of modelling the room used for calculation with ODEON is stored in a file with the par extension e g Room par Tip The par file contains geometry in the Parametric modelling language file format This format can be edited manually in the ODEON text editor see Appendix D 25 Sketchup Our current most preferred program for room modelling is Sketchup SU We offer the SU2Odeon plug in that allows you to make direct use of SU models in ODEON SU is a 3D modelling software which is operated very intuitively and creates geometries directly compatible with ODEON It is available in a free educational as well as a pro version and can be downloaded at http www sketchup com download SU also allows you to create renderings of your rooms for impressive visualization To learn using SketchUp it is a must to watch some of the introduction videos which are found at http www sketchup com intl en training videos html In a day or so you should have learned the basics We have also created a few video on modelling for Odeon with SU at http www odeon dk tutorials acoustic simulation mod To make use of the SU models in ODEON the SU2Odeon plug in must be installed It can be downloaded from _http www odeon dk su2
38. a height of 10 and has an extrusion height of 5 If toggling to absolute extrusion then the same extrusion is displayed as Zi 10 and Z 15 telling that it starts at Z 10 and ends at Z 15 Modelling an array of surfaces Each extrusion surface has a set of array properties associated with it Nx Dist Ny disty Nz Distz one set for each of the three main orientations in the room These properties can be found in the Surface editor and define how many times the surface should be repeated in each of the main directions and the distances between the repetitions This feature is typically used in order to create a number of columns beams tables or chairs with a regular spacing When editing an array surface e g modifying a point all the repetitions of the surface will be changed accordingly 36 i Exploding an array of surfaces If individual changes are needed the arrayed surface must be exploded Once this operation has been carried out the surfaces in the arrayed surface has been turned into individual surfaces which can be modified surface by surfaces e g delete some of them It is not possible to perform the reverse of the explode operation so before exploding an arrayed surface make sure all operations common to the surfaces in the array have been carried out Define Layers For later management of acoustics materials inside ODEON the Extrusion modeller has layers This allow you to assign materials to all surfaces on one layer
39. a typical acoustic model looks which in principle differs substantially from an architectural model where all details mater for a visual impact In an acoustic model only the most important details should be kept In paragraph What to model in section 2 2 you can read more about the basic rules you have to follow Tip Find the pre stored rooms in the Rooms folder in your ODEON installation The usual path is C Odeon13Edition Rooms where Edition stands for Basics Industrial Auditorium or Combined You can see all relevant paths used by your ODEON installation by choosing Options gt Program setup Round Robin rooms In the rooms folder you will find pre calculated results in Elmia RoundRobin2 detailed par and PTB_Studio open curtains detailed model par which were the rooms used as test objects in the 2 and 34 Round Robins Bork 2000 Bork 2005 Geometry absorption data source and receiver positions as well as the measured room acoustical parameters are those supplied to all participants in the Round Robins by the PTB in Germany To compare results calculated by ODEON with those measured in the real rooms e Open the room in question e Open the JobList the Shift Ctri 3 shortcut e Select one of the pre calculated Multi point response jobs job one or two in the Joblist and open it Alt M shortcut e Select the Measured versus simulated tab sheet in the Multi Point display Pe vi N Ni
40. an impulse response arriving at the left ear which may typically have a length of interest of some 2 3 milliseconds approximately 100 samples at a 44100 Hz sample rate or if you prefer a length of 1 metre or so this is what is described by the HRTF s A set of HRTF s used for auralisation will typically contain a library for many different angles of incidence The HRTF s that comes with ODEON are those made available by Bill Gardner and Keith Martin at MIT Media Lab at http sound media mit edu KEMAR html as well as those from the CIPIC Interface Laboratory at http interface cipic ucdavis edu index htm If you have the capability of measuring HRTF s it is possible to import new sets for use with ODEON Scattering Coefficient It has values from 0 to 1 Defines the portion of energy that is reflected in a diffuse manner from a surface The rest energy is assumed to be reflected in a pure specular manner 142 Appendix B Specify Transmission through walls Since ODEON 9 transmission through walls can be simulated taking into account multiple transmission paths allowing walls to have a thickness and to have different materials on either side of the transmission wall Any transmission properties work completely independently from absorption This means that you shouldn t expect a material to absorb more sound because it has transmission properties For any transmission calculation in ODEON only 10 of the rays are let pass through a
41. anywhere on the impulse response to define a new onset time for the current frequency band If needed zoom the area of interest by dragging a rectangle from top left to right bottom The pink dotted line defining the onset time moves now to the new point If Onset snap window is active the onset time tends to snap at the strongest peak Disable Onset snap window when complete manual control is required Immediately when you click on another tabsheet all acoustic parameters are updated for the corresponding band In this way different onset times can be defined for each band To permanently save the changes click Measured Response gt Save Impulse response Manual truncation time Similarly to the onset time you can decide where you want to truncate the impulse response by enabling the Manual truncation time option Choose Measured Response gt Manual truncation time T shortcut The mouse cursor becomes a cross symbol when hovering on the impulse response Click anywhere on the impulse response to define a new truncation time for the current frequency band If needed zoom the area of interest by dragging a rectangle from top left to right bottom The red dotted line defining the truncation time moves now to the new point Immediately when you click on another tabsheet all acoustic parameters are updated for the corresponding band In this way different onset times can be defined for each band Restore automatic onset truncation times Choose the
42. are as follows 1 Visit the room and inspect the walls If no official data for the materials used on the natural walls are available an initial estimation has to be done and the closest materials in the ODEON library should be assigned Remember you can also add your own materials in the ODEON library by pressing the Add Materials CTRL M button in the local toolbar in the materials list 2 A group of impulse response measurements should be performed using the ODEON Measuring System see Chapter 11 or another application see step 7 The measurements have to be done at known positions inside the room The ISO standard 3382 presents the guidelines regarding the number of sources and receivers required for a given precision 125 A virtual model of the existing room should be made in ODEON where sources and receivers have to be placed at the same positions as in measurements Define one multi point response for each source in the room in the Joblist The measured impulse responses should now be loaded in ODEON using the Load impulse response tool see section 11 3 After inspecting the measured results and paying attention to warnings insert the room acoustic parameters into the multi point responses defined in step 4 See section 11 3 for how to import measurements to multi point response If your measured data come from another application as a txt file you can import them directly into the multi point response
43. are not familiar with coordinate manipulations it may be a good exercise to try different manipulations on the sample geometry above and load the geometry into ODEON upon each change Using layers in ODEON The Layer statement allows dividing geometry into separate parts which can be displayed separately and in its own layer colour in the 3DView 3DOpenGL and Materials list This makes it easier to model and investigate selected groups of surfaces When importing geometry from a dxf file e g from AutoCAD where layers are an integrated part the layers included in that file will be preserved in the imported version of the room If layers have been used in geometry the layer can be activated or deactivated in the 3DView 3DOpenGL and Materials list The layers menu is activated from these windows using the Ctrl L shortcut Vocabulary what s a layer Layers are commonly used in CAD modelling programs such as AutoCAD in order to make complicated geometries manageable Layers in CAD programs and some drawing and picture editing programs can be compared to overhead sheets without any thickness You define a number of layers with different names and possibly different line colour thickness etc and draw the different parts of your geometry on the different layers The layers can be turned on or off in the CAD program allowing better overview by hiding parts of the geometry that are not relevant in a part of the drawing p
44. around the Z axis specified in degrees All point definitions made after a UCS call will be created in the specified coordinate system The default coordinate system is defined as UCS 1 1 I 0 The UCS command corresponds to MReset MTranslate lt TranslateX gt lt Translate Y gt lt TranslateZ gt MrotateZ lt RotateZ gt 170 If the UCS command doesn t fulfil your needs for coordinate manipulation you may use the matrix manipulation family MTranslate MRotateX MRotateY MRotateZ MScale MPop and MReset Scale The Scale command is mostly there for compatibility with previous versions of ODEON the coordinate manipulation functions MTranslate MRotateX MRotateY MRotateZ MScale MPop and MReset included from version 4 21 allow far more flexibility For your own sanity it is not adviceable to mix the Scale method with the M family method The Scale command must follow the syntaks Scale lt ScaleX gt lt ScaleY gt lt ScaleZ gt The scale command will multiply scale all the points generated after the scale call using the specified x y and z scale The default setting is Scale 1 1 1 The Scale command evokes scaling of coordinates after all other coordinate manipulation is carried out If you should need more advanced scaling options please use the MScale option Coordinate manipulations the M family Advanced coordinate manipulation can be carried out using matrix manipulation The coordinate manipulation function
45. be used with the relative standing position to specify the distance from the centre of the lowest transducer to the physical bottom of the array loudspeaker The direction of the array is controlled from the acoustic centre of the array see fig 11 8 It may be convenient to define a receiver point to be the aiming point or the OpenGL option will let you look into the room from the centre of the array and the aiming point is the centre of the picture the crossing point of the two diagonals will indicate the exact point lbe tee tex e 1 0 0 0 3 0 5 0 8 1 0 metres 1 0 metres Path lt m gt 0 70 Refl order colour 0 1 2 3 4 5 6 7 8 9 Time lt ms gt 2 Odeon 1985 2009 Licensed to Odeon A S Dead balls 0 Figure 10 7 An array source with 7 units as specified in Table 10 1 The ray tracing in the room acoustic simulations is made from the acoustical centre shown just behind unit T4 The coordinates defining the position of the array refers to the cross on top of the array the hanging option 103 17 Measurement System A measurement system has been integrated into ODEON since version 12 so simulations and measurements of decay curves as well as room acoustic parameters can be done in the same software This allows ODEON to function both as a room acoustics simulation as well as a room acoustics measurement program Indeed both measurements and simulations can be displayed side by side at the sam
46. before the beginning of an impulse response Pressure domain 119 le Distortion Noise 27 14 dB at 1000 Hz use longer sweep Measured response C Odeon13Combined Measurements Errors DistortionOfficeS1 R4 wav Raw Impulse response at 1000Hz Raw decay curve at 1000Hz Decay curves at 1000Hz T 30 1 36 seconds Decay curves all bands Energy parameters Parameter curves Freque 4 gt C Odeon13Combined Measurements Errors DistortionOfficeS 1R4 wav Raw decay curve at 1000Hz M E Measured M Onset time M Truncation time M Noise floor SPL dB Wii pre Time seconds The impulse response shown on the previous figure but squared It is clearly seen that the harmonic distortion leads to a blurred onset of the impulse response Impulse response with distortion at the end Distortion at the end which appears as a notch of noise may have been caused due to presence of spikes in a measurement with the sweep method C Odeon12Combined Measurements B0_04 WAV Raw decay curve at 63Hz M E Measured Noise floor Onset time Truncation time SPL dB 0 0 1 0 2 0 3 0 7 Odeon 1985 2011 Licensed to Odeon A S tme seconds Impulse response with a prominent notch at the end Such a type of hill indicates that strong spikes were present during the measurement When using the Sweep method spikes in the recording are converted to hills after the deconvolution 1
47. by 2 dB per unit from bottom to top 102 10 5 Using the equalizer In addition to the general equalizer for the array there are equalizer options for each transducer This may be used for fine adjustments or to turn off some transducers at some frequencies E g if you want to use a shorter array at the high frequencies you can enter a high attenuation of the transducers in question 10 6 Bringing the array into the room The position of a line source in a room is similar to that for a simple point source but with some extra options e If the transducer coordinate system is Absolute it means that the x y z coordinates of the array indicate the position of the transducer with relative coordinates 0 0 0 if any there need not be a transducer in this position but all transducer coordinates are relative to this point e If the transducer coordinate system is Relative hanging it means that the x y z coordinates of the array indicate the position of the transducer in the top of the array and the coordinates of the other transducers are relative to this e If the transducer coordinate system is Relative standing it means that the x y z coordinates of the array indicate the position of the transducer in the bottom of the array and the coordinates of the other transducers are relative to this e In all three cases it is possible to include an additional offset of the coordinate system For example this can
48. calculated on a theoretical basis taking into account the reverberation time of the room and the background noise but not the explicit shape of the impulse response The STI Expected is the value of STI if the field in the room was totally diffuse and the energy decay was exponential RASTI An abbreviation of Rapid STI it is considered an obsolete parameter in the revised IEC 60268 16 2003 standard It is based only on two Octave Bands 500 Hz and 2000 Hz and provides a faster and simplified calculation of the STT DL Rate of Spatial Decay Rate of spatial decay is the decay of sound pressure level per distance doubling DLz is calculated according to ISO 14257 2001 The DL2 parameter is intended to characterise the acoustic performance of workrooms The values to be expected for the DL2 parameter is according to Ondet amp Sueur 1995 1 3 dB for reverberant rooms and 2 5 dB for ideally treated rooms The design criterion for DLz is set to 3 5 dB or better according to ISO 11690 1 1996 The DL2 parameter is calculated as a part of the Multi Point response if the job only contains one active source the active source is a point source more than one receiver is defined and the distance between the source and the receivers are not the same for all receivers Please notice that one misplaced receiver may ruin the entire DL2 calculation thus it is a good idea to check the receiver positions or even better to check the individua
49. dB is probably sufficient Astop Maximum possible attenuation of octave bands To allow complete attenuation of all reflections of a 16 bit signal 96 dB dynamic range Aso Should be 96 dB however due to auditory masking we are not able to hear such differences so 40 dB is probably sufficient Smaller Asp leads to shorter calculation time of the BRIRs Band overlap in percent Octave bands implemented using FIR filters are not completely rectangular it takes some frequency span before they attenuates completely An overlap between the filters of 100 percent gives a smooth transition between the filters which is probably a more realistic representation of real world reflections than shorter overlaps At the same time long overlap gives shorter calculation time of the BRIRs 57 If you should need to use filters with other filter parameters e g Asi being 96 dB you should create a filtered set of HRTF s with these parameters use the File Create filtered HRTF s option Then from within your room select the new filter bank from the Auralisation Setup If you should need to import other HRTF s than the Kemar Gardner amp Martin 1994 which are installed but do need to be imported or Subject_021Res10deg Algazi 2001 which are installed as well as imported you should create a text file following the same format as used in the files Unity ascii_hrtf Kemar ascii_hrtf and Subject_021Res10deg ascii_hrtf These files can be found in th
50. dependent and as such it is not known before the source and receiver are defined and the actual ray tracing or image source detection takes place An example on this is that a desktop may provide a strong specular component to its user whereas it will provide scattered sound at remote distances The method has several advantages not only does it make life easier because the same scattering coefficient can be used for different surfaces no matter their size it also allow better estimate of the actual scattering occurring at a reflection point because scattering caused by diffraction is not fully known before the actual reflections are calculated thus angles of incidence path lengths etc are known In order to allow these features to be included in predictions the reflection based scattering coefficient sr combining the surface roughness scattering coefficient ss with the scattering coefficient due to diffraction sa is calculated individually for each reflection as calculations take place s l d s d s The formula calculates the fraction of energy which is not specular when both diffraction and surface roughness is taken into account 1 sa denotes the energy which is not diffracted that is energy reflected from the surface area either as specular energy or as surface scattered energy the resulting specular energy fraction from the surface is 1 sa 1 ss S Surface scattering Set of scattering coefficients Surface
51. e g the Lambert Oblique Lambert or Uniform directivity see later 70 Summarizing the calculation method used for point response calculations in ODEON As described above the point response calculation in ODEON is divided into a receiver independent and a receiver dependent calculation part The division into two calculations is solely done in order to save calculation time by reusing parts of the calculation where possible One of the advantages of the late ray tracing method used in ODEON compared to more traditional methods is that rays do not even have to come near to the receiver to make a contribution and a late ray being reflected 100 times has a potential of generating 100 secondary sources Thus even in coupled rooms with only a modest number of rays it is possible to obtain a reasonably number of reflections at a receiver which is required to obtain a result that is statistically reliable This results in a fine balance between reliability of the calculation results and short calculation time A complete histogram containing both early and late energy contributions is generated and used to derive Early Decay Time and Reverberation Time The other room acoustical parameters are calculated on basis of energy collected in time and angular intervals For surface and line sources a number of secondary sources are placed randomly on the surface of the source each emitting one ray and radiating a possible contribution to the receiver T
52. early late and total energy and are described in more detail in Gade 2003 Early Support or ST1 E ST ea rly dB Foto Late Support E ST 1001000 qB Eoo Total Support E ST ota a dB 0 10 STearly or STi is used as a descriptor of ensemble conditions i e the ease of hearing other members in an orchestra STiate describes the impression of reverberance and STota describes the support from the room to the musicians own instrument If the early late averaging is turned ON averaging in time is performed as for the other parameters In case of the stage parameters the following limits of time intervals are used 9 ms 10 ms 11 ms 18 ms 20 ms 22 ms 900 ms 1000 ms and 1100 ms 2 j Average was known as LG80 in earlier versions of Odeon Only available in Odeon Auditorium and Combined 84 IACC Inter Aural Cross Correlation Since the hearing process is binaural the spatial impression in a concert hall can be quantified by measuring the inter aural cross correlation between the left and right ear canal ISO 3382 1 2009 The Inter Aural Cross Correlation is expressed as a coefficient from 0 to 1 First the normalized inter aural cross correlation function is calculated P O p t 2 dt IACF t gt fo nat p tdt where p t is the impulse response at the entrance to the left ear canal p t is the impulse response at the entrance to the right ear canal Then the inter aural cross
53. import a DxF file e Select Files gt Import from file dxf 3ds cad or simply drop the file on the Canvas of ODEON e Specify the input file e g MyCADRoom dxf e Specify the destination file e g MyCADRoom Par Once the file names have been specified the Import DXF file dialog appears allowing miscellaneous import options to be specified By default most of the parameters may be left untouched however it is important that the correct drawing unit is specified If the geometry does not appear as expected you may try other input parameters 30 Unit in input file Unfortunately dxf files are unit less It is important that the correct unit in which the geometry was modelled is selected in the import dialog If the correct unit is not specified the import process may fail because the geometry seems to be only a few millimetres large or several kilometres in size Geometric rules glue surfaces Surfaces imported in the DxF format are put simple by nature surfaces build from three or four sets of coordinates When the glue option is turned on ODEON will try to glue or stitch if you prefer these surfaces in order to form fewer surfaces with larger areas Do note that some surfaces based on poly faces may not import correctly unless the glue option is turned ON and Don t allow subtractive PolyFace algorithm is turned OFF ODEON will not combine surfaces with each other when they are situated on different layers in the CAD drawing
54. in room acoustical modelling However good modelling practice will greatly reduce the time used for modelling and re modelling rooms In order to study a room in ODEON a file containing the room s geometry needs to be created This file has to be imported by one of the ways described below ODEON then creates a group of subsequent room files calculation files and result files All files share the same name the name of the room with a different extension added Three main categories of files are used e Room files which contain all data entered by the user From these files both calculation and result files may be recalculated Some of the most important extensions are par that describes the whole geometry PcS that lists the sources PcR that lists the receivers and Mat that includes the materials used in the model e Calculation files which holds data that are being reused during calculations to speed up calculation of point response calculations These files can be erased once job calculations have been carried out however if they are present future calculations are speeded up e Result files containing all results For example the extension GXX refers to the results in a grid response See the full list of files in the ODEON help by searching for File management Remember Pressing the F1 key reveals the ODEON help menu with a topic associated with the window that is open The file containing the room model must be wri
55. more than one gene The criterion that is used from the evolution process to create an improved generation is the fitness function Properties of the individual that give good fitness between simulated values and target values will have better chances of propagating into the following generations The reason that GAs have become popular is their ability to find useful solutions in a very complex search space having many minima maxima without getting stuck in the first occurring local minimum maximum In our optimization problem there are eight different GAs that run independently for each octave band Translating the foregoing terms to our problem an individual consists of a complete set of absorption coefficients for a particular frequency band corresponding to the different materials in the room Henceforth we shall use the term material for describing its absorption coefficient One material is one gene The terms are shown in Table 1 All frequency dependent GAs start with a random pass Monte Carlo method where all individuals of the population are generated with absorption coefficients that vary randomly according to a specified range This can be called 0th generation After this pass the evolution process is initiated by filtering out the best individuals as parents and producing children that are likely to inherit some of the advantages from their parents The process continues for ever until the solution fully converges criterion is satisf
56. of origin a secondary scattering source is created The Late Reflection method All reflections that are not treated by the early reflection method are treated by the late reflection method At all the reflection points of late part of the receiver independent part of the ray tracing a small secondary source is generated for reflections above the Transition order This secondary source may have a Lambert Lambert Oblique or Uniform directivity depending on the properties of the reflection as well as the calculation settings ODEON checks each secondary source to determine whether it is visible from the receiver The late reflection process does not produce an exponential growing number of reflections with respect to the time as the image source method would suggest but keeps the same reflection density in all of the calculation allowing for reasonable calculation times The attenuation of a secondary source is calculated taking the following into account e Directivity factor of the primary source in the relevant direction of radiation point sources only e Reflection coefficients of the walls involved in generating the image can be angle dependent if Angular absorption is enabled in the Room Setup e Air absorption due to the length of the reflection path e Distance damping due to the distance travelled from the primary source to the receiver is inherently included in the ray tracing process e Directivity factor for secondary sources
57. option Restore automatic truncation times Or Restore automatic onset times and truncation times for restoring the original settings by ODEON for the selected impulse response Save Impulse Response When you make a manual change the option Save impulse response becomes active so that any new onset or truncation times can be saved with the impulse response file Click on the option define a new name and click OK 11 6 Making auralisations with Impulse Responses Auralisation facilities have been included in wana Leal the measuring system since ODEON 13 This ee el J allows convolution of the measured impulse i ee response with an anechoic signal and canion listening to the result as if this signal was og J b played inside the real room However quality Play Convolution might be limited by the loudspeaker or the 114 microphone used for the impulse response recording Choose Play impulse response file P from the Measured response menu to listen to the impulse response Choose Auralisation A to listen to an anechoic signal convolved with the impulse response In the Input signal section you can make your folder and file selection while in the Convolution section you can choose the location of the convolved file The file receives its name from the name of the impulse response in use and by default it is stored in the Temp folder meaning that will be deleted when shutting down Windows
58. proces of obtaining new sets of materials Measured acoustic parameters Fitness The error between simulated and measured data for an individual Should be minimized Fitness function Calculation of fitness according to equation 1 129 Search Space In order to optimize the search process it is important to limit the search space The search space can be limited by telling the GA that some of the absorption coefficients should only vary within certain limits and indeed that some should not be changed at all This variation can be called search range and is given in percentage A value of 100 would lead to a search range from 0 to 1 absorption coefficient regardless the initial values A value of 0 leads to no change at all meaning that the material is excluded from the optimization process A search range between 0 and 100 gives lower and upper limits depending on the initial absorption coefficient Careful estimation of the search range is crucial for achieving realistic solutions For example if it is suspected that two hard parallel walls may cause a flutter echo once extra absorption is installed in the refurbished room then it is important to restrict the absorption coefficients and search range to low values e g maximum 2 If a material is only installed on a small surface area or it is believed to be well known it should be assigned a search range of 0 Omitting some materials from the optimization process will reduce the calc
59. process carried out in order to estimate the room volume assumes scattering coefficients of 1 for all surfaces rather than using the coefficients assigned to the surface in the materials list since the mean free path formula is based on diffuse field assumptions The value of S used here is the sum of the areas of non transparent surfaces taking into account whether one two or indeed none of the sides of a surface are visible inside the room Convergence criterion A certain number of particles must be sent out and followed around the room for a stable estimate to be obtained More and more particles are sent out in random directions until the value of the reverberation time has remained within 1 for at least 50 particles At the end of a run the data on how many times each surface was hit is stored Then if new materials are assigned to the surfaces the reverberation times can be recalculated instantaneously without repeating the particle tracing S Global Estimate This method estimates the global reverberation times T20 Tso using the method proposed by Schroder Schroeder 1970 as well as mean free path and generates estimates of decay curves Particles are sent out in random directions from the source see section 6 8 and reflected using the Vector based scattering method see section 0 ODEON records the loss of energy in each particle as a function of time occurring because of absorption at room surfaces and in the air S
60. should be aware of when upgrading from previous versions Upgrading to version 13 ODEON version 13 runs on Windows 8 32 64 bit editions Windows 7 32 64 bit editions and Windows Vista 32 64 bit editions ODEON 13 is a major upgrade please see section 1 3 Though ODEON 13 is capable of loading and converting older projects into the version 13 format forward compatibility an older version of ODEON may not load a room once it has been loaded into ODEON 13 ODEON 13 includes a large number of enhancements for a fairly complete list see section 1 3 or open the help file from within the ODEON software press F1 from within ODEON then click Whats new in ODEON 13 entry in the Contents menu Upgrading from versions earlier than version 8 When upgrading from versions earlier than version 8 it is essential to learn about the new methods for handling of scattering Chapter 4 covers the material properties to assign to surfaces chapter 6 covers the calculation principles including handling of scattering and chapter 8 covers the choice of calculation parameters Major upgrade If performing a major upgrade typically a full version number or more e g from version 8 to version 9 then ODEON will install to a new directory for that version without changing the existing installation If you have no wishes to use the old version of ODEON then it is recommended to uninstall the version s using the Windows Start Control panell Add remove program
61. specularily and each time a new combination of reflecting walls are found a new image source is added to an image source tree The number of images stored in the image source tree after this process is usually much lower than all combinations of reflecting walls but the images are all valid in the sense that each of them can be seen at least from some receiver position in the room Late part finding the secondary sources The late ray tracing is different A number of Late Rays specified in the Room setup are emitted from the source and reflected according to the vector based scattering method described later taking into account scattering due to surface size and surface roughness At each reflection point of a ray a secondary source is generated if the reflection order is above the Transition order The early and late processes described use ray tracing which is carried out taking into account the Impulse response length and the Max Reflection order as specified in the Room setup This part of the calculation only stores geometrical information and does not involve material data nor receiver position 68 Remember Note that this receiver independent process stores large files containing ray history These files being present after a point response calculation will soeed up future calculations making use of the same sources However if these file are not erased when finishing a project you may end up with a full hard drive Either use the Fil
62. statement to define a series of vertical surfaces from a series of perimeter points plus an elevation height The perimeter points are typically defined using the MPt statement The syntax of ElevSurf is ElevSurf FirstSurfaceNumber gt lt FirstPointNumber gt lt SectionsInElevSurf gt lt Height gt lt Optional name gt Example on use of the MPt and ElevSurf statements First the perimeter points point 1 to 23 at the floor level of an office environment are described using the MPt statement Then the elevation surface is created from these points creating the perimeter walls of the office with a constant height of 2 7 metres Finally the floor and ceiling is created using the Surf statement 0 00 110 00 120 00 130 00 140 00 metres Perimeter points at the floor level Example file MPt and ElevSurf Par Demonstrates the use of MPt multi point Surf Surface and ElevSurf Elevation surface statements In this example the X coordinates are made in absolute values whereas the Y coordinates in most cases are in or de creased using the or options To create a closed ElevSurf that is the first wall joins the last wall first and last point in the series of points handled to the ElevSurf must be identical in this example point 1 and point 23 are identical If an elevation surface has 22 surfaces then 23 points must be made available to the ElevSurf as in this example HHH MPt 1 23 49 62 2 02 164 48 22 1
63. the calculation speed and offer better results without loosing accuracy in the acoustical geometry How to model an audience area Modelling each step between the rows in an audience area is not recommended the audience area can be simplified a lot without compromising the quality of the results in fact using the suggested 24 method below is likely to produce better results e Model the audience area as audience boxes with a height of approximately 0 8 metres above the audience floor In the figure on the right audience boxes are indicated by blue color e Assign appropriate absorption material e g ODEON material 11001 e Assign a high scattering coefficient of 0 7 to the surfaces of the audience box See in Chapter 4 for how to assign scattering coefficient on a surface e Position the receivers around 0 4 metres above the modelled audience box An obsolete alternative solution was to model the audience just as a flat absorptive surface on the floor mainly for simplifying the model for computational issues The main problem with this approach was that the absorption area of the room looked smaller since more rays were likely to hit a hard surface like the isles between audience areas than in real life In order to compensate for that the same absorptive material had to be assigned to the isles too Nowadays there is no need to follow this approach Modelling audience as a box has been proved to be
64. the 3DGeometry debugger in ODEON ODEON will generate a list of warnings and a corresponding illustration in a 3D display whenever an overlap or a warp exceeds the value specified in the Room setup Model Air conditions dialog Overlapping surfaces is a tricky problem because it is usually invisible on 3D projections of the geometries however such errors in the model may lead to unpredictable results so always check models of some complexity for overlapping surfaces Testing Water tightness using 3D Investigate Rays Testing a new model for water tightness i e whether it is completely closed may be done using a 3D Investigate Rays Window The room model may not be watertight if e Surfaces are missing from the model e Surfaces are unacceptable warped e Boundary surfaces have been assigned transparency coefficients greater than zero e Boundary surfaces have been assigned Material 0 transparent e Sources are located outside the room Before investigating ray tracing you will have to e Make the boundary surfaces of the room solid by assigning materials to them For the moment it does not matter what the materials are as long as they are not transparent Material 0 or fully absorptive Material 1 Go into the Materials List and assign e g 20 absorption to all surfaces use Ctri Ins to do this in one keystroke e Place a source somewhere inside the room Sources are defined from the Source receiver List At first it may b
65. the factor is one and if the angle is 90 the factor becomes its maximum of two because half of the balloon is outside the room Factors for angles between 0 and 90 have been found using numerical integration Oblique Lambert attenuation factors Cc Oo O eb O O Cc O 5 O a So O O 20 30 40 50 60 70 80 90 Oblique angle Figure 0 3 Correction factor for Oblique Lambert When the oblique angle is zero Oblique Lambert corresponds to traditional Lambert and the correction coefficient is one When the oblique angle is 90 corresponding to grazing incidence on a smooth surface the correction factor reaches its maximum of two A last remark on Oblique Lambert is that it can include frequency depending scattering at virtually no computational cost This part of the algorithm does not involve any ray tracing which tends to be the heavy computational part in room acoustics prediction only the orientation of the Oblique Lambert source has to be recalculated for each frequency of interest in order to model scattering as a function of frequency 6 7 Diffraction over screens and round objects Screen diffraction When no direct sound is received from a point source at a receiver position ODEON will try to detect a one or two point diffraction path using the methods suggested by Pierce in Diffraction of sound around corners and over wide barriers Pierce 1974 By only calculating the diffracted contribution
66. this could be out door recordings of machinery trains etc or studio recordings of music Auralisation auralization The term auralisation was invented by Mendel Kleiner who gives the following definition Auralisation is the process of rendering audible by physical or mathematical modelling the sound field of a source in a space in such a way as to simulate the binaural listening experience at a given position in the modelled space When used in ODEON one may think of auralisation as the art of creating digital simulations of binaural recordings in rooms which may not be build yet The aim is to provide the same three dimensional listening experience to the listener as would be achieved in the real room at the given receiver position with the simulated source position s and signals Binaural recording Humans usually listens using two ears This allows us to perceive sound as a 3D phenomenon To create a binaural recording it is not enough to create a two channel recording stereo also the colouration created by diffraction from the human body has to be included This is usually done by using a dummy head with a microphone mounted at the entrance of each ear canal this recording may be recorded using an ordinary stereo recorder but is now refereed to as binaural Binaural recordings are usually played back through headphones to avoid colouration from the room in which it is played as well as avoiding diffraction from the human
67. to another PC with a different material library To make such a new material take effect in the room reassign the material e g using the Global Replacement option Another option is to change the absorption coefficients of a room material using the edit fields below the surface list in the left side of the materials list doing so will change the absorption of all surfaces which have been assigned that material the material will not change in the material library in the right side of the Materials List Scattering coefficient While thw absorption coefficient controls the amount of sound energy that is absorbed from a surface the scattering coefficient determines the way sound energy is reflected We can group relfections in rooms in three categories e Specular reflections where the angle of reflection is equal to the angle of incidence 51 e Diffuse reflections where the angle of reflection is independent of the angle of incidence memoryless reflection e A mixture of the cases above A scattering coefficient is assigned to each surface This scattering coefficient accounts for the roughness of the material at the mid frequencies around 700 Hz and it is expanded during calculations in order to take into account the frequency dependent behaviour of scattering using typically frequency functions for scattering coefficients This coefficient is taken into account during the ray tracing if Room setup gt Calculation parameters gt Scatterin
68. trick is that the Invert phase option has been applied to half of the transducers in the Dipol Quadropol and Octopol arrays in order to get the special behavior of those source types The Dipol_Domain_Frequency xml file is essential identical to the Dipol xml and will give same results when used inside ODEON but this 146 file has been defined using Domain Frequency which is the option that can be used when defining beam steered arrays Nodes Array Source at Position Transducer za 1 OffSet Description i Z ama D O e Domain Frequency Domain Octave Tree structure for the Array source node The Domain Data nodes are described separately in figure E2 Domain Frequency Octave Octave Octave 63 Hz 125 Hz 250 Hz Sub Re Img Domain Octave Ea C octaves Sub Tree structure with the details of the Domain Data node This note comes in two versions depending on weather Domain Octave or Domain Frequency 147 x NG D Images of the Mono Dipol Quadropol and Octopol directivity patterns installed with Odeon Auditorium and Combined as viewed in the Far field balloon tab in the Odeon Array Source Editor 148 Viewing the sample files in ODEON The XML files are text files and can be viewed in a text editor such as ODEONEdit Because the sample files follow the format outlined in this description they may also be viewed in the ODEON Array Source
69. uk 3daudio html Gade A C 1997 The influence of basic design variables on the acoustics of concert halls new results derived from analysing a large number of existing halls Proceedings of the Institute of Acoustics Vol 19 Pt 3 Gade A C 2003 Room acoustic measurement techniques Chapter 4 In Room acoustic engineering Note 4213 Lyngby Denmark Acoustic Technology Technical University of Denmark Gardner B amp Martin K 1994 May 18 HRTF Measurements of a KEMAR Dummy Head Microphone Retrieved June 16 2011 from http sound media mit edu resources KEMAR html Gerzon M A 1992 General Metatheory of Auditory Localisation Preprint 3306 of the 92nd Audio Engineering Society Convention Vienna Ingerslev F 1949 L rebog 1 bygningsakustik for Ingeni rer in Danish Copenhagen Teknisk Forlag Ingerslev F amp Petersen j 1953 Lydabsorberende materialer in Danish Arkitektens Ugehefte no 3 Insul n d Retrieved June 16 2011 from Marshall Day Acoustics http www insul co nz Knudsen V O amp Harris C M 1950 1978 Acoustical Designing in Architecture Acoustical Society of America Kristensen J 1984 Sound Absorption Coefficients Measurement evaluation application Note No 45 in Danish H rsholm Denmark Statens Byggeforskningsinstitut Malham D 2005 January 21 Home page for Ambisonics and related 3 D audio research Retrieved June 16 2011 from Music Technolog
70. use ODEON S Genetic Material Optimizer for calibrating a room model towards a real room The Genetic Material Optimizer makes use of genetic algorithms in order find the list of materials absorption coefficients that lead several simulated parameters to match with measured ones END USER LICENSE AGREEMENT This legal document constitutes the complete agreement between you the end user either an individual or a single entity and Odeon A S All rights not expressly granted in this License Agreement are reserved by Odeon A S The right to use the Odeon software is sold only on the condition that the user agrees to the following License Agreement COPYRIGHT PROPRIETARY PROTECTION The Odeon software auxiliary applications and documentation are owned by Odeon A S and are protected by copyright laws and international copyright treaties as well as other intellectual property laws and treaties The SOFTWARE PRODUCT is licensed not sold You shall be held legally responsible for any copyright infringement that is caused or encouraged by your failure to abide by the terms of this Agreement OWNERSHIP OF ODEON Odeon A S retains sole title and ownership of the ODEON software regardless of the form or media in or on which the original and other copies may exist As the licensee you own the magnetic or other physical media on which the ODEON program is subsequently recorded or fixed This Agreement is not a sale of the original ODEON software o
71. user is able to edit the room acoustic parameters calculated in the point response or to define his own ones The Room acoustic parameter list window allows several manipulations such as change of the limits for the reverberation time or editing the formula corresponding to each room acoustic parameter An additional feature is the manual definition of the grid range The user can define the type of a parameter Reverberation T30 SPL Centre Time Formula Dietsch The values of the energy intervals on the left of the window can be used as variables in a parameter expression When someone wants to use an energy interval from a specific time point t to the end of the response he has to make use of the predefined whole response length and subtract the interval from t 0 sec to t e g E_Omni50ms_Total E_Omni_Total E_Omni50ms For a formula type the equation can be written explicitly in the white box A list of all mathematical expressions available in the edit formula mode is given in Appendix E 87 G Calculation Parameters Room Setup and Define Grid The Calculation parameters specify how calculations are carried out in ODEON and as a starting point ODEON suggests values that can be considered suitable for most calculation parameters Select one of the presets Survey Engineering or Precision representing the quality of results desired ODEON will adjust the calculation parameters accordingly There are a few additional parameters in t
72. very fast in the Material list in ODEON Create layer with a name and colour in the layer box above the surface editor You can deactivate layers to make them invisible in the extrusion modeller but deactivated invisible layers will still be loaded into ODEON The current layer ticked as Current is the one that new objects are created in By checking Move all layers on layer it is possible to multiple surface in one operation It is possible to change the layer of an extrusion surface in the surface editor allowing you to group surfaces on one layer prior to moving them Odeon extrusion modeller version 3 0 C Users Claus Documents Odeon filer Rooms oes Office inventory OES 9 boh File Edit Help Z 20 h30 h20 h s50 hs h70 Modeling plane 9 r YZ Vertical cross sechon xr Horizontal XZ Verica ength Grid and snap options J Snap to gnd Ortho snap Lock equal ponts Honzontal Vertical Grid spacing 0 20 0 20 Metres Snap sze 02 0 25 Metres Sutlace ty zo nz gR zs Ms ss T EEL E g 7 O Minor Vest miror at 0 000 Metred Vertical Heewornta Horiz Mince at 0 000 Metreg Z Poirt editor 20 1 30 j140 p 50 j1 0 11 70 2 Pore 7 Y Extrusion Surface x x X 1 75 metres 1 00 metres Point input Locked Mouse Drag Selection Scroll RMB Zoorn Alt LMB Select Point C The easiest way to model a chair like the one in the left Figure is to simply enter the same data as displayed in the screenshot
73. which are elements in the boundary of the room such as windows doors paintings blackboards etc one should however not expect these elements to provide effective scattering down to infinitely low frequencies From diffuser theory it is found that typical behaviour is that the effectiveness of a diffuser decreases rapidly below a cut off frequency which can roughly be defined from the depth of the diffuser wall construction being less than half a wave length Two octave bands below the cut off frequency the diffuser is no longer effective At the lowest frequencies however the dimensions of the room will provide some diffraction therefore the dimensions of the reflecting panel as used in the formulae for f and fw are substituted with the approximate dimensions of the room at the lowest frequencies and a combination of surface and room dimensions are used for frequencies in between high and low frequencies It is worth noticing that it is not only the depth of the wall construction which enables the elements of the wall construction to provide diffraction also angling between the surfaces offsets e g the door being mounted in a door hole or the surfaces being made of different materials provides the phase shifts which results in diffraction Therefore it may be reasonable to assume that the boundary walls have a minimum depth of say 10 cm in order to account for such phase shifts Boundary Furniture Interior Margin Ground plan of a room wit
74. 0 5 The transparency coefficient can be assigned values between 0 and 1 e 0 0 is assigned to all solid walls This value should always be assigned to the boundary walls of the room otherwise rays will escape from the model e Very small transparency coefficients should be avoided unless the number of rays is increased substantially Instead consider modelling the surface as solid Using a transparency coefficient greater than zero will cause the Image source method to be discarded for rays hitting such surfaces only relevant for point sources Another problem is that only very few rays will be transmitted making the results on the other side of the surface statistically unreliable e Very large transparency coefficients e g 0 95 should also be avoided Instead consider removing the surface from the model An easy way to do this is to assign Material 0 transparent to the surface The Transparency value should not be used for modelling sound transmission through walls instead use the wall type for this purpose see below Tip If assigning scattering and transparency coefficients between 1 and 100 ODEON will assume that the values are in percent and divide by 100 Type The wall Type can be set to Normal Exterior Fractional Transmission The type Normal Exterior and Fractional relates to the way Reflection Based Scattering is calculated for sound reflecting from the surfaces Transmission is for walls which transmit sound to anothe
75. 0 for more information about sources and directivities 16 1 5 How to upgrade or update your current license For ODEON 9 1 and later versions your ODEON software license is stored in a Smart Card based hardware key the Rockey 6 Smart dongle As an ODEON user you can have 4 ways of updating or upgrading the program after the installation of ODEON 9 1 Version number can be updated free of charge in between the full versions at www odeon dk E g from version 9 1 to version 9 2 Your version can be updated to a full version number E g version 10 by requesting a remote licence update For time limited licenses time limitations can be removed or the Run time hours available can be updated by requesting a remote licence update The Edition can be upgraded to Combined if the current edition is Basics Industrial or Auditorium by requesting a remote licence upgrade Your wishes for update or upgrade could be a combination of the above 4 points which can be fulfilled by generating a remote request file and following the steps described below Remote License Update or Upgrade is divided into four steps 1 2 Generate a license request file and send it to sales odeon dk You will receive a mail with the license file when your update or upgrade has been generated Download the received license file to your dongle Update your ODEON software installation as needed from the www odeon dk homepage For more info on dongle update
76. 00 0 14000 0 20000 0 06000 0 08000 0 20000 0 10000 0 10000 123 03 9 x z Current s 06 urre lt Low Limit 3004 Wooden foor on joists Ingerslev 0 15000 0 15000 0 12000 0 10000 0 07000 0 06000 0 07000 0 07000 187 57 5 05 ge 1949 v High Limit 10003 Doudle gazing 2 3 mm glass 10 0 10000 0 10000 0 07000 0 05000 0 03000 0 02000 0 02000 0 02009 46 11 5 2 04 mm gap Kristensen 1954 2 4 Auto scale EE 40 adsorbent 0 40000 0 40000 0 40000 0 40000 0 40000 0 40000 9 49900 0 40000 1 53 En 03 3068 Plywood paneling 1 om thick 0 278000 0 28000 0 22000 0 17000 0 09000 0 0000 0 11000 0 11000 236 60 5 z z com 08 63 125 250 500 2000 8000 nanna aaseae PP nannan nanana a annaa p ws Frequency fiz Error as a function of Generation 63 Hz 125 Hz 250 Hz 500 Hz 1000 Hz 2000 Hz 4000 Hz 8000 Hz Average Error in JND s H EN u m B 0 Generation Individual Receiver independent ray tracing Tracing rays Job Rernaining time Live display of error Absorption coefficients The following main areas are available in the interface List of Jobs Jobs that have been assigned a task in the job list are listed here For this example 12 jobs have been used in the job list while two of those concern Multi point response calculations Since only Multi point response are used in the optimization procedure they are automatically activated Keep the fist 2 jobs activated List of parameters These paramete
77. 0deon plugin google sketchup The download includes installation instructions and last minute instructions Once you have installed SU and the plug in SU2Odeon you can start creating models for ODEON using SU Just save the SU model in the ODEON room folder and press the ODEON icon in SU menu bar Then SU2ZODEON will create a par file which is saved in the same folder and can be opened directly in ODEON Warning Note that for large models the SU application may emit the message SketchUp is not responding while it is in fact still working on creating the par file normally it only takes a few seconds but in some extreme cases it may take hours The models created in SketchUp and exported by SU2ZODEON are inherently compatible with the requirements set by ODEON e Plane surfaces e No repeated points e No surfaces without area e Multi loop surfaces are converted into single loop surfaces e Layer support e Automatic explosion of blocks and components e Unique surface numbers Extra lines appearing in the ODEON model When a surface is created in the middle of another surface eg a window on a wall SketchUp automatically cuts this surface from the other surface so that no overlap occurs An extra line will appear when the model is imported into ODEON see figure below An extra line by ODEON indicates a healthy surface that has been cut to accommodate the window in the middle Odeon 1985 2013 Licensed to Odeon Restri rs
78. 1 45 3 2 68 43 85 1 45 42 40 40 98 1 45 0 34 5 2 30 13 2 13 000 ElevSurf 1 1 22 2 7 walls Surf 200 Floor 1 gt 22 Surf 201 Ceiling 24 gt 24 22 1 HHH Elevation surface 2 Use the ElevSurf2 statement to define a series of vertical surfaces from a series of perimeter points plus an elevation height The perimeter points are typically defined using the MPt statement The syntax of ElevSurf2 is ElevSurf2 lt FirstSurfaceNumber gt lt FirstPointNumber gt lt SectionsInElevSurf gt lt Height gt lt T B N gt lt Optional name gt The ElevSurf2 only differs from ElevSurf in that a top and bottom surface may be specified the T B N option lt FirstSurfaceNumber gt A unique number from 1 to 2 147 483647 for identification of the first surface in the ElevSurf2 surface Using the same number but with negative sign define the surfaces and their counter parts mirrored in the XZ plane Y 0 lt FirstPointNumber gt First point number in the floor perimeter The floor points are typically MPt statement See example below lt SectionsInElevSurf gt The number of surfaces to be created by the ElevSurf2 statement If creating a cylinder a number between 16 and 24 is suggested if it s a column only use six to eight surfaces lt Height gt Height is oriented in the Z direction lt T B N gt The T B N parameter specifies whether the ElevSurf2 should have a top and or a bottom
79. 2 1 ISO 3382 1 2009 As reported in the standard the source should be able to produce a sound pressure level sufficient to provide decay curves with the required minimum dynamic range without contamination by background noise Some loudspeakers come with a built in amplifier active loudspeakers They can be very useful as no external amplification is needed When the source has no internal amplification passive source an external amplifier is needed External amplification is also needed for the microphone Many times the audio interface sound card offers sufficient amplification so it is worth checking the capabilities of your audio interface before investing in an expensive microphone amplifier 11 3 Measurements and Processing Important settings Before starting a measurement make sure that your equipment and settings fulfils the following prerequisites SNR of sound card and microphone The sound equipment you use for the measurements sound card microphone and amplifier should provide a decent S N otherwise you will need long measurement times in order to suppress background noise caused by the equipment In principle the longer is the sweep signal the higher is the suppression of the background noise For every doubling of the sweep length a suppression of background noise by 3dB is achieved 106 Analog filters of sound cards Some cheap sound cards may have very poor realisation filter which should cut off high freque
80. 20 High Non linearity in the decay When the energy decay in not linear calculation of reverberation time may be questionable as by definition reverberation time is calculated for an exponential energy decay translated to a linear decay in a dB scale Not only reverberation time but a group of many other parameters require an exponential energy decay as described in the ISO standard 3382 ISO 3382 1 2009 The XI parameter is used to quantify the degree of non linearity in an impulse response Values greater than 10 o indicate a questionable nonlinearity in the decay This may indicate that there is something wrong with the captured impulse response e g Impulsive noise during the measurement or indeed that the behaviour of the room is nonlinear As an example assume a rectangular room with uniform low absorption on all surfaces except from the ceiling which is highly absorptive The resulting decay curve experiences a knee point that changes the slope and leads to a double curved decay with a high XI value In the following case ODEON has placed the following warning message in the title bar WARNING max XI 158 o at 8000 Hz Similar messages are displayed for other frequencies If you are suspicious about the quality of measurement and if you believe the high value of XI is due to the recording repeat the measurement Check if the errors like high XI are reproducible In addition check if there is any sign of harmonic disto
81. 66 6 3 l ick FS UMAC erorii neneeese EEE ENEE E EE EESE EE EEE EEE AERE 66 6 4 Calculation of Response from Sources to Receivers seeseeseseesereererrererrsrrererrerrersresresrerees 67 6 5 The Reflection Based Scattering coefficient eeeesseseeseesrreresressersersresresrersersresresrerseesresresreses 72 6 6 CS Lam Dett csecsen E sadeuscapietcadeaouciesnpiaaedetesoleacneer 77 6 7 Diffraction over screens and round objects Screen diffraction 78 6 8 Radiated rays romi a SOU CCC sco ccese secssesacoeqeatensusssacenqutvaneneteeetueeionaengionanceeiacavenetab coeeennaeeneeeeeoe 79 6 9 Processing reflection data for auralisation in Single Point Response 80 6 10 Calculation method for Reflector Coverage sseseesererrereereerersrrsrrsresrersersresresrerrersresresresees 81 7 Calculated Room Acoustical parametels 0 00000 cesses 82 8 Calculation Parameters Room Setup and Define Grid 88 9 Achieving good results oo ccccccssssesssssessssesssssssseessssesssseessseesssessseessseeesse 92 9 1 PC OEEO dace cae acsearle asses E aetna ev ceadannoa sade ance does tea tl 93 10 Line array Sources oles cssssssessssnsesssssesssssssssssssesssssessssssssssssesssssesessseesessieseasseeten 97 10 1 a Ee E E E A E EE 97 10 2 apaa E Ay E E E sesame anyon euaanealuneaveraaasuete 99 10 3 PETA WAU NC VC oars te deca acne E A tie A T E E ss 101 10 4 Combining delay and level adjustments sacssncssceannssessysnsiaedecensetiisna
82. 8_ASCII_input_file_samples directory created at the installation of ODEON When the complete directivity characteristics are known FULL The first non comment line of the file should start with the word FULL In the full case there are 36 lines of data for each frequency The first 36 lines are for 63 Hz the next 36 lines for 125 Hz and so on As a minimum there must be 1 36 8 lines in a full input file e 1s line is vertical upper plot 0 12 o clock plot when looking from the front of the source towards it e g at a loudspeaker membrane e 10 line is horizontal left plot 90 9 o clock plot e 19 line is lower vertical plot 180 6 o clock plot e 28t line is right horizontal plot 270 3 o clock plot An example FullOmnidat on the full input format can be found in the DirFiles So8_ASCII_input_file_samples directory created at the installation of ODEON When the directivity pattern is rotationally symmetric SYMMETRIC The first non comment line of the file should start with the word SYMMETRIC In the symmetric case there is one line of data for each frequency As a minimum there must be 1 8 lines in a symmetric input file The symmetric sources could be a Trumpet or the Omni directional source An example Symmetric_Omni dat on the symmetric input 47 format can be found in the DirFiles So8_ASCII_input_file_samples directory created at the installation of ODEON Creating a new directivity pattern using a te
83. Auralisation both in measurements and simulations is played back though the device chosen in Windows playback devices panel Source power spectrum ODEON comes with two types of source power spectra that can be used with the calibration of the measuring system A flat frequency spectrum source called G ISO 3382 1 and a speech spectrum source called Speech ISO 3382 3 The source type G ISO 3382 1 should be chosen for almost all room acoustic cases covered in the ISO 3382 1 auditoria concert halls etc On the other hand source type Speech ISO 3382 3 should be chosen if the measurement is or has been 112 carried out in an open plan office ISO 3382 3 standard Apart from the default settings the user can define a custom source spectrum Background noise spectrum This specifies the absolute RMS dB value of the background noise in the room per octave band The values is obtained from the Room setup if a room has been assigned otherwise these values are used The level of the background noise is used in the calculation of the Speech Transmission Index and is important for the ISO 3382 3 results presented in the multi point response of ODEON Decay curve settings Here you can adjust the resolution of the impulse response as it appears at the Decay curves display see Load Impulse Response section When no room is assigned to ODEON the default setting for the resolution is 3 0 ms With a room assigned the value is taken directly from
84. Average none 0 00 99 00 10 none Average none 0 00 99 00 overloaded by 62 36 dB Setting the recording level down to 63 dB will lead to an output level just below 0 dB This process is very practical for single files as the absolute level of the playback can be further adjusted by the windows volume control However when mixing different convolutions you probably need to preserve their individual levels For example if you make an auralisation of an orchestra where all instruments have been recorded separately their individual dynamics levels should remain unchanged To avoid clipping reduce the Overall Recoding Level in the Auralisation setup Remember When the output level is greater than O dB clipping occurs In this case you have to reduce the input level Rcd lev column or the Overall Recording Level in the Auralisation Setup a Play wave file through headphones Once the calculations have been carried out click the Play wave result button and listen to the result through headphones If you have selected the Signal Sub Path or the Signal file column in the Convolve BRIR and Signal file table the anechoic input file is played If any other column in this table is selected the convolved result file is played Remember The play wave function is sensitive to the cell chosen in the table When the input signal cell is selected the play functions reproduces the input signal without convolution You have to selec
85. Curve 4 NAMAN AK MRotateZ 90 Cylinder 1000 N W 2 180 H N 10 Storres PListA for later use with floor 11 PList0 PListA 0 12 MTranslate 0 0 H 13 Dome2 2000 N W HCurve 180 Halfdome 14 MReset 15 MRotateZ 90 185 16 MRotateY 90 17 MTranslate 00 H 18 Cylinder2 3000 ONVert W HCurve L n Cylindric ceiling 19 MReset 20 Ptl 0W 20 21 Pt2 LW 20 22 Pt30W 2H 23 Pt4 LW 2 H 24 Surf 1 Floor 25 2 PList0 2 26 done with PList0 its a good habit to Reset it 27 ResetPList0 28 Surf 2 Side Walls 29 1243 30 Surf 3 Back Wall 2 PListB 2 31 F Defining surfaces with concave edges Most surfaces in the geometries used with ODEON will probably have convex edges rectangles cylindrical surfaces etc however in ODEON it is possible to define surfaces with cavities even surfaces with holes Such surfaces are defined just like any other surface by creating a list of corners where the listing is obtained by travelling around the surface s edge in either direction Below are two examples one with a donut shaped balcony floor and another with a cylindrical window opening in a ceiling In the donut example two rings of corners are created using the CountPt statement notice that the point 100 is equal to point 112 and point 200 is equal to point 212 The donut surface is created simply by connecting the inner and outer ring of points into one surface It doesn t matter whether one of t
86. Data can also be imported from Insul Sound Insulation Prediction software Insul simply by copying the spectrum for Bastian format and pasting into ODEON EJ Microsoft Excel Lydisolation eks xls Ioj xj BS File Edit Yiew Insert Format ool Data Window Help Typ on for helb X A2 fe Light double wall A BEC Spe SEa 6s SH eee Ee ae tS 1 Sound Transmission Loss 5 53 80 100 125 160 200 250 315 400 500 630 83 2 iLiaht double wall T umy AA OA AETS ae 2 AT ES o Er a gt NN Sheet1 Sheets Z Sheets ca Ready Suny 825 Z Copying transmission data of the reduction index from MsExcel A wall with different material on either side and walls modeled with thickness Sometimes a transmission wall may be composed by two individual surfaces two separate surfaces in the MaterialsList as shown in the illustration from the 3bBilliard display below It is possible for ODEON to link together such a pair of surfaces if they are almost parallel allowing transmission through walls with different absorption properties on either side The reduction index itself takes into consideration the wall thickness So for a wall persisting of two parallel surfaces the reduction Index should only be used once To accomplish this it is important that e Transmission type is assigned to both surfaces in the MaterialsList e Same set of reduction indexes are selected on either surface in the Transmission dialog e Double sided w
87. Denmark DTU will be used as an example A wire frame is shown in the following figure Load the room Open the room Auditorium 21 at DTU par that is pre stored in the ODEON s rooms folder eg C Odeon13Combined Rooms Inspect the room in the 3DView se and have a look at the Source Receiver list fl Two sources and five receivers have been set giving 10 combinations Open the Material list EA There are 11 different materials assigned to surfaces The back wall which is some kind of resonator panel with unknown resonance frequency and ceiling are set to 50 absorption for all frequencies In the real room it was not possible to inspect the ceiling material so an initial estimate is 2 x 13 mm gypsum board with mineral wool back Material list P Surface List Room material library auditorium21 at DTUSreceiversGkMaterials28may2014_ 7parOpSrnized_ 16June Lis Humber Material Scatter Transp Type Surface name Layer ania g amber Spaufmtisn 46 x04 0 050 0 000 Noma STAIRS a 4407 0 300 0 000 Normal Auderce tables 48 m27 0 320 0 000 Normal Audence tables 49 xO 0 080 0 000 Normal STAIRS so m27 0 200 0 000 Normal Audernce tables 51 D 0 080 0 000 Normal TABLES 52 1207 0 050 0 000 Normal Aa 5 320 0 200 0 000 Normal ALL 54 ww 0 050 0 000 Normal TAS 4 yi d 21 first guess SO 55 140 0 300 0 000 Noma Audience tables OOo opaa 6 3 0 050 0 000 Nomai STAIR WN heen 57 14307 0x0 0000 Noma Audence tables E Hoa se n07 0 900 0000 M
88. Drection towards main axs X none Remember In the Measured versus Simulated tab sheet you can change receiver by hitting the R key frequencies by using the Up Down arrows and parameters by using the Left Right arrows na Aulti point sponse parameters job 1 Simulated mode 3D Sources and Receivers Energy parameter curves 1 Energy parameter curves 2 Statistics Energy parameters Measured versus Simulated D 50 at 1000 Hz Receiver 1 D 50 D 50 N La wt wo p lp a O oO o ad N D S N lp oO oO O O N wt fee Receiver Frequency Hertz Ko Multi point response parameters job 1 Simulated mode 3D Sources and Receivers Energy parameter curves 1 Energy parameter curves 2 Statistics Energy parameters Measured versus Simulated T 30 at 1000 Hz Receiver 1 Simulated X Measured T 30 s T 30 s N m T w ise Ww oS Oo oO O oc w oc a N D N le oO i N E ioe Receiver Frequency Hertz Tip Double click on the graphs to get a table displaying the deviation of measurements from simulations in JNDs Just Noticeable Differences 132 Choose Room Acoustic Parameters Before starting the Genetic Material Optimizer you may want to limit the optimization process to several selected parameters not the default bundle To do so open the Room Acoustic Parameter list ee Initia
89. HRTF s which was measured on a live subject with blocked ear canal by the CIPIC Interface Laboratory Algazi 2001 along with our matching ee hph headphone filters The later contains the well known Kemar HRTF s which includes ear canals along with our matching ec hph filters One of the directories can be selected in the Options gt Program setup gt Auralisation setup dialog once this is done matching HRTF s and headphone filters can be selected in the same dialog as well as in the Auralisation setup specific to the individual room If installing new HRTF s or installing some of the additional CIPIC data which can be downloaded from http www odeon dk HRTF headphone filters are acceptable in the wav format In that case the filters should contain the impulse response s of the headphone as measured on a dummy head of the same type as the one selected in the HRTF drop down menu i e with or without ear canal or a corresponding ear coupler The filter may be one or 58 two channels two channels are desirable if compensating a specific headphone that particular headphone rather than that particular model of headphone A measuring program such as DIRAC Dirac may be suitable for the measurement of the impulse response of a headphone The creation of the inverse filter to be used is taken care of by ODEON Adjusting levels Sound Pressure Level is one of the most important room acoustical parameters so it is important that levels
90. L4 No description Line source at x y z 1 00 1 00 1 00 Length 5 0 m g W P2 S02 Point source at x y 2 3 05 1 69 1 60 El MS No description MSurf using surface 10 12 14 16 18 e F P3 No description Point source at x y z 9 00 1 00 2 20 ed e Receiver pointing towards source C a gt 1 No Gescnption Pi S01 Point source at x y z og Eg none amp 2 No Gescrption P2 S02 Point source at x y z X none a 3 No Gescnption Drection towards main axs X 1 R01 x y z 11 30 3 73 2 82 is 4 No description Draction towards man axs X 2 R02 xyz 6 78 2 84 1 47 Re 5 No description Drection towards main axs X 3 R03 x 2 5 92 2 51 1 20 amp 6 No description D rection towards main axs X 4 R04 x y 2 13 08 3 55 3 36 7 No descriotion Direction towards main axs X 5 R05 x y 2 10 30 0 15 2 55 Py No Gescnption Drrection towards main axs X 1 R01 x 2 11 30 3 73 2 82 9 No description Drection towards man axs X 2R02 x v 2 6 78 2 84 1 47 30 No descrption Draction towards man axs X 3 R03 x y z 5 92 2 51 1 20 11 No description Drection towards man axs X 4 R04 x y 2 13 08 3 55 3 36 12 No description Drection towards ran axs X 5 R0S x y 2 10 30 0 15 2 55 13 No description Direction towards main axis X none 14 No description Direction towards main axs X none 15 No Gescnption Drection towards main axs X none 161o Gescnption
91. Length TanD 30 Defining and reassigning variables The definition of variables must follow the syntax Var lt Name gt lt OptionalValue gt Example 1 defining the variable FloorLevel Var FloorLevel Example 2 defining the variable FloorLevel and assigning the initial value 0 Var FloorLevel 0 Example 3 reassigning a variable adding 1 metre to the FloorLevel 159 FloorLevel FloorLevel 1 Remark The predefined variable NumbOffSet may be used like any other variable but has a special meaning because it offsets point and surface numbering This variable is useful if copying a part of a geometry from another geometry file it is also useful in connection with the for end statements Auto can also be assigned to NumbOffSet in doing so ODEON will automatically increment the value of NumbOffSet to be greater than any point and surface number previously defined This has the advantage that repeated point and surface numbers can easily be avoided without having to keep track on the numbers used the drawback is that slight changes in the geometry file may change numbers on many subsequent surfaces ruining the relationship between surface numbers and the material assigned to that surface inside the ODEON program The Auto option is very useful in combination with loop constructions see description of the for end constructs later on Typing PtAbsRef after the value assigned to NumbOffSet forces absolute number references for points while u
92. N on the other hand is the READER ODEON will normally be the reader but being able to export array data itis a WRITER as well When developing export facilities for array loudspeaker data you may need to inspect or manually edit the files XML files can be browsed and edited in simple programs like Notepad the ODEONEdit editor will however perform syntax highlighting The ODEONEdit editor is installed with ODEON 10 or later including the free demo version Decimal point can be either or in the XML files ODEON will convert the XML files upon import to conform to the regional settings of the PC on which it s running The format is case sensitive therefore it is important that Attribute and Node names are spelled with upper and lower case as defined use one of the examples installed with ODEON and cut and paste directly from there to your code Encoding and formatting of XML documents Encoding should be set to UTF 8 lt xml version 1 0 encoding UTF 8 gt or in a coding environment XMLDocument Encoding UTF 8 Most parsers should be able to read a number of other encodings though To allow reading the XML document by eye consider to set the XMLDocument Options doNodeAutoIndent if that option is available in your programming environment XML document content Many of the data given in the XML file is optional if not given in the XML file ODEON will set a default value The nodes and Attributes of
93. N when loading the geometry and it has no effect on the geometry If Debug sOn is set to FALSE then debug lines are ignored Contents in the Debug strings which can not be evaluated are displayed in quotes Example When loading the following par file into ODEON HHH DeBuglIsOn TRUE debug option turned on if DebugIsOn is set to false then Debug lines are ignored const L 6 Debug L debug a single constant const W 4 Debug const W 4 const H 2 7 Debug L W H Debug values of L W and H Box 1 L W H TB Walls floor and ceiling Debug Box 1 L W H TB Walls floor and ceiling Debug a complete line HHH 180 ODEON will create this Debug window as a response O x Debug window DebugleUn is TRUE at places in par file thus user defined Debug messages may be displayed in this window In order to prepare the geometry for calculations and skip debug messages the DebugleUn flag must be FALSE everwhere in the par file Line amp L 6 We4 H 2 O000000000000032 rn 11 Bos 1 1 L 6 W 4 H 2 0000000000000032 TE walls floor and ceiling Creating a new Par file time saving hints The golden rule when creating a Par file to model a room is to think carefully before you start typing For very simple rooms it is not too difficult to keep track of things but for realistically complex rooms a systematic approach is desirable You will typically have a set of drawings which have to be use
94. ODEON The demo version of ODEON or a full version of ODEON Auditorium or Combined can be used for this purpose If you are reading this you have probably already installed it if this is not the case you may download the demo version from www Odeon dk To model and view the array in ODEON you only need to use a very limited number of features there is no need knowing all features in the ODEON software In a full version you may import your array into any room In the demo version you should open the ArraySpeakerTestRoom Par room located in the AppendixD folder e g C ODEON11Combined Rooms AppendixD ArraySpeakerTestRoom Par The ArraySpeakerTestRoom par room is a large box measuring 100x100x20 metres by default all surfaces have been assigned 100 absorbing material as its not intended for room acoustics predictions e To load this room use the File gt Open Room menu entry e Once loaded open the Source receiver list shortcut Shift Ctrl S e Inthe Source receiver list define an array shortcut A 149 e Inthe Array source editor import a XML array file shortcut Ctrl O That s it Hopefully the array imports without problems otherwise load it into the ODEONEdit editor to study the XML code When the file is loaded you have the chance to study the array in various tabs of the Array source editor Viewing and editing XML documents If your program writes data to be imported into ODEON then we refer to it as WRITER ODEO
95. ON has now been re built offering easier construction of directivity balloons Fe Toobar Options Tools Window Help SSBS DRL OG CB AHKSSHUEt Ht SH RVG LS 63 Hz Fixed 125 Hz Fixed 250 Hz Faxed 500 Hz Fixed 1000 Hz Fixed 2000 Hz Fixed 4000 Hz Fixed 8000 Hz Fixed oo i F45 0 25 0 15 0 15 0 dB at 1 metre Azimuth lt j r 2 260 J om J eo 2 re 2 10 ial a 4 Jia oo we z 20 ia a a a a ow erg I eS 30 id a a a a aco D 0 e 40 ia a a ow 1 g 50 a a a aj ow C i 60 a e a Aaji a ow O Thy 70 a a a Ta a oc gt z so u a a a owo O d Beez cl m ia besa 10 A i 70 100 m a a Ta a ow O A 110 a a a a a ow ar 1 120 al a al a owo C y 1 130 a a a o a 000 D j 140 w Saj a 0 00 i 150 id p Tia owo 120 160 FE 000 oe a eto Uevation 63Hz Fixed 125Hz Fixed 250Mz Fired SOO Mz Fixed 1000tz Foced 2000Htz fixed 4000HMz Fixed 8000 Hz Fhoed a jevati ET a a D 0 00 J f 70 100 n a ow x E a k C ace a a 0 00 a gt a 0 00 ry i ry aco ry _ ooo a x owo a P 0 00 100 a m E oo 110 s i L oco 120 a ry m ooo 130 a a a oo 140 a i a aco 150 a M 2 00 160 a a i a oo 170 owo M ooo ii Minge plot Fip piots Coy Display optons Lett Fight Top gt Batom Front gt Back Revolve Let Lak Right Dest band Al bands Tl Bands to depla in d
96. ON is presented 6 2 Global decay methods ODEON features two methods for calculating the Global decay of rooms e Quick Estimate which is available from the Materials List is the fastest method allowing quick evaluation of the effect of changes to materials This method should be considered only as a tool for preliminary results e Global Estimate is the most precise of the two global methods allowing high quality results 6 3 EA Quick Estimate This method estimates a mean absorption coefficient which is inserted in the Sabine Eyring and Arau Puchades formulas to give an estimate of the reverberation time Instead of simply taking the areas of the surfaces and multiplying by the corresponding absorption coefficients to obtain the total absorption in the room ODEON also sends out particles from the source assuming diffuse conditions thus reflecting them in random directions keeping a count on how many times they hit each surface Surfaces that are hit very often then carry greater weight in the overall mean absorption coefficient of the room Surfaces which are not detected at all in the ray tracing process are left out of all calculations and surfaces which are hit on both sides are included twice in the calculation As a result the estimated reverberation time corresponds to the sub volume in which the selected source is located Note however that if a part of the area of a surface which is present in the sub volume is located outsid
97. Pt 14 6 2 2 7 Surf 1 floor 1 2 3 4 Surf 2 ceiling 11 12 13 14 Surf 3 end wall 1 2 12 11 Surf 4 end wall 1 2 12 11 Surf 5 side wall 1 4 14 11 Surf 6 side wall 2 3 13 12 HHH Below the box shaped room is modelled using constants for the definition of W L and H Some of the advantages of using parameters in modelling rooms are that it makes changes to a model much easier allowing reuse and often it will also improve the clarity of a model data Parametric sample BoxFromParameters par The box measures are Width 4 metres Length 6 metres Height 2 7 metres HHH const W 4 const L 6 const H 27 Pt 1 0 W 2 0 Pt 2 0 W 2 0 Pt 3 L W 2 0 Pt 4 L W 2 0 Pt Il 0 W 2 H Pt 12 0 W 2 H Pt 13 F W 2 H Pt 14 L W 2 H Surf 1 floor I gt 4 Surf 2 ceiling 182 11 gt 14 Surf 3 end wall 1 2 12 11 Surf 4 end wall 1 2 12 11 Surf 5 side wall 1 4 14 11 Surf 6 side wall 2 3 13 12 HHH Below the box shaped room is modelled using parameters and symmetric modelling syntax signs on point and surface numbers The symmetric modelling syntax means less typing and less typing errors Parametric sample BoxFromParameters UsingSymmetricModelling par HHH const W 4 const L 6 const H 2 7 Pt 1 0 W 2 0 Pt 2 L W 2 0 Pt Il 0 W 2 H Pt 12 L W 2 H Surf 1 floor 1 2 2 1 Surf 2 ceiling 11 12 12 11 Surf 3 end wall 1 11 11 1 Surf 4 end wall 2 12 Mirror Mirror works just as well defines point 12 and 2 Surf 5 sid
98. TI Jnd ref J Acoust Soc Am 2001 Apr 109 4 1474 82 STl_Femal 0 00 1 00 Speech Transmision Index specfically STl_Female Jnd ref J Acoust SocAm 2001 STI_Male 0 00 1 00 Speech Transmision Index specfically STI_Male Jnd ref J Acoust SocAm 2001 Apr 109 4 1474 82 RASTI 0 00 1 00 Speech Transmision Index specfically RASTI Jnd ref J Acoust SocAm 2001 Apr 109 4 1474 82 Wide band parameter formula 1 SPL_A 10 Logi0 Power 10 SPL_63 AW_63 0 1 Power 10 SPL_125 AW_125 0 1 Power 10 SPL_250 AW_250 0 1 Power 10 SPL_500 AW_500 0 1 a Power 10 SPL_1000 AW_1000 0 1 Power 10 SPL_2000 AW_ 2000 0 1 Power 10 SPL_4000 AW_ 4000 0 1 Power 10 SPL_8000 AW_ 8000 0 1 Select the active visible parameters B Room acoustic parameter list D ODEON FILES ODEON VERSIONS TASKS ODEON_13 Auditorium21 at DTU PcE GEEA Energy intervals Omni microphone Room acoustic frequency parameters E Name Start millisec Stop millisec Type Visible Decimal Y Origin Unit Manual grid Min Grid Max Grid Meas 7i E_Omni_Total 0 00 INF Reverberatio v 2 m s 0 00 2 50 pH it E_Omni_Direct 0 00 0 10 Reverberatio v 2 W S 0 00 2 50 W fu E_Omni7 0
99. The options are T B TB and N for none If Top or bottom may only be included if all of the points in the floor in the elevation surface are in the same plane 165 Example on use of the MPt and ElevSurf2 statements First the perimeter points point 1 to 23 at the floor level of an office environment is described using the MPt statement Then the elevation surface is created from these points creating the perimeter walls of the office with a constant height of 2 7 metres In this example the X coordinates are made in absolute values whereas the Y coordinates in most cases are in or de creased using the or options To create a closed ElevSurf2 that is the first wall joins the last wall first and last point in the series of points handled to the ElevSurf2 must be identical in this example point 1 and point 23 are identical If an elevation surface has 22 surfaces then 23 points must be made available to the ElevSurf2 as in this example HHH MPt 1 23 000 7 48 2 2 38 8 35 26 10 76 4 64 64 78 2 68 51 52 49 62 2 02 48 22 1 1 45 3 2 68 43 85 1 45 42 40 40 98 1 45 0 34 5 2 1 30 13 2 J3 000 ElevSurf2 1 1 22 2 7 TB walls HHH Defining a number of surfaces using the CountSurf statement The CountSurf is mostly here for backwards compatibility In most cases it will be easier to use the Surf statement along with a for
100. a Remember Structures with high roughness should be modelled as planar surfaces The lack of detail can be compensated by choosing an appropriate scattering coefficient according to the depth of the structure 2D diffuser 1D diffuser N wv QO c lt b oy co Sen a c 2 od O Q On c lt ro p Q N 10 20 50 100 200 500 1000 Depth of structure mm Scattering coefficient according to depth of structure Assign scattering to a surface according to the depth of s Scattering due to diffraction In order to estimate scattering due to diffraction reflector theory is applied The main theory is presented in Rindel 1986 Rindel 1992 The goal in these papers was to estimate the specular contribution of a reflector with a limited area given the basic dimensions of the surface angle of incidence incident and reflected path lengths Given the fraction of the energy which is reflected specularily we can however also describe the fraction sa which has been scattered due to diffraction A short summary of the method is as follows For a panel with the dimensions l w above the upper limiting frequency fw defined by the short dimension of the panel the frequency response can be simplified to be flat i e that of an infinitely large panel below fw the response will fall off with by 3 dB per octave Below the second limiting frequency fi an additional 3 dB per octave is ad
101. a row e g hand clap recordings in a sequence a sufficiently short Noise floor window length 111 is required in order for ODEON to successfully discriminate the different impulse responses In cases where magnetic feedback introduces strong spikes at the beginning of the impulse response this measurement error happens if microphone and loudspeaker cables lie parallel and close to each other for some distance this setting can help the user get rid of it Sweep model Two types of sweep signals can be used for a measurement in ODEON Linear sweep that has frequency energy spread equivalent to white noise and Exponential sweep which has frequency energy spread equivalent to pink noise In other words the Exponential sweep provides longer playback time for low frequencies thus more energy at this range while the Linear sweep provide longer playback time at mid and high frequencies For most room acoustic measurements the Exponential sweep is preferred against the Linear sweep On the other hand Linear sweep may be preferable in measurements of sound transmission between rooms because partitions often provide attenuation at high frequencies Receiver model Two types of receivers microphones can be specified in the measuring system e 1 Channel Omni e 2 Channels Omni Figure8 The omni directional receiver is sufficient in order to calculate parameters such as EDT T 30 SPL C 50 C 80 etc Remember 1 In order for par
102. a very realistic and reliable assumption without affecting the computation cost How to model the podium on stage Same guideline as for the audience area goes here Rather than modelling each step of the podium on stage the podium can be simplified into a few sloped surfaces Should furniture such as tables chairs and shelves be included in a model of an office If a table plate is close to a source or receiver point then it is likely to produce a strong specular reflection at the receiver so if this is the case then it should be included Furniture such as shelves and screens in large office environments which subdivides the room breaking up long reflection paths and introducing extra absorption and scattering should neither be omitted Furniture at more distant locations in the room which does not produce any strong early reflections to the receiver can be greatly simplified or even omitted from the model as long as the extra absorption and scattering produced by that furniture is somehow included on other surfaces in the same regions of the room Orientation of surfaces does tilt of a surface have any significance on room acoustics Small changes to the orientation of surfaces can indeed cause dramatic changes Making dominant surfaces slightly off angle can cause extra scattering in the room almost as if extra scattering had been assigned to the surfaces in the room A classic example on this is the box shaped room where
103. able plate with circular ends is modelled a First a circle is created b Then half of the circle is deleted c Finally the surfaces is mirrored in a horizontal mirror at y 1 Creating a circular surface To create a circular surface first draw a line a surface with two points in order to specify centre and radius of the circle then use the Ctrl O shortcut to activate the circle tool and accept to create a circle from 12 points If the circle tool is clicked when the selected surface contains less than two or more than three points then a help text is displayed this text will also explain about ellipses Make a circle half a circle Delete the 5 upper points in the circle in order to reduce the circle to half a circle At this point you should have created the half circle in middle of Figure above fi Mirroring the surface In order to create the complete table the mirror functionality can be used select the horizontal mirror and specify the coordinate of the mirror line in the Figures the coordinate was 1 00 Select the first of the two points to be connected across the mirror line and finally use the Mirror shortcut Ctrl M to create the full table The position of the mirror is easily changed if holding down the Shift key while pressing Right mouse and moving it if performing a very significant move in the horizontal or vertical direction this will toggle between a horizontal and a vertical mirror line blue dashed line Mirror m
104. al with an internet link is chosen the material homepage will come up by a click on the web button from within the Materials List shortcut W The absorption coefficient a for each octave band should have a value between above 0 00 and 1 00 If no data is available for upper or lower bands do not write 0 00 rather use the value of the neighbouring band e g use the 4 kHz value for the 8 kHz band 4 2 Surface List Left side of Material List window The surface list lists the material specifications assigned to the surfaces starting from the left to the right Surface number The unique number assigned to the surface in the geometry file Material number The number of the material assigned to the surface from the material library This number and material corresponds to the number listed in the material display except when e The material has been edited in the material library after the material was assigned to the surface e g its absorption coefficients have been changed e The materials were assigned on another computer where another material library was available with different definitions of the material having this number Remember Once a surface has been assigned a material this material stays the same for that surface even though the material has been changed in the material library the Material Li8 file thus calculated results stay in consistency with the materials assigned to the room even if the room is moved
105. all at a power plant Measured SPL A is included in that example Apart from the common point sources surface sources have been used as well that can model the radiation of sound from an entire surface This is due to the assumption that an engine consists of a group of surfaces that radiate sound as they vibrate The example Transmission rooms par found also in the main Rooms folder includes some walls with transmission properties and it serves as a good demonstration of insulation between rooms You can read more about transmission modelling with ODEON in Appendix B The room Oil rig par is an example of an offshore oil platform where loudspeakers have been installed as part of an alarm system Noisy sources have been taken mene TTT TT LT the oe oe me m ei oe p o a Be eee a Sh S HERRERSUERRRSRGRRRROGOROOO into considerations and they are modelled as surface sources This room has also been abalysed in the application note Calculation of Speech Transmission Index in Rooms which can be downloaded from the application notes page http www odeon dk application notes 22 Tip Some of these rooms have been set up for very accurate calculations that can take longer time If you wish to carry out faster calculations you may enter the Room setup ah and select the Engineering setting This will provide results approximately 2 times faster without much loss in quality in results Public Ad
106. all check mark is selected for both surfaces in the Transmission dialog Most of the above can usually be accomplished if the Update double sided wall upon exit is checked when Transmission data for the first wall is edited It is recommended checking and rechecking 144 the data entered for transmission data before making calculations a check may involve using the 3DBilliard Or 3D Investigate Rays to ensure that walls do in fact transmit sound and that double sided walls has been set up correctly 0 2 4 6 8 10 12 14 metres 6 metres Oo Path lt m gt 8 50 Refl order colour 0 1 2 3 4 5 6 7 8 9 10 11 gt 12 Time lt ms gt 25 Odeon 1985 2008 Licensed to Odeon A S Dead balls O Figure B2 3DBilliard display illustrating Transmission through a double sided wall as can be seen ODEON understands correctly that balls should jump through the wall from one surface to another The principle of calculations is shown in Figure D2 Statistically 10 of the balls rays are transmitted and 90 are reflected However this is compensated for in the calculations so the energy losses in the two rooms are determined by the absorption coefficients and reduction indexes as used Odeon 1985 2014 Licensed to Odeon Restricted version research and teaching only The Transmission Rooms par Sample is installed with ODEON The Figure illustrates how transmission data are assigned to transmission walls wh
107. ameters like LF 80 and LFC 80 to be calculated option 2 including the figures microphone must be selected and both the omni and figure8 microphones must be connected to the sound card If option 2 is not selected then you should uncheck the Measured option for those parameters in the Define room acoustics parameters 2 In order to obtain meaningful values for a parameter such as SPL it is essential that meaningful values are entered for the Source Power Spectrum described below and that a system calibration has been performed if you do not perform any system calibration then you should uncheck the Measured option for SPL and other parameters that may rely on calibration in the Define room acoustics parameters 3 There are parameters such as Steady State Diffusivity Diffusivity ss which can be simulated in ODEON because of a large number of different types of microphone directivities are available for simulations however only the omni and figure8 microphones are currently possible options for real measurements Therefore you should uncheck the Measured option for any such parameters in the Define room acoustics parameters Audio devices All audio devices installed in your PC should appear in this menu You should choose which devices you prefer to use for recording Input device and which for playback Output device Remember The output device in the measurement setup affects only the sweep signal during the measurement
108. and 1 always When 2 parents reproduce and create a new child their chromosomes are combined using Crossover and then the child s chromosome may be mutated Elitist percent Describes the percentage of elite parents used every generation for evolution When this percentage is 0 it means no elite parents are chosen When this is 50 it means that half of the parents are selected as elite The best 50 are chosen 137 References ISO 8879 1986 Information processing Text and office systems Standard Generalized Markup Language SGML Anechoic Orchestral Music Recordings 1988 Denon PG 6006 Music for Archimedes 1992 CD B amp O 101 B amp O Denmark ISO 11690 1 1996 Acoustics Recommended practice for design of low noise workplaces containing machinery Part 1 Noise control strategies ISO 14257 2001 Acoustics Measurement and modelling of spatial sound distribution curves in workrooms for evaluation of their acoustical performance ISO 9613 1 2001 Acoustics Attenuation of sound during propagation outdoors Part 1 Calculation of the absorption of sound by the atmosphere IEC 60268 16 2003 Sound system equipment Part 16 Objective rating of speech intelligibility by speech transmission index Third edition ISO 3382 2 2008 Acoustics Measurement of room acoustic parameters Part 2 Reverberation time in ordinary rooms ISO 3382 1 2009 Acoustics Measurement of room acoustic param
109. anipulation Mouse operation Move mirror line toggle between vertical and horizontal mirror line Shift Right mouse button y Creating a mirrored copy Uses the current mirror line in order to create a copy of the currently selected surface Create scaled copy of surface Alt Ctrl C shortcut As above but a dialog appears allowing input of a scaling factor 38 Create copy of surface Will create a copy of the selected surface The copy is offset slightly from the original one to make it visible amp Rotating a surface To rotate a surface select the point in the surface around which the surface should be rotated then activate the Rotate dialog using the Ctrl R shortcut and enter the number of degrees to rotate the surface positive rotation angles are always counter clockwise CCW A surface can not be rotated around a point which is not included in the surface However this trick will do it insert the point of rotation into the surface e Select the surface e Bring it into editing state Esc or Insert shortcut e Add the rotation point it is not important where it is inserted in the sequence of points e Rotate the surface e Delete the rotation point from the surface Del shortcut I eon 4 ioo hoo Ie oo mmm dee ee ee ee 2 00 To rotate a surface around a point which is not included in the surface a Insert a rotation point in the surface b Use the Rotate surface shortcut Ctrl R to activate th
110. ary loss for any direct indirect consequential or incidental damages resulting from any defect in the software update version or its documentation or arising out of the use the results of use or inability to use the product even if ODEON has been advised of the possibility of such damage or claim LIMITED WARRANTY ODEON A S make no warranty or representation either expressed or implied including but not limited to implied warranties or merchantability and fitness for a particular purpose with respect to this product As a result this software is sold as is and you the licensee are assuming the entire risk as to its quality and performance No ODEON A S Distributor agent or employee is authorized to make any modification extension or addition that will increase the scope of this warranty UPDATE POLICY ODEON A S may create from time to time update versions of the ODEON software Odeon A S reserves the right to make changes to the software or its documentation without notice Major updates e g from Odeon 11 to Odeon 12 will be made available to registered users who have a valid Maintenance and Support Agreement or who purchase an upgrade Minor service updates e g from v 12 0 to 12 01 or 12 1 may be made available for download and may also be used by the licensee without a valid Maintenance and Support Agreement if the hardware key permits running the update USE OF ANECHOIC ORCHESTRA RECORDINGS DELIVERED WITH THE SOFTWARE
111. ase and cepstrum plots spectrum of the frequency response Yistortion Noise 25 30 dB at 500 Hz use longer sweep Measured response D Measured Impulse responses Auditorium2 March S2R5_sweep4000 wav Raw decay curve at 1000Hz Decay curves at 1000Hz T 30 1 87 seconds Decay curves all bands Energy parameters Parameter curves Frequency Response ita ma D Measured Impulse responses Auditorium21_8March S2R5_sweep4000 wav E e sai Response M Complete M Smoothed aaa wie dB 63 125 250 500 1000 2000 4000 8000 Frequency 14 floor and the onset time a group of manual operations are now possible for those who wish to have full control on the impulse response file Together with the existing crop function editing an impulse response file becomes an extremely time efficient task Very often a room model needs to have several openings in the geometry open windows open ceiling etc Moreover there are projects that deal with pure outdoor sound propagation such as an atrium courtyard a street canyon etc In all these cases a bounding box needs to be constructed by the user Now an easy to use command has been added in the ODEON editor which builds a Although ODEON offers superb automatic truncation of recorded impulse responses at the noise bounding box to automatically include the whole geometry Manual definition of Onset and Truncation times in a measured Impulse Response Fast Constructio
112. at which auralisation samples are presented are realistic If playing a simulation of voice at an unrealistic high level the speech intelligibility may be over rated it does not help that Clarity or Speech Transmission Index is satisfactory if the Sound Pressure Level is too low If play back levels are too high echo problems may be exaggerated because echoes that would be below audible threshold or at least at a very low level are made audible The levels presented in auralisation samples created by ODEON are influenced by e The HRTF s e Level in input signal file e g the RMS value or Leqa e Calculated Sound Pressure Level which is based on geometry sources receiver positions materials etc e Overall recording level in the Auralisation setup e Rcd Lev in the Auralisation display of the JobList if off line convolution is used e Mixer levels Mix Lev in the JobList if off line convolution is used e Gain in the Streaming convolution dialog if the real time convolution option is used e Output gain of the soundcard the volume setting e Sensitivity of the headphones e Coupling between headphone and the subjects ear Maximised play back levels for maximum dynamic range If you are only interested in the best sound quality in your auralisation files you may focus on getting an Output Level Out Lev in the auralisation display within the Job list below but as close to 0 dB as possible in the Convolve BRIR and Signal file tab
113. atement Cylinder2 is a cylinder shell of the calotte type Rather than specifying the radius and revolution angle Cylinder2 is specified in terms of the width and height Cylinder2 is typically used for cylindrical curved ceilings The syntax for Cylinder is Cylinder2 lt Number gt lt NumberOfSurfaces gt lt Width gt lt Height gt lt Length gt lt T B N gt lt optional name gt lt Width gt Y 176 Width is oriented in the X direction on the Figure lt Height gt Height of the cylinder shell is oriented in the Y direction on the Figure and may be positive concave shell as well as negative convex shell Height must be different from zero and less or equal to 2 Width lt Length gt Length of the cylinder shell is oriented in the Z direction the Figure If Length is negative the orientation is inverted Insertion point The insertion point of Cylinder2 is always foot point of the calotte floor bottom surface Connection points The foot points in Cylinder2 are stored in PlistA The top points in Cylinder2 are stored in PListB The example shown was generated with the following code HHH Const N 10 Const W 5 Const H 1 Const L 10 Cylinder2 1 N W H L TB Cylinder calotte HHH Hint The cylinder can be made elliptical using the MScale statement The Cone statement The Cone statement models a cone Typical use of the Cone statement is for modelling half cone or cone shaped ceilings The syntax for Cone is
114. ave to use the Offline Auralisation by pressing the toggle button on the right of the Joblist Audio files produced in the offline auralisation may be used on the WEB CD ROMs Power Point presentations etc Toggle button for offline auralisation W JobList Binaural mode Headphone Subject_021ReslOdeg_diffuse wav a 00 9 0 i yuon b 1 v Agora 0 41 05 Average none 0 00 99 0 e mee i No desorption 39 00 1 none 0 00 none Average of 0 00 none 0 00 Convolve Bin a Mi corwotved wave results into one wave fle LOnvolve a Kene e Wea Corwotutions in mix no 1 ERIE cnabied SAI SEPE STN RRS CRSA CRE TEE WEEENTEF SSUES ped lev Maxo M0 no Enabled WEEN Description tax out 7 No descrptin 99 00 Wo Mbc Lev Delay in sec Job deser Signal file 0 0 0 A Jt d e w i a a b The offline convolution display is divided into a left and a right part In the left display mono signals are convolved with Binaural Room Impulse Responses BRIR s which have been calculated as part of the Single Point responses this process may be compared to a binaural recording of a mono signal played through simulated source s in the room other types of impulse responses can be used for different playback techniques The right part of the auralisation display two tables is a mixer allowing convolved results to be combined into one wave file allowing multi channel simulations e g stereo setups singer ver
115. ave zero strength Whereas rays perpendicular to the surface carry the maximum of the energy Spherical radiation introduces uniform strength to the rays regardless of their direction Lambert radiation is preferable for ordinary surface sources such as noise emission from machinery Spherical radiation is recommended when several point sources are modeled as a surfaces source instead of each one individually e g many people distributed over a large canteen hall Spherical radiation is also a good option for modelling traffic noise which can be regarded as a dense array of vehicles acting as individual point sources Lambert left and spherical right types of radiation applicable to line and surfaces sources 6 9 Processing reflection data for auralisation in Single Point Response A typical point response calculation in ODEON includes some 100 000 to 1 000 000 reflections per source receiver The reflections are calculated in terms of time of arrival strength in 8 octave bands and angle of incidence azimuth and elevation The information on size of the reflecting surfaces and absorption coefficients are also available as a part of the calculation Binaural filters for headphone playback When the Auralisation setup Create binaural impulse response file option is turned on reflections are processed in order to create a binaural impulse response BRIR First of all it is determined whether a phase shift should be applied to the ref
116. ays end up as 0 dB the other bands shifted accordingly NON calibrated sources Electro acoustical sources machinery natural sources etc Press NO to preserve the sensitivity of an electro acoustical source or the absolute level of natural source e g a human voice When selecting the NON CALIBRATED source you are allowed to enter equalising electric losses zero for natural sources and a sensitivity at a selected frequency band zero for natural sources The addition of electrical sensitivity electrical input power and electrical loss values completes the data necessary to generate a source directivity file directly readable by ODEON 3 or later 48 Materials The next most important step after importing a room geometry and after assigning sources receivers is to set the materials you would like to test on the surfaces of the geometry This chapter covers material properties and the facilities available from within the Material List You can access the Material List either by pressing EE or by typing the SHIFT CTRL M shortcut The material list consists of a window containing two lists the surface list on the left and the material library on the right When selecting a surface in the surface list the surface is automatically highlighted in the corresponding 3D Materials window G ODEON 13 uM as ustrial Roots Ek File Toolbar Materials Options Tools Window Help SES OULSF OMaT CBP RASANT SY SB RRS Lod 30 Maternal I
117. ber of sections in the cylinder by simply changing the N constant Parametric sample a cylinder RevSurfCylinder Par HH const N 16 const R 15 const H 10 CountPt 100 N 1 R CosD PtCounter 360 N R SinD PtCounter 360 N 0 CountPt 200 N 1 R CosD PtCounter 360 N R SinD PtCounter 360 N H RevSurf 300 100 200 Sections cylinder walls Surf 100 Circular floor 100 gt 100 N 1 Surf 2 Circular ceiling 200 gt 200 N 1 HHH Modelling a box shaped room with columns in two dimensions using two level For End constructs When modelling geometries having more than one level of symmetry it is advantageous to use For End constructs This example shows how to model columns in two dimensions in a room using a two level For End construct Each column is created using 8 points and 4 surfaces thus the numbering used by points and surfaces is incremented by 8 each time a column is created This is done by incrementing the predefined variable NumbOffSet by eight for each column in order to make surface and point numbers unique The different positions of the points used for each column are obtained using MTranslate and MReset Parametric sample BoxColumnRoom Par HHH const L 10 const W 4 const H 3 const NumColX 4 184 const NumColY 3 const ColumnW 0 3 mTranslate 1 2 0 0 Box 1 lw h tb walls in the room modelling the columns for ColYCnt 1 NumColY MReset MTranslate L NumColX 1 w 2 ColYCnt W NumColY 1 0 for ColXCnt 1 NumC
118. between 5 and 35 dB obtained from the backward integrated decay curve e Early Decay Time EDT is obtained from the initial 10 dB drop of the backward integrated decay curve 82 Sound Pressure Level Cso D50 LF so STearly STiate and STtotal The energy of each reflection is added to the appropriate terms in the formulas for all the energy parameters according to its time and direction of arrival After the response calculation Cso Clarity Dso Definition Centre Time SPL Sound Pressure Level LFso Lateral Energy Fraction STearly STiate and STtotal are derived In the following formulae Ea is the sum of energy contributions between time a and time b after the direct sound time t is the end of the calculated response and t is for the reflection arriving at time t the angle between the incident direction and the axis passing through the two ears of a listener Only the few most common of the parameters that are available in ODEON are listed here in fact you may also look up the definitions of parameters from within the Room Acoustic Parameter list in ODEON where the formulas are listed and where you may even enter new formulas It is worth to notice that parameters that rely on the Ea term assumes that time a and b are well defined however this is not the case when there are no direct sight between source and receiver This does make results questionable in simulations as well as in real measurements when direct sight betw
119. body to be included twice at the recording and at the playback If one has measured or indeed simulated the BRIR s see below in a room it is possible to simulate a binaural recording BRIR Binaural Room Impulse Response The BRIR is the key to binaural room acoustic auralisation The BRIR is a set of impulse responses detected at the left and right entrance of the ear canals of a dummy head or indeed at blocked entrenches of the ear channels of a living person residing in a room when a sound source or some sound sources has emitted an impulse The BRIR should include all the necessary information on receiver position and orientations source s position s and orientations room 141 geometry surface materials and the listener s geometry described by the HRTFs Convolving the left channel of the BRIR and the right channel of the BRIR with a mono signal a binaural signal is created which when presented to the listener over headphones gives the impression of the three dimensional acoustics at a particular position in the room It is also possible to simulate the recording of the BRIR s which is what ODEON does HRTF s Head Related Transfer Functions In short terms the HRTF describes how an impulse arriving at a person dummy head is smeared out by diffraction phenomena from head and torso of the person While an incoming impulse is only 1 sample long this will result in an impulse response arriving at the right and
120. ce is completely symmetric as above then the symmetric points can also be specified using the Mirror word which should be the last component in the corner list Surf 1000 Symmetric surface 1 2 Mirror Note You should not try to define the point 2 in the geometry file it is automatically generated Symmetric double surface Symmetric double surfaces are pairs of surfaces symmetric around the XZ plane Y 0 e g a right and a left wall Surf 2 Right wall Left wall 12 22 23 13 will appear as two surfaces inside the ODEON program Thus you will have the following two surfaces inside the ODEON program 2 Right Left wall containing the symmetric points 12 22 23 13 and the surface 2 Right Left wall containing the points 12 22 23 13 as they are defined in the geometry file Note You may not define surface 2 if you are using the symmetric double surface 2 because ODEON automatically generates surface number 2 The Box statement The Box statement defines a Box with or without top and bottom The Box statement may typically be used for Box shaped rooms and columns A special case of the Box statement is when one of the dimensions Length Width or Height is zero in this case only one surface is created 174 The syntax of the Box statement is Box lt Number gt lt Length gt lt Width gt lt Height gt lt T B N gt lt optional name gt lt Number gt A unique number from 1 to 2 147 483647 for identificatio
121. ces that 67 radiates energy locally from the surfaces of the walls Responses from line and surface sources are carried out using the special ray tracing method Energy Reflection order TO The calculations carried out are divided into a two step process a receiver independent part and a receiver dependent part Receiver independent ray tracing The purpose of this process is to find virtual sources that radiate energy into the room Rather than thinking of reflections the idea is so to speak to substitute the room with a number of sources at various positions and delay in space The process is divided in to an early part based on the image source method for point sources and a late part based on a secondary source method ray radiosity Early part finding the image sources Image sources can be found by just taking all combinations of walls up to a given reflection order e g Source Wall1 Wal2 Wall Wall3 and so on For simple rooms with few surfaces this is easily accomplished however for rooms with thousands of surfaces this can lead to huge number of images sources which for geometrical reasons are not valid anyway Instead another approach is used in ODEON early rays up to the transition order specified in the room setup are used in order to detect possible image sources A number of early rays are generated ODEON suggest a number based on room dimensions transition order and surface sizes The rays are reflected
122. correlation coefficient is calculated as IACC max IACF for Ims lt 7 lt 1 ms STI Speech Transmission Index Speech Transmission Index known as STI is calculated according to IEC 60268 16 2003 The STI parameter takes into account the background noise which may be adjusted from the Room Setup For the STI parameter to be valid it is very important to adjust the background noise accordingly remember that background noise must be set in a relative level if relative source gains are used It should be mentioned that it is not stated in IEC 60268 16 2003 what kind of directivity the source in the STI measuring system should have so if using a source with directivity different from the one used in the real measurements in the simulations results may not be comparable The subjective scale of STI is given below Subjective scale STI value 0 00 0 30 Poor 0 30 0 45 0 45 0 60 0 60 0 75 0 75 1 00 In the point response and multi point response results five different speech transmission indexes are shown STI This is the most known form of STI derived for an average gender independent voice spectrum STI Male and STI Female These are STI values adapted for male and female speakers according to their speech spectrum Both the contribution of each Octave Band to the STI and the information overlap from one band to another is taken into account 85 STI Expected The expected STI value is
123. d as the basis for the ODEON model It pays to spend quite a long time working out how the room can be simplified to a manageable number of sensibly shaped plane surfaces sketching over the drawings These ideas will have to be modified when you start to work out the actual coordinates to ensure that the surfaces really are plane Here are some ideas that may help you to create correct surface files faster e Exploit symmetry If the room has an axis of symmetry place the coordinate axis on it Then use the sign convention for symmetric semi symmetric modelling e If there are vertical walls and or features which repeat vertically e g identical balconies use the CountPt CountSurf RevSurf statements or indeed For End constructs e Build the room gradually testing the Par file at each stage of growth by loading it into ODEON and have a look at the result e Use hybrid statements such as Box Cylinder etc Where it is difficult to get surfaces to meet properly without either warping or using lots of small surfaces to fill the gaps allow the surfaces to cut through each other a little This will usually ensure a watertight result and has only minor drawbacks These are i the apparent surface area will be a little too big affecting reverberation times estimated using Quick Estimate Reverberation and the room volume estimated by Global Estimate ii crossing surfaces can look odd and hinder clarity in the 3D displays Do not try
124. d be sufficient The scatter rays are traced from the reflection order of the images source up to the selected Transition order Select calculation methods Angular absorption Job calculations and Global Estimate When this option is on angle dependent reflection factors are taken into account The calculation method uses the diffuse field absorption coefficients as input but generate angle dependent ones taking into account angle of incidence and the size of the surface The method is based on the work in Rindel 1993 In some situations this option does not improve the results but on the other hand it does not seem to harm but in many situations it means a significant improvement Three options are available Disabled All materials and Soft materials only The All materials is the thorough method whereas the Soft materials only is a trade off between calculation time versus prediction quality The Soft materials only method is usually almost as fast as the Disabled option and yields a quality almost as good as the All Materials method which is somewhat slower Screen diffraction If this option is checked ODEON will try to calculate the sound diffracted around objects when the source is not visible from the receiver The diffraction method is the one suggested by Pierce 1974 This method includes one and two point diffraction allowing diffraction over or around single surface screens wedge shaped screen and two sided objects The Scre
125. dd HLA Miler Taschenbuch der Technischen Akusak Springer 1994 PEPI Orchester Chor Pubibum dcht strend Ref M Mechi H A Muller Taschenbuch der Technischen Akustk Springer 1994 sod rsa brandara inenen unbesetrt durchschrttich mt Stoff bezogen gut gepolstert Unterseiten der Stre perforert Ref DNA 1968 Tabele o 1 1 63 Hz 125 Hz 250 Hz 500Hz 1000Hz 2000Hz 4000Hz 8000Hz o w Class a r ES on aa ult 0 45000 0 45000 0 60000 0 73000 0 30000 0 75000 0 64000 0 64000 0 75000 c 4 63Hz 125 Hz 250 Hz 500 Hz 1000Hz 2000Hz 4000Hz 8000 Hz aiw Class m 4 1 Material Library Right side of Material List window In the material database every 1000 numbers have a category e g gypsum or wood In each category the first 100 materials should be reference materials within the category ODEON has suggested a few to begin with Hereafter there will be room for 9 different manufactures lists e g with number 1100 1199 In this way the consultant decide which manufacturer should be having what 100 numbers of absorption by writing the first 1 3 numbers in front of the manufacturers numbering in the material library editor The numbering of the main categories is given below e 1 99 Special materials e 100 999 Concrete e 1000 1999 Brick e 2000 2999 Ceramics e 3000 3999 Wood 49 e 4000 4999 Gypsum e 5000 5999 Steel e 6000 6999 Vinyl Plastic Plexiglas e 7000 7999 Carpets e 8000 8999 Curtains and blinds e 9000 9999 Natural materials e
126. dd or behaves strangely Solution If the Origo is situated in a point far away from the geometry the coordinate system may not display properly when projection is turned on in that case turn off the projection using the P shortcut In order to fully solve the problem the position of the origo should be altered as described in the example above The geometry displayed in the 3D View appears to be too small or large after re import If the geometry was initially imported using an incorrect unit then ODEON has defined its default view list in the 3DView in order to display that initial version of the geometry correctly To reset the view list use the ctri DEL Shortcut from within the 3Dview 32 2 5 The ODEON Extrusion Modeller A small modelling program the ODEON Extrusion Modeller is included with ODEON The program is found in the Windows start menu along with the ODEON program It can also be launched from within the ODEON editor Odeon 1985 2004 Odeon 1985 2004 For some geometry it may be more appropriate to draw the geometries in one of the other main planes XZ and YZ planes As an example the auditorium par model in the Figure has been modelled in the XZ plane using separate extrusion surfaces for the room the wall with windows holes the table and the windows The Extrusion modeller allows modelling so called extruded geometries in a graphic environment or in other words to draw geometries using the mo
127. ded resulting in a fall off by 6 dB per octave In the special case of a quadratic surface there will only be one limiting frequency below which the specular component will fall off by 6 dB per octave The attenuation factors Ki and Kw are estimates to the fraction of energy which is reflected specularily These factors takes into account the incident and reflected path lengths for ray tracing we have to assume that reflected equals incident path length angle of incidence and distance for reflection point to the closest edge on the surface all information which is not available before the calculation takes place 74 l for f gt f l Jr f gt f F Jor f lt f c a c a 2cw cos O d c d rer ZFL Qa where a Ao If we assume energy conservation then we must also assume that the energy which is not reflected specularily has been diffracted scattered due to diffraction This leads to the following formulae for our scattering coefficient due to diffraction Ss 1 K K xd s In order to compensate for the extra diffraction which occurs when a reflection appears close to an edge of a free surface the specular component is reduced by a factor 1 se The edge scattering coefficient is defined to be 0 5 if the reflection occurs at the edge of a surface saying that half of the energy is scattered by the edge and the other half is reflected from the surface area If the reflection point is far from the edge the edge scatte
128. deller is always in the par format and as such they may be combined in the ODEON Editor When combining different geometries from different sources some facilities in the parametric modelling format may be quite useful NumbOffset CoordSys Unit MTranslate MRotatex MRotateY MRotateZ MScale MReset and MPop Below is an example outline which illustrates how a number of geometries can be merged together into one parametric file HHH CoordSys X Y Z re 1 model data ee 1 model data NumbOffset 1000 _ avoid reusing point and surface numbers which have already been used CoordSys Y X Z swap coordinate axes if needed MTranslate 01520 Translate move geometries as needed ee 2 model data ae 2 model data NumbOffSet 2000 MReset restoring default origo CoordSys X Y Z restoring default coordinate system More model data HHH Li Using the 3DOpenGL display for model verification This display is very useful for detecting holes in the geometries Especially if stepping outside the model Arrow back shortcut and rotating the model using the Ctrl Arrow shortcut See the corresponding 3DOpenGL dropdown menu for more shortcuts When materials have not been assigned to all surfaces in the room surfaces will appear in random colours making holes easier to spot If materials have been assigned the colours will by default reflect the acoustic properties of the surfaces however it is possible to turn on the
129. dress Systems The model Copenhagen Central Station Array par is available in the Auditorium and Combined editions of ODEON This example demonstrates a PA system installed by Duran Audio which utilizes 20 beam steered line arrays Several measured parameters are available including STI 2 2 Guidelines on room modelling Whether you choose to import your rooms from SkecthUp by using a CAD program by typing your rooms directly into a text file or by using the ODEON Extrusion Modeller there are considerations that are common to either case Some guidelines of general nature are given below Default coordinate system To make it as easy as possible to operate ODEON the following orientation of room geometries should be applied using a concert hall as the example e X axis pointing towards the audience e Y axis pointing to the right as seen from the audience e Z axis pointing upwards Recommended size of a surface An important theoretical consideration concerns the size of surfaces in a room model The classical laws of geometrical acoustics are such that for the purpose of calculating how much energy is reflected all surfaces are considered to be infinitely large in comparison to the wavelength For practical room models surfaces are not infinitely large and ODEON is to some degree able to take into account the limited size of surfaces in calculations using the Reflection Based Scattering method see section 6 5 S
130. e e g 44100 Hz or 48000 Hz Not in use yet Specifies the sample rate if the filter for a transducer is a FIR filter PhaseCF2 false True if directivity balloon containing phase angles should be used for the transducer See section on phase balloons at the end of this appendix For transducers which have been assigned a directivity pattern in the Common Loudspeaker Format CF2 version 2 the transducer may belong to multi part speaker e g having separate balloons for the low frequency unit and the high frequency unit in a loud speaker I that case there may also be assigned a filter e g low pass and high pass to each transducer This is what CLFPartNo CLFPartMax and CLFFilter is for CLFPartMax 2 152 How many transducers with CLF2v2 balloons belongs to this speaker array CLFPartNo 1 Which part of a multi part CLF is this transducer e g part 1 or two of a two way speaker CLFFilter true Is a filter assigned to the transducer in the CLF2v2 file e g a low pass or a high pass filter Domain Frequency So far options are Frequency for beem steered otherwise Octave future should include Time in which case an attribute SampleRate is needed e g 44100 Hz or 48000 Hz Nodes Position Orientation DomainData DomainData node There are two kinds of domain data so far also described in figure E2 If Domain Octave simple octave band equalization is applied to each transducer Thi
131. e HeadAndPhone EE and HeadAndPhone EC directories To import a set of HRTF s select the Tools Create filtered HRTF s option then select the specific HRTF ASCII file e g C ODEON HeadAndPhones EE Kemar ascii_hrtf finally specify miscellaneous parameters when the import dialog appears for help on these parameters please press F1 from within that dialog You may also desire to import a set of HRTF which has already been imported in order to specify an alternative filtering approach e g to enhance the HRTF s in different ways Remember If you want to obtain a neutral impulse response from the Joblist without HRTF filters you have to select the Unity ascii_hrtf The BRIR s in the single point response will be exactly the same for both ears in this case This is a result like the one you would obtain with a microphone during a measurement Even when no HRTF is used filtering is still applied In addition since all calculations in ODEON are in energy domain artificial phase is added to make the energy based impulse response look like a pressure real impulse response Therefore an impulse response obtained in ODEON will never match a real impulse response from measurements in detail The overall energy behaviour will be the same but reflections might vary a lot Headphone filters It is possible to compensate for non linear frequency response of headphones When a headphone is selected ODEON will filter the output through a m
132. e 2 Log2 8 3 Square MOE Square root Sqrt 2 1 41421356237309515 Sign 2 Sien 0 Sin T ala 188 Appendix F Importing geometries compatibility g f of an ArchiCAD IntelliCAD AutoCAD Revit sS SAL ATS 3dstudioMAX Ss ql ar gt SONarchitect Odeon Import Odeon _ Editor Odeon a IFC2SKP Google SU2Odeon SketchUp Extrusion modeller Recommended Odeon Export gt Optional Oe 3D modelling program Converter cad CATT geo 189
133. e a good idea to define a point source somewhere in the middle of the room Open a 3D Investigate Rays display and run it with e g 1000 rays with a Max reflection order of zero This tests whether any holes can be seen from the source position and should reveal any gross problems The tracks of lost rays will show outside the room boundaries and indicate whereabouts in the room problems occur If rays are being lost and you have an idea of which part s of the room is are leaky a number of things may be done e Reduce the value of Max accept warp in the Room setup at the Model Air conditions page Then run the 3DGeomtry Debugger Warnings will appear if surfaces have a warp or an overlap above the acceptable range This may reveal slight warps of surfaces in the leaky region of the room which then have to be reduced as far as possible by revisions to the geometry file e Use the 3DvView or 3DOpenGL for inspection of the model to study the region s under suspicion It may turn out that a surface is missing or does not join to its neighbours in the expected manner It may help to zoom regions in question with the Highlight surfaces Show corner numbers and coords and Modelling options switched on 42 Sources and Receivers Once the geometry of a room has been loaded into ODEON sources and receivers have to be placed as described in the Quick Start Guide Three types of sources are available generally in ODEON Point sources line
134. e file Jnn option is activated in the Auralisation Setup 7 The Auralisation setup offers various control options for auralisation and HRTFs It looks like the screenshot in the following figure Remember Always press F1 when you need to learn more about specific functions in ODEON The smart help will open to a page related to the active window Many times the ODEON help file complements with the manual and offers extra information To learn more about Surround playback and how to define a speaker rig which is not covered here click the Define speaker rig button and press F1 55 Once a Single point response is calculated it is possible to play the BRIR clicking the Play Single Point BRIR button The BRIR may give a first clue as to how the room sounds and it also allows some evaluation of the quality of the calculated point response e g whether to use a higher Number of late rays in the Room setup Although the BRIR may sound a little rough it may work quite realistic when convolved with a signal less transient than an ideal Dirac delta function can be googled To get a more realistic presentation of a BRIR as it would sound in the real world you might want to convolve it with the Clapping signal file an anechoic recording of hands being clapped which is eventually a less transient signal than an ideal impulse by Auralisation setup lolle General auraligation settings Apply dither and noise shaping Wave result f
135. e gt Delete files gt Delete calculation files When all calculations in a room has been carried out or use the File Tidy directory when disk space becomes small Tidy directory will clean a complete directory and its sub directories and delete all calculation files for all rooms A directory in this context can also be the c drive in which case it will be a lengthy process Single Point Multi Point and Grid Response the receiver dependent part Having traced rays around the room and stored the data of ray histories the next step is to place the receiver at a specific point and so to speak collect the reflections there These point response calculations are the receiver dependent part of the calculations at this point the contributions of direct and reflected sound are collected at the receiving point allowing the calculation of the results known as Single Point Multi Point and Grid Response When more than one receiver is involved the receiver dependent part of the process is simply repeated for each receiver When more than one source is involved the response at a given receiver is simply the sum of the responses from the individual sources each delayed appropriately if a delay is applied to the source ODEON automatically takes care of handling which of the calculations and result files are currently consistent with user entered data erasing those that are no longer valid Thus in some situations you may experience that Rays tracin
136. e response button As the energy of a handclap is limited ODEON may not be able to derive energy at all in particular lower frequency bands due to background noise still this may be better than nothing You can read more on how to make a sketch measurement online on the application notes page http www odeon dk application notes 11 1 Technical Background High quality measurements using the sweep method The room impulse response measurements in ODEON can be performed using linear and exponential sweep signals which is superior in suppressing background noise leading to a high signal to noise ratio Muller 2001 and high immunity to distrotion A sweep is simply a pure sinusoidal signal with frequency increasing monotonically in time The sweep method for measuring impulse responses is the least sensitive method to time variance and distortion and it is expected to work well on most hardware systems ODEON uses sweeps to excite a room and records the response to this signal at a microphone receiver position in real time This sweep 104 response is then deconvolved to give the impulse response between the source and the microphone For this process upward sweeps are used instead of downward meaning that low frequencies are played first The energy of low frequency sound usually takes more time to decay than the energy of high frequency sound in a room Therefore upward sweeps are preferable over downward ones in order to ensure tha
137. e rotation dialog and specify the rotation angle in this case 25 degrees finally delete the rotation point from the surface Examples A few examples on extrusion models are installed with ODEON the examples are located in the ODEON rooms oes Directory The best way to learn about benefits as well as limitations of the extrusion modeler may be to load the examples investigate the surfaces e g scrolling the point and surface tables and to load the models into ODEON in order to investigate the models when they become extruded Special extrusions There are a few extrusion surfaces which are treated differently by the extrusion modeler 1 A surface with an extrusion height of zero will produce one and only one horizontal surface no matter if a bottom or top surface is selected 2 An extrusion surface which only contains two points will only produce one vertical surface neither bottom or top surface is produced only a single extruded surface if the extrusion height of this surface is zero then no surface is produced an exception to the exception 2 6 Model check in ODEON The geometry file is the first file used by ODEON when assigning a file from Files Open Room model When assigning a new or modified room its validity is checked 39 The check performed by ODEON involves checking whether data is consistent and in the correct format but not whether a meaningful geometry is being defined If the geometry passes
138. e that sub volume e g if two sub volumes share the same floor surface then area and surface estimates for the statistical calculations may not be entirely correct In that case it is possible to set a volume manually The classical mean absorption coefficient is given by gt Sa DS A where S and are the area and absorption coefficient of the it room surface respectively The modified mean absorption coefficient as experienced by the particles is gt Ha l LH a where Hi is the number of hits on the itt room surface In ODEON both of these mean absorption coefficients are inserted in the Sabine and Eyring formulae to calculate reverberation times the classical values are labelled Sabine and Eyring and the values using the modified mean absorption coefficient are labelled Modified Sabine and Modified Eyring The mean absorption coefficients used for the Arau Puchades formula are derived in similar ways except that separate values for surface hits area and the corresponding mean absorption 66 coefficients are calculated as projections onto each of the main axis of the room The Sabine Eyring and Arau Puchades formulae require a value for the room volume which ODEON estimates from the mean free path experienced by ray tracing using the well known relation for the mean free path in a 3 dimentional room where V is the room volume and S the total active surface area From version ODEON 6 5 the ray tracing
139. e the airborne propagation of sound through a wall It cannot calculate the structure borne sound propagation nor flanking tranmission ODEON requires the transmission properties coefficients of the wall as input data exactly like for absorption coefficients When a composite structure such as a sandwich panel is used in ODEON simualtions the transmission properties shall be obtained externaly in four ways e By the datasheet from the manufacturer of the structure e By measurements e By hand calculations using analytic formulas e By using software dedicated for calculating transmission coefficients of different materials eg INSUL by Marchal Day Acoustics Copying transmission data from MsExcel A material to be copied may contain a leading comment name and must end by 24 floating point values denoting the 24 reduction indexes from 50 Hz to 10000 Hz the first value is always assumed to be for the 50 Hertz one third octave band if the number of bands is less than 24 then the last value is used for the bands above if more than 24 values the values above 10 kHz are discarded To copy the values from MsExcel mark the relevant cells and press Ctri c to paste the values into Transmission dialog simply press ctri v Likewise to copy data from the Transmission dialog to MsExcel press ctri c while the dialog is the selected window in ODEON then select the first cell in MsExcel or which ever cell is relevant and press ctri v
140. e time for easy comparison of results as ODEON to a large extend aims at displaying measured and simulated results in similar displays It has been made easy to transfer measured room acoustics parameters to the Multi point response displays available from the Jobiist facilitating comparison of measurements and simulations This allows getting statistic values for a number of measured receiver positions even if no simulations have been made Two buttons are included in the toolbar in ODEON 1 Measure impulse response By 2 Load impulse response p These buttons are always active whether or not a room has been assigned The Room Acoustic l Fe i Parameter List button remains always active too in order to make it possible to edit parameters that are displayed in the Load impulse response window there are limited options in the Basics amp Industrial edition Sketch measurements If you are not bringing the full set of measurement equipment PC amplifier loudspeaker etc you can record a hand clap the popping of a balloon or a paper bag or indeed the gunshot of a start gun e g using a smart phone with an App installed that supports recording wave files such applications are available from the relevant application stores for Android as well as IPhones The clapping hands is the source and the phone acts as the receiver microphone The files containing impulse responses can be loaded into ODEON using the Load impuls
141. e wall 1 2 12 11 HHH Below the box shaped room is modelled using the Box statement which is the easiest way to create this simple geometry A MTranslate statement is used to insert the Box at the same position as in the three other examples Parametric sample BoxStatement Par HHH const L 6 const W 4 const H 2 7 MTranslate 1 2 0 0 Box 1 lw h tb Walls and floor HHH 183 Modeling a cylinder This example shows two different ways to create a cylindrical room with a floor and a ceiling In the first example the room is modelled using the Cylinder statement Parametric Sample CylinderStatement Par HHH const N 16 const R 15 const H 10 Cylinder 1000 N R 360 H TB Cylindrical room HHH The Cylinder statement is of course the easiest way to model a cylinder however sometimes more flexibility is needed e g different radius in top and bottom In the second example the corners in the room are modelled using the CountPt statement and the cylindrical surfaces are modelled using the RevSurf statement Notice that the number of points created by the CountPt statement is one higher than the number of sections in the RevSurf statement The bottom and top of the room is modelled using the Surf statement notice that points used by these surfaces are referenced using the statement 100 gt 100 Sections 1 rather than writing each of the sequential points this is not only a faster way to write things it also allows a rapid change to the num
142. easured result with different transition orders and number of rays lol ex5 JobsGroupBox Test option 1No description Transition order 0 2No description Test number of late rays and transition order automatic number of early rays C 3No description Number early of rays 0 a tis l aves 4No description ee 5 No description Number late of rays 20000 Pate hg ae EENT ee CA ae C 6 No description Test number of early rays and transition order fixed number of late rays LOI 7 No description 8 No description 9 No description 10 No description Max transition order to test E Error as a function of transition order and late rays Average error for frequency range 500 to 2000 Hz M Transition order M Transition order M gt Transition order M Transition order wW Transition order M Transition order Average error in JND s 1 6 1 4 100 200 500 1000 2000 5000 10000 Number of rays Receiver independent ray tracing Tracing rays 10653 of 40000 Job 1 Remaining time him s 00 00 08 Two step Sound Source Calibration This tool removes the requirement of fixed conditions in the measuring chain which is common rule in ordinary calibration procedures such as diffuse field calibration and free field calibration The two step sound source calibration adds a second step to the normal reverberation room or free field calibration methods in order to allow compensation for any shi
143. ed see p 1 7 Remember to inform the end user to use headphones when listening to the samples 61 Publishing results on an ordinary audio CD If a CD R drive is installed on your PC it is quite easy to transfer the wave result files into an ordinary audio CD most CD R drives comes with the necessary software for this purpose Most people have access to a CD audio player so publishing results on an audio CD makes it easy to send demonstrations to clients etc without worrying about whether they have a PC with a soundcard of a reasonable quality Again when publishing examples make sure that copyrights are not violated You are free to publish examples which are calculated using the anechoic examples supplied with ODEON For the orchestra recordings the conditions as specified in the software license agreement 8 must be observed see p 1 7 Remember to tell the end user to use headphones when listening to the samples 5 2 Making offline auralisations creating WAV files In the ODEON Auditorium and ODEON Combined Quick Start Guides detailed instructions are given for performing a streaming convolution Streaming convolution is practical for making fast auralizations but it cannot store any WAV files Moreover you can only listen to a single convolution at a time For greater flexibility mixing different convolutions together making individual adjustment of each channel delay and level and storing auralisation results in wave you h
144. een source and receiver is not present Cso Clarity C 10lo Eas 80 2 dB 80 20 Ds0 Definition D Eoso oe Ts Centre time SPL Sound Pressure Level SPL 10log E dB The value of SPL becomes equal to the value of G which is the total level re to the level the source produces at 10 m in free field as defined in ISO 3382 1 2009 when an OMNI directional source type and a power of 31 dB Octave band is selected from within the appropriate Point Source Editor LFso Lateral Energy Fraction The LFso parameter has a high correlation with the apparent source width ASW as shown in Bradley amp Soulodre 1995 L Average Ni 00042 oO Lj Average lt 10log gt gt E cos Z dB Ni 25Hz t 80 The calculated value of Lj Average is according to ISO 3382 1 2009 a source with an OMNI directional source type and a power of 31 dB Octave band is selected from within the appropriate Point Source Editor This parameter is suggested in Bradley amp Soulodre 1995 and has a very high correlation with the subjective parameter Listener envelopment LEV Stage Parameters Stage parameters are calculated as a part of the Single Point response Auditorium and Combined versions only if the job only contains one active source the active source is a point source and the distance between receiver and source is approximately 1 metre 0 9 to 1 1 metre The parameters are called Support for
145. eflections from own head and torso The other difference is that reflections are added with random phase 6 10 Calculation method for Reflector Coverage 25000 rays are send out from the selected source if the rays hit one of the surfaces defined as reflector surfaces at the Define reflector surfaces Menu a cross is painted where the reflected rays hits the room surfaces Note that the value of the Transition order is taken into account if it is zero and the Lambert scattering is active the chosen reflectors will exhibit a degree of scattered reflection corresponding to their scattering coefficients Sound from line and surface sources will always reflect scattered if the Lambert scattering is on 81 Calculated Room Acoustical parameters This chapter will shortly describe the derivation of energy parameters for Single Point Multi Point and Grid response calculations for the Industrial edition only EDT Tso SPL SPLa and STI are available All the parameters are derived on the assumption that the addition of energy contributions from different reflections in a response is valid This manual will not cover the use of the individual parameters in depth and suggestions on ideal parameters choice should only be sought of as a first suggestion instead refer to relevant literature e g some of the following references for a further discussion on parameters and design criterions Auditorium acoustics as Concert Halls Opera Halls Multipurpo
146. eft side vertical points in Dome2 are stored in PlistB 179 In the special case where the revolution angle is 180 all points are stored in PlistA and the number of vertical subdivisions is stored in ONVert The example shown was generated with the following code HHH Const N 16 Const W 10 Const H 3 Const L 10 Dome2 1 N W H 270 Dome calotte HHH Hint The cylinder can be made elliptical using the MScale statement DebugIsOn and Debug The debug options are useful when creating large or complicated geometries in the ODEON par format Using these facilities can speed up geometry loading when loaded for preview only and allow debugging of parameter values in geometry files DebugIsOn is a Boolean which can be set to TRUE or FALSE the syntax is DebuglIsOn lt Boolean gt Typically you will insert the DebugIsOn flag in the beginning of the geometry file in order to investigate parameter values when loading geometry When this Boolean is set to TRUE e ODEON will not prepare the geometry for calculation as result the loading of rooms is speeded up e ODEON will enable debugging of parameters with the Debug statement The syntax for Debug is Debug lt debug string gt In effect anything can be put after the Debug keyword i e you may put a complete copy of a line in the par file there The contents following the Debug keyword is evaluated or if it can t be evaluated then echo ed directly to the debug window in ODEO
147. eld preferable a reverberation chamber A group of source receiver combinations is required According to ISO 3382 2 two source positions and three receiver positions give 6 source receiver combinations and correspond to a moderate precision Engineering precision Down to 2 source receiver combinations correspond to the lowest acceptable precision Survey precision while up to 12 source receiver combinations correspond to a high precision The impulse response for each combination should be recorded and saved Afterwards the files can be loaded using Tools gt Calibrate measurements gt Diffuse field calibration You should select and load all impulse responses obtained for the same calibration at once use the Ctrl key and mouse in the Open calibration files menu when displayed then ODEON will calibrate the level according to the average of the values The volume of the reverberant room must be stated 110 Free Field Calibration Ideally this should be carried out in an anechoic room but in many cases it can be carried out in a relatively dry room The receiver should be placed close to the source and the impulse responses should be recorded ODEON will try to capture the very early part of the response corresponding to the direct sound According to the distance between the source and the receiver and their heights from the floor the arrival time of the 1 reflection is calculated so that the impulse response is truncated at this place Idea
148. elling entities such as CountSurf Box Cylinder etc and in particular to make the automatic surface numbering work without any problems when the NumbOffSet is set to Auto If having problems loading a room due to the reasons just mentioned ODEON will either give an error message that surfaces are repeated in the geometry file or that materials are not applied to all surfaces In these cases you may wish that ODEON use the old numbering mechanism this can be done using the Version4 par file As the first line in the geometry file just after the sign type flag in the Version4 TRUE Once the old incompatible code ends Version4 may be set to FALSE again 1 3 What s new in ODEON 13 The new version of ODEON is packed with a bundle of attractive features and enhancements Improved Interface All lists and input fields have been freshened up in ODEON 13 Materials list source receiver list job list etc they all have been equipped with shorting functions better display colors and easier editing properties such as multi delete copy paste etc oom matenal library Elmia RoundRovdin2 detailed L8 z pa Humber Speoficabon o Transparent Spedal material j 30004 z a 100 absorbent Special material 30004 Z 2 100 reflecting Spedal material 30008 ril 10 10 absorbent 3000s St 11 Some thick plasterboard 30004 2 20 20 absorbent zons Qa 30004 Normal 30004 Normal 30004 065 0 000 Noma
149. en Sometimes this is quite acceptable because we are just interested in rough results at other times we are interested in results as good as possible In any case being aware of the sources of error may help getting the maximum out of ODEON The sources of error or at least some of them are e The approximations made in the ODEON calculation algorithms e Inappropriate calculation parameters e Imprecise material absorption coefficients e Imprecise material scattering coefficients e Not accurate or optimal geometry definition for use in ODEON e Not accurate measured reference data to which simulations are compared Approximations made by ODEON It should be kept in mind that algorithms used by software such as ODEON are only a rough representation of the real world In particular the effect of wave phenomena are only to a limited extend included in the calculations There is very little to do with this fact for you the user except to remember that small rooms and rooms with small surfaces are not simulated at high precision Optimum calculation parameters A number of calculation parameters can be specified in ODEON These settings may reflect expected reverberation time a particular shape of the room or a trade of between calculation speed and accuracy Number of Late rays ODEON by default specifies a suggested Number of late rays to be used in point response calculations This number is derived taking into account the aspect
150. en diffraction is only carried out for the primary source not for its image sources when it is not visible from the receiver Screen edges are detected automatically by ODEON and the shortest path around an object is used for the calculations For more information on Figure 4 Example of a two point diffraction path displayed in Single point response gt 3DReflection paths as it has been detected by Odeon methods and limitations please see chapter 3 If one or more diffra ction contributions are included in a point response they can be viewed as part of the Single point response gt 3DReflection paths when all reflections are displayed the A shortcut Surface scattering Job calculations Global Estimate and Quick Estimate If Surface scattering is set to Actual all directions of late reflections are calculated using the scattering coefficients assigned to the surfaces in the Materials list or according to the reflection based scattering method see below If the scattering coefficient is 10 the new ray direction will be calculated as 90 specular and 10 scattered random direction due to the Lambert distribution If the Scattering method is set to None s 0 scattering is not taken into account thus all reflections are calculated as specular If the surface scattering is set to Full scatter S 1 all surfaces will scatter 100 The Full and None scattering methods ARE NOT RECOMMENDED except for educational or research purposes
151. en transmission walls are composed from a single surface two surfaces and three surfaces Once the room has been loaded into ODEON additional information is available in the Notes editor using the Shift Ctri N shortcut In the Joblist some example setups have been prepared for calculations in source rooms as well as in receiver rooms 145 Appendix C Description of XML format for import of array loudspeaker data The ODEON XML format for array loudspeakers allows import of array parameters that specifies how a number of transducers are combined together to form an array loudspeaker Arrays can be beam steered a digital filter being assigned to each transducer or a more conventional array where each transducer is feed directly or through an equalizer with a delay etc External files needed a directivity pattern for each transducer Currently the following directivity formats are supported Common loudspeaker format CF1 CF2 and ODEON s native format So8 The frequency range needed by ODEON covers the full octaves from 63 Hz to 8 kHz however in order to make the format as versatile as possible e g for future use or for use in other programs frequencies from lower and higher bands can be included ODEON will just ignore those bands XML eXtensible Markup Language has been chosen as the format for import of data for arrays loudspeakers because e It s astandard ISO 8879 1986 See also http www w3 org XML e It allow
152. eometry Click the ODEON icon inside the editor in order to save the modified geometry and reload it into ODEON Other coordinate manipulations to the geometry may be desirable in particular the CoordSys statement described in section 0 may be useful Trouble shooting Below common problems when importing are described Problem with zoom or translation in the 3DView Model appears in a strange position on the screen and zoom translation does not work as expected This problem is probably caused by some small invisible and irrelevant surface s located at odd position s in the imported model Solution 1 Try importing the geometry once again with some of the entities unchecked turned off it may be that some of the entities such as 3DPOLY or the like were not intended to be surfaces Solution 2 Removing the unwanted surface e Inthe 3Dview turn on the Modelling options M shortcut look out for odd positioned points e Move the mouse cursor to the position of the odd point read one of these point numbers e Click the ODEONEditor icon to open the par file remove the point and try to reload the room by clicking the ODEON icon in the Editor Now ODEON will hopefully report an error stating a surface is referencing the point which no longer exists e Remove that surface along with ALL the points it is referencing out comment them e Reload the room Problem with display of coordinate system The blue coordinate system looks o
153. erences in JNDs lt Multi point response parameters job 1 Simulated mode 3D Sources and Receivers Energy parameter curves 1 Energy parameter curves 2 Statistics Energy parameters Measured versus Simulated D 50 at 1000 Hz Receiver 1 0 8 Simulated X Measured 0 7 0 6 0 5 D 50 D 50 0 4 v 0 3 0 2 0 1 oO R4 R5 ise a R1 R2 125 250 500 1000 2000 4000 8000 Receiver Frequency Hertz 136 Multi point response parameters job 1 Simulated mode e 3D Sources and Receivers Energy parameter curves 1 Energy parameter curves 2 Statistics Energy parameters Measured versus Simulated gp gp gp T 30 at 1000 Hz Receiver 1 2 2 1 8 1 8 1 6 1 6 1 4 1 4 12 12 z ol o 0 8 0 8 0 6 0 6 0 4 0 4 0 2 0 2 0 0 z fr oe oe o3 amp 8 8 8 8 2 8 N fp a O O O N t Receiver Frequency Hertz Expert s parameters Crossover Method Defines the way individuals mix together to form a new individual Two options are available Gene exchange and Vector mixing With the Gene exchange option new individuals are created completely genetically by re arranging Frequency optimization Evolution Method Controls how parents are selected for generating new individuals Five options are available Roulette Parents are selected based on a roulette wheel selection Each parent is assigned a portio
154. erenrenrenreseesees 156 Appendix E Mathematical expressions amereree ee cre memes errr rer rrmeerre rere etter errerer ye reer etre serene er Te 188 Appendix F Importing geometries compatibility vresicssscsssseveinetencesdecsnewsoasvareawesseensousaeaeneressessvarsbeneveess 189 1 ODEON Installations 1 1 Installing and running the program To run ODEON your PC must be running one of the operating systems supported by ODEON e 32 bit 64 bit Windows Vista e 32 bit 64 bit Windows 7 e 32 bit 64 bit Windows 8 ODEON comes on a USB drive containing the edition of ODEON purchased Basics Industrial Auditorium or Combined As an alternative you may download the most resent updated version including service updates that has been made since the production of the USB from www odeon dk updates To install the program a Double click on the file with the name of the edition you wish to install e g InstalLOdeon12Combined exe to install the program b To run the program the supplied hardware key Rockey 6 Smart R6 Smart must be inserted into the USB port on the PC If you start the program without the hardware key it can only be used in viewer mode c When you have installed the ODEON software please check for the most resent updates at www odeon dk updates 1 2 Upgrading from previous versions If you are upgrading from previous versions of ODEON read on to learn about the changes in ODEON Below is a list of issues which you
155. es This section gives an introduction to the use of the line array option in ODEON The examples are chosen in order to demonstrate some basic properties of line arrays and they are not representing recommended solutions It is a delicate process to adjust and optimize a line arrays sound system for a particular room and the knowhow and technique needed for that is beyond the scope of this manual 10 1 Stacking the units By stacking a number of loudspeaker units on a vertical line with a constant distance d between the centres of the units the first thing to note is that the splay angle changes The sound is radiated in a more or less concentrated beam and the splay angle narrows in when the array gets longer With N units the length of the array is L d N 1 and this should be longer than one wavelength in order to obtain the narrowing of the spay angle This can be expressed by a lower limiting frequency fi lt L d N where c 344 m s is the speed of sound However the concentration into a single beam only works at frequencies below the upper limiting frequency i e when the distance between the units is short compared to one wavelength Above fu the sound radiation breaks up into a number of directions In fig 10 1 is shown the near field radiation at 1 kHz with different number of units in the array from 1 to 11 The unit in the example is SLS_LS8800 CF2 imported from the CLF collection of loudspeakers and the distance be
156. es for use in ODEON This ensures that it is easy for loudspeaker 44 manufactures to make these data available Links to loudspeaker manufacturers currently providing binary distribution files can be found at the download page at www clfgroup org If apparently the data of interest is not available from the manufacturer of interest then assist the Clf group by encouraging the manufacturer to make such data available free tools for this purpose can also be obtained at the CLF group s homepage File location for directivity files No matter if files are in the CF1 CF2 or in ODEON s native So8 format the files should be stored in ODEON directivity directory which is specified inside ODEON at Options Program setup Directivity files location The files may be stored in subdirectories to this directory allowing loudspeaker directivities of different brands to be located in separate directories e g C ODEON DirFile ManufacturerA or C QDEON DirFile ManufacturerB We have taken the opportunity to create a number of folders for manufactures which do supply loudspeaker directivity files in the CLF format Using these predefined directories it is easier to move a room from one PC to another without breaking file linkage 3 4 Creating new directivity patterns in the ODEON So8 format Tools for creating directivity patterns in the ODEON So8 format can be found at the Tools gt Directivity Patterns gt Creat directivity pattern in plot edit
157. eter curves Freque D ODEON 13 PicturesFromimpulseResponses IR_BasementCropped wav Raw decay curve at 1000Hz M E Measured lv Onset time Truncation time V Noise floor SPL dB 1 2 Time seconds A typical impulse response with a well defined healthy noise floor that is easy to be detected by ODEON Impulse response with fluctuating noise floor The tail of the impulse response in the following figure is polluted by a high degree of non flat noise floor full of dips and notches This can be an indication of presence of strong background noise during the measurement and the signal to noise ratio is insufficient for the calculation of all room acoustic parameters Many times ODEON displays a fluctuating noise floor by a dashed blue line However in this particular case this prolonged fluctuating noise floor cannot be detected as fluctuating and ODEON places a normal blue line instead In such a case it is wise to visually inspect the fluctuating envelope of the noise floor and repeat the measurement under less noisy conditions 116 C Odeon12Combined Measurements Elmia RoundRobin2 detailed ImpRespFileO wav Raw decay curve at 1000Hz Jv E Measured Jv Noise floor Onset time Truncation time 02 03 04 05 0 6 f P 0 9 1 1 1 time seconds Odeon 1985 2013 Licensed to Odeon An impulse response with highly fluctuating noise floor Series of impulse responses The next
158. eters Part 1 Performance spaces Affronides S 1996 Introduction to Signal Processing J Prentice Hall International Algazi R 2001 August 12 The CIPIC HRTF Database Retrieved June 16 2011 from The CIPIC Interface Laboratory Home Page http interface cipic ucdavis edu Bamford J S 1995 An Analysis of Ambisonics Sound Systems of First and Second Order MSc Thesis Waterloo Ontario Canada University of Waterloo Barron M 1993 Auditorium Acoustics and Architectural Design London E amp FN Spon Barron M amp Marshall A H 1981 Spatial Impression Due to Early Lateral Reflections in Concert Halls The Derivation of a Physical Measure Journal of Sound and Vibration 77 pp 211 232 Beranek L L 1962 Music Acoustics and Architecture New York John Wiley Beranek L L 1996 Concert and opera halls how they sound Acoustical Society of America Beranek L L 2004 Concert halls and opera houses music acoustics and architecture Acoustical Society of America Beranek L L amp Hidaka T 1998 Sound absorption in concet halls by seats occupied and unoccupied and by the hall s interior surfaces J Acoust Soc Am 104 3169 3177 Bobran H W 1973 Handbuch der Bauphysik in German Berlin Verlag Ulstein Bork I 2000 A Comparison of Room Simulation Software The 2nd Round Robin on Room Acoustical Computer Simulation Acta Acustica 86 943 956 Bork I 2005 Report on t
159. extended at will by the user using the material editor available from the Material list If you should wish to add several materials e g by copying them from some other file this is possible by editing the file using the OdwkEdit editor which is also available from within the Materials list and following the ODEON material format Furtherit is possible to import multiple material from a datasheet from the manufacturer This is further described at www odeon dk acoustic absorption data Special Materials There are three special materials in the library e Material 0 transparent e Material 1 totally absorbent e Material 2 totally reflective Although the material library Material li8 may be edited materials 0 1 and 2 must remain as originally defined Data format for materials in Material Lis The data format for a material in Material Lis is very simple each material is described by two lines 50 ID_Number Descriptive text up to rest of line a63 al25 a250 a500 alk a2k a4k ask ID_Number must be a unique number between 0 and 2 147 483 647 Absorption coefficients on second line must be floating point within the range 0 1 the line containing 8 floating point values Descriptive text should consist of a description of the material and a reference to the source where the absorption is documented This reference can be a link to manufactures internet page or the measurement report written as href http link When a materi
160. f NAM 86 pp 257 260 Aalborg Denmark Rindel J H 1992 Acoustic Design of Reflectors in Auditoria Proceedings of the Institute of Acoustics vol 14 Part 2 pp 119 128 139 Rindel J H 1993 Modelling the Angel Dependent Pressure Reflection Factor Applied Acoustics 38 223 234 Rindel J H 1995 Computer Simulation Techniques for Acoustical Design of Rooms Acoustics Australia pp 81 86 Rindel J H 1997 Computer simulation techniques for the acoustical design of rooms how to treat reflections in sound field simulation Proceedings of ASVA 97 pp 201 208 Tokyo Rindel J H amp Christensen C L 2008 Modelling Airborne Sound Transmission between Coupled Rooms Proceedings of BNAM 2008 17 19 August Reykjavik Iceland Rindel J H Nielsen G B amp Christensen C L 2009 Diffraction around corners and over wide barriers in room acoustic simulations Proceedings of the 16th International Congress of Sound and Vibration 5 9 July Krakow Poland Rindel J H Nielsen G B amp Christensen C L 2009 Multi source auralisation of soundscapes in large buildings Proceedings of Euronoise 2009 26 28 October Edinburgh Scotland Schroeder M R 1970 Digital Simulation of Sound Transmission in Reverberant Spaces J Acoust Soc Am 47 424 431 Stroem S 1979 Romakustisk prosjektering in Norwegian Anvisning 20 Oslo Norges Byggforsknings institutt Vigeant M C Wang L
161. fect of the units being more than one wavelength apart from each other If the phase is shifted by one period another angle of radiation is found Geometrically the angles that correspond to 7 periods phase shift can be calculated from 6 arag ATMS where fis the frequency and i 1 2 etc The theoretical angles of radiation corresponding to this example are shown as a function of the frequency in Table 10 2 below The direction of the main lobe relative to the horizontal direction is 100 10 for At 0 1 ms and 19 At 0 2 ms The results for 7 1 explain the upward side lobes seen at 2 and 4 kHz in fig 10 4 and 10 5 Table 10 2 Calculated angles of radiation from the example array with d 0 20 m Positive angles are downwards and negative angles are upwards Calculated for two different values of the delay per unit At Delay for transducers in an array should not be confused with delays assigned to the overall delay assigned to a point source or array source in order to benefit from the HAAS effect although the devices used in order to obtain the delays may be identical e Delays assigned in order to benefit from the HAAS effect are mainly supposed to have effect in the time domain because loudspeakers are supposed to have significant distance between each other and multiple reflections will blur the effect in the frequency domain anyway e Delays assigned to the individual transducers in an array on the other hand are
162. ft in the broadband gain from the calibration to the field measurement condition Improved Bass Response in Auralisation Enhanced Bass response is audible in the auralisation algorithms of ODEON 13 The bank of octave band filters has been extended towards 16 Hz which provides a better low frequency auralisation Auralisation in Measured Impulse Responses Record an impulse response in an existing room Go home and try how different signals would sound inside the room Speech music noise anything now can be convolved and auralised exactly as itis done so far in simulations 13 Distortion Noise 13 85 dB at 2000 Hz use longer sweep Measured response D CAHRISMA Impulse responses Saint Sophie S2 S2R6sbroad wav oE Raw Impulse response broad band Raw decay curve broad band Decay curves broad band T 30 10 35 seconds Decay curves all bands Energy parameters Parameter curves a gt D CAHRISMA Impulse responses Saint Sophie S2 S2R6sbroad wav Ray Impulse response broad band oO p x E 1 Ny OO ND Kk OO Signal source directory root directory X Signal fie Agora shortwav o 4 6 Convolution a Wave fie folder c Jsers george AppData Local Temp z Play Convolution on Frequency Phase and Cepstrum Response A recorded impulse response can be further edited in the ODEON measuring system An FFT frequency response is available now as well as ph
163. g sand grass water e 10000 10999 Doors Windows furniture inventory e g organ pipes ventilation grills bookshelves e 11000 11999 Audience areas people e 12000 12999 Mineral wool e 13000 13999 Wood wool and alternative porous absorbers e 14000 14999 Slit absorbers Micro perforated absorbers miscellaneous Special Materials Material 0 transparent Assigning Material 0 to a surface corresponds to removing the effect of the surface completely from all calculations Hence surfaces with this material assigned e Offer no hindrance to rays either in energy or direction e Are excluded from the calculated active surface area of the room and therefore do not affect the estimate of the room s volume produced by Global Estimate or Quick Estimate Reverberation This facility can be used to temporarily remove surfaces such as doors or reflectors from the room or to define a phantom surface over which an energy map a grid is to be plotted Material 1 totally absorbing The totally absorbent material Material 1 may be used for modelling outdoor situations e g an open roof This is the only material which will stop the rays during ray tracing and no reflections are generated from surfaces assigned this material Editing and extending the Material Library The materials displayed in the left side of the Materials List window resides in an ASCII file called Material Li8 This library provided with ODEON may be altered and
164. g calculations have already been done are still valid in other cases they have to be recalculated The Early Reflection method Early reflections in ODEON are reflections generated by point sources while the 26 reflection order is less than or equal to the Transition order specified in the Room setup For each of the image sources found in the Early part of the receiver independent calculation ODEON checks each to determine whether it is visible from the receiver Images may be hidden because walls in the room block the reflection path to the receiver or because the receiver falls outside the aperture formed between the image source and the surface generating it er The figure on the left illustrates the concept i of visible and hidden image sources If an 5 image is found to be visible then a g 9s reflection is added to the reflectogram Image sources are split into a specular Visible and invisible images Images S1 and S21 are visible from contribution and a scattering tree which R while S2 S3 and S12 are not consists of secondary sources on the image source surfaces allowing a realistic calculation of early scattering The attenuation of one particular Image source is calculated taking the following into account e Directivity factor of the primary source in the relevant direction of radiation 69 e Reflection coefficients of the walls involved in generating the image can be angle depende
165. g method is set to Lambert The scattering coefficient can be assigned values between 0 and 1 see section 6 5 Table 1 Warning Some acousticians call the scattering coefficient diffusion coefficient but this is wrong The diffusion coefficient has to do with the degree of diffusivity in the sound field i e in the space while the scattering coefficient determines the way sound is reflected from a surface High scattering coefficient can lead to high diffusion coefficient but the two values are not directly linked With the suggested scattering coefficients it is assumed that Diffraction surfaces and Oblique Lambert has been enabled in the Room Setup If this is not the case then a minimum scattering coefficient of 0 1 is suggested 0 3 may be more appropriate for disproportionate rooms such as class rooms If some details are not modelled in a room then the scattering coefficient may also need to be increased a coffered ceiling where the coffered cells have not been modelled may typically have a value of 0 3 to 0 4 for the mid frequencies around 700 Hz Transparency coefficient semi transparent surfaces A transparency coefficient is assigned to each surface this is a way to make the surface semi transparent This feature may be used for modelling many small surfaces in real rooms E g a reflector panel built from many small surfaces with space in between can be modelled as one large surface having a transparency coefficient of e g
166. ght side menu Energy parameters All calculated room acoustic parameters are displayed in a table of the same format as that for the single and multi point responses see section 2 1 Parameter curves The room acoustic parameters are displayed in bar graphs You can scroll among the different parameters by using the Left and Right arrow keys in the keyboard 109 Inserting measured room acoustics parameters into a multi point response If a room model has been loaded into ODEON then it is possible to insert the room acoustics parameters derived from a measurement file which has been opened with the Load Impulse Response Use the Add measured parameters INS shortcut or Add measured parameters and close Ctri INS shortcut available from the Measured response menu and specify appropriate job and receiver number in the dialog that appears Once parameters has been added you may view them in the Multi point response with the given job number and if the receiver number has been specified in the Source receiver list available from the JobList e g together with simulation results see section 2 1 Even if no simulations have been carried out you may benefit from the Multi point response display as it can provide results that related to multiple receivers e g statistics or results according the new ISO 3382 3 standard on open plan offices You can toggle between simulated and measured mode in the Multi point response by pressing the m
167. gth Sin PI 4 where Length is a user defined constant or variable Mathematical expressions may not contain any SPACE or TAB tabulation characters To get a complete overview of the mathematical functions available please refer to appendix F Points A point is made up from an unique point number and its X Y and Z coordinates Use the Pt MPt and CountPt statements to define points Points can also be defined implicitly using one of the hybrid statements Surfaces A surface is made up from a unique number an optional descriptive text and a number of points connected to one another All the points must be coplanar otherwise the surface cannot be constructed To define surfaces use the Surf MSurf CountSurf ElevSurf ElevSurf2 and RevSurf statements Surfaces may also be defined implicitly using the hybrid statements Hybrid statements Hybrid statements are Box Cylinder Cylinder2 Cone Dome Dome2 and ElevSurf The hybrid statements create the points and surfaces needed to model the specified shape The points and surfaces created must always have unique numbers Coordinate manipulation functions A set of functions for coordinate manipulation and surfaces made up from coordinates is included This includes rotation around the various axes scaling and translation These functions are needed in order to insert shapes defined by the hybrid statements in the geometry with the correct position and orientation Comments and empty l
168. h poor sound radiation Table 10 1 Coordinates and elevation angles of the 7 units in the array example in fig 11 3 Transducer Elevation 0 000 0 000 0 000 0 000 0 000 0 000 0 000 10 2 Playing with delay One advantage of the line array is the possibility to control the direction of the main lobe of sound by means of small phase shifts different for each of the units So instead of physically tilting the loudspeaker the line array can be mounted in a vertical position and still direct the sound towards the audience If the units all have the same distance d and the delay from one unit to the next is At the angle of sound radiation relative to the normal direction perpendicular to the array line is 0 arag AE d In fig 11 4 is shown an example with At 0 1 ms per unit and in fig 11 5 the same with At 0 2 ms per unit Note In order to turn the beam downwards the delays should be set from 0 ms in the upper unit to N 1 At ms in the lower unit e g 1 2 ms in the case of N 7 and At 0 2 ms 99 Figure 10 3 The radiation in octave bands from 250 Hz to 8 kHz for a line array with 7 units and a time delay 0 1 ms per unit ma He Figure 10 4 Same as fig 11 4 but with a time delay 0 2 ms per unit At the 2 kHz band and higher frequencies i e above the upper limiting frequency 1720 Hz in this example a rather strong side lobe of radiation is seen at various directions slightly upwards This is the ef
169. h the interior marin marked as a dashed green line Surfaces between the boundary and the interior margin provide low scattering at low frequencies while surfaces inside the interior margin provide high scattering throughout the whole frequency range The typical depth of geometry s wall construction should be specified in the Interior margin in the room setup ODEON will use this number in order to distinguish between interior and boundary surfaces Once the margin has been entered and the room setup dialog has been closed the 3Dview will display surfaces which are considered to be interior in a greenish colour while the exterior is displayed in black Diffraction from the exterior will be calculated taking into account that diffraction is limited towards the lowest frequencies because of limited depth of the wall constructions Limitations In special cases the Reflection Based Scattering Coefficient may overestimate the scattering provided by small surfaces which are only fractions of a bigger whole e g small surfaces being part of a curved wall or dome Such surfaces should not cause diffraction due to their individual area 76 because the individual surfaces do not provide any significant edge diffraction In these cases the method can be bypassed by setting the surface Type to Fractional in the Materials List see chapter 4 When setting the Type to Fractional the surface area used for calculating the Reflection Based Scattering Coeffic
170. he 3rd Round Robin on Room Acoustical Computer Simulation Part II Calculations Acta Acoustica United with Acoustica 91 753 763 Bradley J S 1986 Predictors of speech intelligibility in rooms J Acoust Soc Am 80 837 845 Bradley J S amp Soulodre G A 1995 Objective measures of listener envelopment J Acoust Soc Am 98 2590 2595 Christensen C L amp Rindel J H 2011 Diffusion in concert halls analysed as a function of time during the decay process Proceedings of the Institute of Acoustics Vol 33 Pt 2 pp 98 105 Dublin Christensen C L Nielsen G B amp Rindel J H 2008 Danish Acoustical Society Round Robin on room acoustic computer modelling Retrieved from Odeon A S http www odeon dk pdf Classroom 20RR pdf Cremer L amp M ller H 1982 Principles and Applications of Room Acoustics London Applied Science Publishers 138 Dietsch L amp Kraak W 1986 Ein objektives Kriterium zur Erfassung von Echostorungen bei Musik und Sprachdarbietungen in German Acustica 60 205 216 Dirac n d Retrieved June 16 2011 from Acoustics Engineering Dual Input Room Acoustics Calculator http www acoustics engineering com dirac dirac htm Fasold W amp Winkler H 1976 Bauphysikalische Entwurfslehre Band 4 Bauakustik in German Berlin Verlag f r Bauwesen Furse R n d 3D Audio Links and Information Retrieved June 16 2011 from http www muse demon co
171. he General settings that may also need attention the Impulse response length and Number of late rays The Specialist settings need normally not be adjusted Impulse response length Job calculations and Global Estimate Always specify the Impulse response length It should be at least 2 3 of the reverberation time the longest reverberation time over all 1 1 octaves It determines how many milliseconds of decay curve are to be calculated It is a key parameter if it is shorter than app 2 3 of the reverberation time in the room then Tso cannot be calculated because the dynamic range of the decay curve is less than 35 dB then it will be displayed as in result displays For reliable result it is recommended to use an Impulse response length which is comparable to the estimated reverberation time To get an estimate of the reverberation time in a room use the Quick Estimate available in the Materials list The maximum allowed Impulse response length is 32000 ms Number of late rays Job calculations The number of Late rays determines density of reflections in the late part of the decay If the predicted decay curves Decay curves are displayed in the Single point response results in ODEON Auditorium and Combined has suspicious spikes you may try to increase the number of late rays accordingly The actual number of reflections per millisecond used in the calculation of a point response is included in the calculated parameters of the Si
172. he Material List and increase the absorption of the ceiling so that the simulated T30 becomes lower and approaches the measured one better Genetic Algorithm Optimization Instead of following the tedious manual optimization procedure described above you can make use of the ODEON s new feature called Genetic material optimizer The tool makes use of Genetic Algorithms and tries to modify the absorption coefficients of selected materials in the room in order to achieve a good agreement between simulations and measurements This optimization is done on the basis of acoustic parameters The user can determine which acoustic parameters should be involved in the process by changing their visibility in the Room Acoustic Parameter list Ey Before performing any calculation it is worth getting familiar with the basic principles of Genetic Algorithms 12 2 How Genetic Algorithms Work Introduction Genetic algorithms GA are widely used for optimization processes in diverse areas such as industrial design artificial life systems and economics GA start with an ensemble of individuals chromosomes and evolve new and improved individuals by applying principles found in molecular genetics and biology crossover recombination mutation etc In any stage of the evolution the ensemble of individuals is called population and corresponds to one generation An individual is essentially a candidate solution to the optimization problem and normally consists of
173. he keyboard e Repeat Scattering coefficient assigns the scattering coefficient last entered in the surface list to the current selected surface e Assigning Transparency coefficients select field at the surface and enter the transparency coefficient directly using the keyboard e Repeat Transparency coefficients assigns the transparency coefficient last entered in the surface list to the current selected surface 53 e Assigning wall Type allowing either calculation of Transmission through walls or a Normal Exterior or Fractional surface property Select the field at the surface and enter the wall type using the mouse e Repeat wall Type Select surfaces for which you want to repeat wall type by holding down shift and use the arrows and repeat wall type with ctrl V e Quick Estimate for fast evaluation of reverberation times and listing of summarized absorption areas while assigning materials etc e Add Edit Delete a material Four buttons allow you to create new materials or to edit existing ones in the material library The material editor available assists in mixing different materials into one It is also possible to edit materials directly in the material library file e Find surface or Material type Three buttons makes it easier to find a certain surface in the material list or material in the material library The surface search material button for finding material in the library and the toggle interior exterior and the layer butt
174. he rays emitted from these source types generate an independent secondary source each time they are reflected Compared to the calculation principle applied to the point sources one might say that transition order is always zero and that only late energy contributions are collected for these source types or rather that calculations are based on a sort of ray tracing Vector Based Scattering reflecting a Late ray Vector based scattering is an efficient way to include scattering in a ray tracing algorithm The direction of a reflected ray is calculated by adding the specular vector according to Snell s law scaled by a factor 1 s to a scattered vector random direction following the angular Lambert distribution of ideal scattered reflections Sin2 Rindel 1995 which has been scaled by a factor s where Sis the scattering coefficient If sr is zero the ray is reflected in the specular direction if it equals 1 then the ray is reflected in a completely random direction Often the resulting scatter coefficient may be in the range of say 5 to 20 and in this case rays will be reflected in directions which differ just slightly from the specular one but this is enough to avoid artefacts due to simple geometrical reflection pattern Incident KS i q Specular weight 1 s KA Z Resulting A _ e i Pp Scattered weight s Vector based scattering Reflecting a ray from a surface with a scattering coefficient of 50 re
175. he rings are created clock or counter clockwise The surface is created from the following list of points 100 101 102 110 111 112 200 201 202 210 211 212 DonutSurface par HHH Const RI 10 Const R2 15 Const N 12 CountPt 100 N I R1 CosD 360 PtCounter N R1 SinD 360 PtCounter N 0 CountPt 200 N 1 R2 CosD 360 PtCounter N R2 SinD 360 PtCounter N 0 Surf 100 Donut surface 100 gt 100 N 200 gt 200 N HHH The window example shows how a cylindrical window opening is created in ceiling surface The interesting surface in this example is surface 1 the ceiling surface The surface is created from the following list of points 1 100 101 102 103 111 112 1 2 3 4 186 HHH Ceiling With WindowTube par HHH Const R1 0 75 Const R2 0 5 Const N 12 Pt1110 Pt2 1 10 Pt 3 1 1 0 Pt 4 110 CountPt 100 N 1 RI CosD 360 PtCounter N R1 SinD 360 PtCounter N 0 CountPt 200 N 1 R2 CosD 360 PtCounter N R2 SinD 360 PtCounter N 2 Surf 1 Ceiling 1 100 gt 100 N 1 gt 4 RevSurf 2 100 200 N Window tube Surf 100 Window glass 200 gt 200 N 1 187 Appendix E Mathematical expressions Here mathematical expressions used in the Par modelling format and in the Room acoustic parameter list are presented Constants variables point numbers surface numbers and coordinates may be defined using mathematical expressions Where integer numbers are expected Counter ranges in for end loops point surface numbers etc the res
176. hich might then act like a reflector Concave curves naturally focus sound energy and since focussing is a fault we wish to model we must try to arrange so that it is preserved However this does not mean that a large number of subdivisions is the solution Using many surfaces in the model will e make the model visually complex and increase the probability of errors in the model typically small leaks may become a problem e not comply with the image source theory used for the early reflections point sources See more about calculation principles in Chapter 6 e increase the calculation time In order to calculate focusing from concave surfaces the wall type of surfaces forming a concave shape should be set to fractional in the Materials List otherwise the concave surface will scatter sound too much taking into account the small areas of the individual surfaces forming the concave shape rather than the total area of the concave shape The Reflection based scattering method would produce too much scattering in this case Remember Subdivisions about every 10 to 30 will probably be adequate to reproduce focussing trends of concave surfaces without excessive number of surfaces thus walls in a cylindrical room may be modelled from 12 to 36 surfaces A cylindrical column which disperses energy may probably be modelled from say 6 to 8 surfaces What to model Some simple tricks during modelling will save you a lot of time improve
177. i 4 vi Orchestra Mozart MOZART va 21 _ 7 3 4700 Average 5 Ltowards P46 Voi 0 00 __ 14 20 E 0 00 000 miwon MOTETO gj 5 Vi Orchestra Mozart MOZART vc 21 3 4700 Average 6 ltowardsP46 Cello 0 00 7 99 7 4 0 00 0 00 Vols MOZART via Pv 6 Vi Orchestra Mozart MOZART db 21 3 4700 Average 7 1towadsp46 Double bass 0 00 7 91 3 s s 0 00 0 00 Celo MOZART vc 4 7 v Orchestra Mozat MOZART fl 21 3 4700 Average 8 itowardsP46 Fute 0 00 i795 7 6 6 0 00 0 00 Double bass MOZART db 8 vi Orchestra Mozat MOZART d 21 3 4700 Average 9 1towadsP46 Cornet 0 00 _ 13 59 y 7 7 0 00 0 00 Flute mozart 9 x Orchestra Mozart MOZART bmn 21 __ 3 4700 Average 10 1towadsP46 Bassoon 0 00 9 44 s 0 00 0 00 Clarinet 10 v Orchestra Mozart MOZART comol_21__ 3 4700 Average 11 1towardsP46 1st Fr hom 0 00 1I 7 ss 0 00 0 00 Bassoon MOZART bsn 11 Vi Orchesta Mozart MOZART comol 2 3 4700 Average 12 1 towards P46 2nd Fr hom 0 00 __ 12 92 j 10 10 0 00 0 00 1t Fr hom MOZART com 12 v Orchestra Mozart MOZART sopr 21 3 4700 Avenge 13 1towardsP46 Soprno pos 2 0 00 35 64 7 uj un 0 00 0 00 2nd Fr hom MOZART com 13 v Orchestra Mozart MOZART sopr 21 34700 Avenge 14 itowadsP46 Soprano pos 3 0 00 33 02 12 none 0 00 0 00 m x none Average none 0 00 99 00 y 13 none 0 00 0 00 15 v none Average none 0 00 99 00 14 none 0 00 0 00 16 none Average none 0 00 99 00 7 15 none 0 00 0 00 z g none Average none 0 00 99 00 7 16 none 0 00 0 00 am Pa
178. ich should be presented over headphones the objective being to reproduce the same sound pressure at the entrance of the ear canals and at the eardrums for that matter of the subject as would be obtained in the real room if it exists Soundcards A sound card is required in order to play back the auralisation results and may also be useful if you wish to transfer anechoic signals to the hard disk As a minimum the sound card should be capable of handling signals in stereo in a 16 bit resolution at a sampling frequency of 44100 Hz To transfer signals without loss in quality to from a DAT recorder the soundcard should be equipped with digital input and output and the soundcard should be able to handle a sampling frequency of 48000 Hz It should also be considered whether the card is immune to electromagnetic noise which is always present in a PC and whether its analogue output for headphones is satisfactory For surround auralisation obviously a multi channel surround soundcard is needed along with the necessary loudspeakers and amplifiers Input signals for auralisation anechoic recordings For auralisation you will be using input signals to convolve with the calculated BRIR s Usually the signals will be anechoic signals although it may also be other types of signals e g if you are simulating an ordinary stereo setup in a room you will probably be using a commercial stereo recording The input signals to be used with ODEON are stored i
179. idea whether a room has strong decoupling effects you may try to run the Global Estimate calculation If e Global Estimate stabilizes slowly e The Global decay curve make sudden jumps like steps on a stair e The Global decay show hanging curve effect this could be an indication that more rays are needed Let the Global Estimate run until the decay curve seems stable then use say 1 10 1 times the number of rays used in the Global estimate to specify the number of rays to be used in the calculation of the point responses specified in the room setup Transition order The Transition order applies only to point sources Current recommendation is TO 2 for most rooms For rooms that are heavily packed with various fittings or rooms where no or few image sources are visible from the receivers a lower Transition order of 0 or 1 may be used Materials absorption data Wrong or imprecise absorption data are probably one of the most common sources of error in room acoustic simulations This may be due to lack of precession in the measurements or limitations to the measurement method itself of the absorption data or because the material construction assumed in the simulations are really based on guesswork in any case it is a good idea to remember this and to estimate the size of error on the material data as well as the impact on the simulated results It has been seen that absorption coefficients outside the range 0 05 to 0 9 shou
180. idth gt lt Height gt lt RevAngle gt lt optional name gt lt Number gt A unique number from 1 to 2 147 483647 for identification of the first point and surface in the Cone Using the same number but with negative sign defines the surface and its mirrored counterpart in the XZ pane Y 0 A Cone will take up several point and surface numbers which must all be unique lt NumberOfSurface gt Specifies the number of surfaces in one horizontal ring of the dome around 16 to 24 surfaces per ring is suggested ODEON will automatically calculate the number of subdivisions in the vertical level If the revolution angle is 180 the number is stored in the ONVert variable would have been 9 in the example above The ONVert variable may help when connecting a Dome2 to a Cylinder2 in order to specify the correct number of surfaces in the cylinder lt Width gt Width at the beginning of the dome The width must always be greater than zero lt Height gt The Height must be different from zero If the height is less than zero the orientation of the dome is inverted Height must be different from zero and less or equal to 42 Width lt Revangle gt Revangle must be within the range 360 and different from zero If RevAngle is 180 a half cone is generated if its 360 a full cone is generated Positive revolution angles are defined counter clockwise Connection points The right side vertical points in Dome2 are stored in PlistA The l
181. ied or until the user terminates the calculation It is unlikely that the solution will converge fully in the acoustic optimization problem so in practice calculations are terminated by the user Target Value and Fitness Function In order to evaluate how well a room simulation model with a given set of materials an individual matches against the measured room one needs a fitness function see Table 1 The fitness function returns a number fitness to the GA that allows it to determine which individual 128 candidate solution is better than others controlling the genetic evolution In our problem the GA seeks for individuals that minimize the fitness value while in other GA problems the criterion might be a maximization of the fitness In principle it can be considered to evaluate directly how well the simulated impulse responses match against the measured ones However there is bound to be differences between them as neither measurements nor simulations are perfect Christensen C L 2014 Instead we have chosen to compare how well some of the room acoustic parameters match as these are supposed to be good measures for important attributes of the acoustics in a room Indeed a group of room acoustic parameters is one of the most important tools of the room acoustician for the evaluation of the acoustic quality In order to evaluate the fitness of a set of materials point responses for a number of source receiver pairs are simulated and
182. ient is determined from the box subscribing the room rather than the individual surface if the construction part which the fractional surface is part of is considerable smaller that the room box scattering might be underestimated and a higher scattering coefficient should be assigned to the surface A typical diffuser depth interior margin entered by user determines which parts of the geometry should be considered exterior of a room thus limiting the scattering of the boundary surfaces which can not be considered freely suspended at the lowest frequencies 6 6 Oblique Lambert In the ray tracing process a number if secondary sources are generated at the collision points between walls and the rays traced It has not been covered yet which directivity to assign to these sources A straight away solution which is the one ODEON has been applying until version 8 is to assign Lambert directivity patterns that is the cosine directivity which is a model for diffuse area radiation However the result would be that the last reflection from the secondary sources to the actual receiver point is handled with 100 scattering no matter actual scattering properties for the reflection This is not the optimum solution in fact when it comes to the last reflection path from wall to receiver we know not only the incident path length to the wall also the path length from the wall to the receiver is available allowing a better estimate of the characteris
183. ile 16 bit PCM ha Binaural settings Parameters for B Format F Create binaural impulse response file Jnn aa 4000 lt eB dB HRTF Subject_ 21Res1Odeq_M3 0_SRate44100_Apass l50 Astop40 00 BOwLapl00 PPHRTF258 Pas O90 ip Band overlap 100 amp Headphone Subject_021Res1Odeg_diffuse way Sample rate 44100 Hz Low cut filter 10 Hz T Low cut filter 10 Hz Overall Recording level 0 00 dB Encoding Phase approximation phase shift at surfaces tilterphase 1 order ambisonics _B Format Create impulse response Overall recording level 0 00 dE 20 Surround sound C Create impulse response SurFioundnn Overall recording level 0 00 dE Compensate speaker delays Parametrization 20 l Define speaker rig l Use program defaults settings Make surround settings program defaults Save speaker ig and make archive file Get speaker rig from file The auralisation setup in ODEON The left hand side refers to headphone binaural reproduction while the right hand side refers to reproduction over a speaker rig Make sure the Create binaural impulse response file Jnn option is activated when you want ODEON to derive BRIR s for auralisation Similarly make sure the Create impulse response SuRoundnn is activated for a surround playback A note on directivity patterns for natural point sources With natural sources we refer to sources such as human voice an acoustical instrument or similar
184. ile and re importing the exported file 29 e Export the geometry into a 3D Studio file using the 3DSOUT command this does not change your current CAD drawing e Import the 3DStudio file just created back into a new clean drawing in AutoCAD using the 3DSIN command In this new drawing the above entities has been converted to Polyface entities which are supported directly by ODEON At the same time all entities contained in BLOCK s have been exploded making them appear explicit thus directly compatible with ODEON If ODEON reports of any of the unsupported entities when the dxf file has been imported this is because some 3D data is available in the pxF file in a format which cannot be converted by ODEON Consider following the steps above in order to create a dxf file which can be converted by ODEON Do note that Acts solid modelling extensions may not be available in all editions of the various modelling programs Using LAYER s in the CAD drawing Surfaces will when imported to ODEON carry the name of the layer on which they were drawn Use the layer name to give the different parts of the geometry different names e g draw the stage floor surfaces on a layer named Stage floor the sidewalls on a layer named Sidewalls etc If you are modelling subdivided surfaces such as Upper wall and Lower wall because you wish to be able to assign different materials to these parts of a surface it is advisable to model these parts on different la
185. in better calculation results for the room acoustical parameters In such cases create an extra copy of the room using the Files Copy files menu and conduct the special investigations on the copy The Transition order applies only to point sources Current recommendation is TO 2 for most rooms For rooms that are heavily packed with various fittings or rooms where no or few image sources are visible from the receivers a lower Transition order of 0 or 1 may be used If a Transition order of zero is selected then point responses will be calculated using ray tracing ray radiosity only which has better robustness to cases where the wave front is cut into small pieces e g by apertures in the room a situation which is not handled elegantly by the Image source method Number of early rays Job calculations only The Number of early rays is normally best left for ODEON to decide ODEON will estimate a Number of early rays to use in order to detect important early reflections Image sources based on the number and size of surfaces the dimensions of the room as well as the user defined Transition order It is not given that more is better 89 Number of early scatter rays Job calculations only If the specified Number of early scatter rays is greater than 0 the early reflections from point sources are split into a specular part an image source and a scattered part given by the scattering coefficients 100 scatter rays per image source shoul
186. ine contains 19 values 0 10 20 160 170 and 180 Lines containing comments and empty lines may be inserted anywhere in the ASCII input file as long as they do not come between data items which should occur on one line Comment lines must begin with a colon a semicolon or an asterisk When only horizontal and vertical polar plots are known POLAR The first non comment line of the file should start with the word POLAR In the polar case there are four lines of data for each frequency band The first four lines are for 63 Hz the next four for 125 Hz and so on For a given frequency the first and last values must agree on all four lines since all the polar plots meet at the front and back polar axe The first line of a group of four is the upward vertical polar plot as seen from in front of the source 12 o clock plot Then comes the left horizontal plot 9 o clock plot downward vertical plot 6 o clock plot and finally the right horizontal plot 3 o clock plot As a minimum there must be 1 4 8 lines in a polar input file Elliptical interpolation When ODEON translates the polar input file it has to interpolate values between the four polar planes given in the input data This is done using elliptical interpolation independently for each frequency band creating the 8 x 4 plots missing between the four input plots An example Polar_Omnidat on the polar input format can be found in the DirFiles So
187. ines Lines containing comments and empty lines may be inserted anywhere in the file as long as they do not come between data items which should occur on one line Comment lines must begin with a colon a semicolon a slash or an asterisk The semicolon can also terminate a non comment line allowing the non comment line to be followed by a comment A series of one or more comment lines are started with a and terminated by a Y at any position on a line This is useful to disable a section of lines 157 Syntax highlighting in the ODEON editor makes it easy to spot comment lines Reserved keywords predefined counters and constants The following keywords are reserved by ODEON and have a special meaning in the parametric modelling language Line folding markers BeginBlock EndBlock Constant and variable statements Const Var Point statements Pt MPt CountPt Point list statements PlistO Plist9 ResetPListO ResetPList9 PlistA PListB Surface statements Surf MSurf RevSurf CountSurf ElevSurf ElevSurf2 Hybrid statements Box Cylinder Cylinder2 Cone Dome Dome2 Loop statements For End Transformation statements Mreset MPop MScale MTranslate MRotateX MRotateY MRotateZ and for compatibility with earlier releases of ODEON Scale UCS Predefined constants PI 3 14159265358979312 Predefined variables NumbOffSet ONVert Predefined Counters PtCounter Coordinate syste
188. inimum phase filter with a frequency response inverse to that of the headphone A number of headphone filters are supplied with ODEON in ODEON s hph format Filters ending on ee hph are measured on a dummy head at the entrance of a blocked ear canal whereas filters ending on ec hph are measured at the end of the ear canal at the ear drum so to speak The selected headphone filter should match the HRTF used for example the Subject_021Res10deg hrtf doesn t have an ear canal therefore an ee filter should be used whereas the Kemar hrtf does have an ear canal so the ec filters should be used If the corresponding filter for the headphone used is not available then a generic filter matching the set of HRTF s may be used e g Subject_021Res10deg_diffuse hph If a diffuse field filter equalization is selected the results are filtered in order to obtain an overall flat frequency response of the HRTF s that is the average frequency response of all the HRTF filters is calculated and the auralisation results are filtered with the inverse of that If using headphones which are diffuse field equalized most headphones attempt to be and a matching headphone filter is not available then the matching diffuse filter headphone filter can be used For your convenience two directories with matching HRTF s and headphone filters are installed with ODEON namely the HeadsAndPhones EE and HeadsAndPhones EC directories The first directory contains a set of
189. inition for each of the corners referred to in the corner list of the CountSurf statement In the above example the points 1000 1004 1100 1104 1200 1204 and 1300 1304 need to be defined Sample room files Beams Par BeamBox Par BeamBoxWithWindows Par Revolution surface RevSurf RevSurf must follow the syntax RevSurf lt FirstSurfaceNumber gt lt CurveStartl gt lt CurveStart2 gt lt SectionsInRevSurf gt lt Optional name gt The RevSurf command is typical used together with two CountPt statements to create a revolution surface using two curves of points The curves must contain the same number of points The RevSurf command will always create a number of surfaces each build from four points lt FirstSurfaceNumber gt A unique number from 1 to 2 147 483647 for identification of first surface in the revolution surface Using the same number but with negative sign defines the surface and its counter part mirrored in the XZ plane Y 0 lt CurveStartl gt First point number in the first revolution curve The curve of points is typically created using the CountPt statement and the curve must contain one more point than number of sections in the RevSurf lt CurveStart2 gt First point number in the second revolution curve The curve of points is typically created using the CountPt statement and the curve must contain one more point than the number of sections in the RevSurf The second curve must always contain the same n
190. io tracks will always be in two channels if you know that signals in fact are mono signals it will be a good idea to convert the resulting wave files into mono signals this will save space on the hard disk and avoid confusion whether a signal is stereo or in fact mono A standard wave file editing software which is usually included with the soundcard should be capable of doing the job If you have recordings which you have created yourself e g using a DAT recorder you should use a wave file recording software in connection with your soundcard in order to transform the recordings into wave files Most soundcards comes with a software program for recording and editing wave files which should be capable of this job Please note that the connection between the CD ROM drive and you soundcard is often an analogue one so if you record from this drive you ll not benefit from digital inputs on your soundcard resulting in a loss in quality Output signals The output signals from all binaural auralisation are stored in two channel wave files and will have the same leading name as the room The result files being in the wave format makes it easy to edit and publish the results e g on the Internet or on audio CD s The binaural impulse responses files have the extension JInn Wav where nn refers to the relevant job number The wave files created as results from the Convolve BRIR and Signal file table will have the extension ConvAuralnn Wav where
191. ion research and teaching only 26 This extra line is justified by the fact that each surface in ODEON needs to be formed as a series of points towards a fixed direction clockwise or counter clockwise As a result when placing a window in the middle of a wall an extra line appears as a natural reaction by ODEON to help continouing the series of points The line is not part of the actual geometry and it is not a fault It is just an indication of a healthy geometry As can be seen in the next figure 11 points form surface No 4 highlighted by red color The line formed by points 9 and 10 serves as a connector between the points at the main wall circumference and the points at the window circumference You can get this display by opening the 3DView and pressing the N key shortcut File Toolbar 3D_View Options Tools Window Help Remember Always check the dynamic menu e amp E a aes that appears at the menu bar between the Bes Right Alt Right Toolbar and the Options menus for all functions 2 Up Alt Up available in the window you are working Down Alt Down with For the picture on the right the 3DView Rotate Left Ctri Left window is active and therefore the 3D_View ica ach Rotate Up Ctrl Up menu appears at the menu bar l Rotate Down Ctrl Down 3 3D View Interior Exterior mode L Talg orners in surface no 4 yz 8 470 10 596 3 770 6 470 14 240 3 770 6 470 14 240 0 000 6 470 0 000 0 000 6 470 0 000 3 770
192. ioritized in the order listed below In these cases it recommended to perform a visual investigation of the impulse response remember that there is more than one octave band to look at and check that e itis without significant distortion products e impulse squared impulse responses look healthy i e without suspicious bumps e onset and truncation has been estimated correctly e acoustic parameters of interest were derived The warning messages concern frequencies between 125 and 4000 which are the most important according to the ISO standard 3382 ISO 3382 1 2009 Impulse response with well defined noise floor The following figure displays the raw decay curve for a typical impulse response filtered at 1000 Hz The particular impulse response has a well detected noise floor solid blue line with well defined onset pink line and truncation times red line In the ODEON measuring system the noise floor is automatically calculated individually for each octave band Subsequently a truncation time is automatically derived If another truncation time is needed you can do so manually following the instructions in section 11 5 be Decay range less than 45 dB 41 46 dB at 4000 Hz Measured response D ODEON 13 PicturesFromImpulseResponses IR_BasementCropped wav gE ll sz Raw Impulse response at 1000Hz Raw decay curve at 1000Hz Decay curves at 1000Hz T 30 1 46 seconds Decay curves all bands Energy parameters Param
193. ith different surface materials on either side When the Transmission type is assigned to a surface 90 of the rays hitting the surface will be reflected and 10 will be transmitted the energy calculations are accomplished by multiplying each ray or particle with the appropriate frequency dependent energy parameters Transmission is covered in appendix C Surface name Lists the name given to the surface in the surface file if given any name Area Lists the calculated area for each surface 4 3 Manage material library and material list Material Toolbar Some of the functions available at the local toolbar as well as from the toolbar dropdown menu allows you to manage the global and local material library and surface properties in the surface list AZOBDA EAM ATAT E AGE KO e Assign Material assigns the material selected in the material list to the surface s selected in the surface list e Global Replacement replaces all appearances of the material assigned to the selected surface in the surface list with the material selected in the material list This is useful if you wish to replace all materials of one type with another type e Assign Material for all surfaces assigns the material selected in the material list to all the surfaces in the room or in the selected layer e Assigning Scattering coefficient is done a little different You simply select the field at the surface and enter the scattering coefficient using t
194. k So8 file as natural directivity When creating new directivity patterns this information is part of the input data Always use the _Natural versions of the directivity files when defining new natural point sources The old versions of the files are kept in Old_So8 a subdirectory to the Dirfiles directory If you wish to use the old directivities in old existing projects then open the Source receiver list and click the Repair broken directivity links button shortcut Ctrl L Samples on natural directivity patterns TLKNORM TLKRAISE and Soprano ref 42 The TLKNORM source type corresponds to a male talker with a normal vocal effort The gain and EQ fields in the Point source editor inside ODEON should be set to zero This source is also a reasonable approximation to a female talker except that the 63 and 125 Hz band should be ignored To simulate a trained talker addressing an audience in a raised voice use the TLKRAISE source This has the same directivity as TLKNORM but the levels in the eight octave bands are respectively 2 2 5 7 9 8 6 and 6 dB higher The directivity pattern of Soprano ref 42 is the directivity of a soprano singing opera Parati amp Otondo 2003 3 3 Common loudspeaker format CF1 and CF2 files ODEON 8 and later supports the Common Loudspeaker Format which is an open format for loudspeaker data supported by several loudspeaker manufacturers as well as manufactures of software programs such as ODEON The C
195. key Figure 11 0 2 shows an example of measured and simulated Tso in the Multi point response s Multi point response parameters job 1 Simulated mode zobe Energy parameter curves 1 Energy parameter curves 2 Parameter versus distance Statistics Spatial decay curves STI versus distances Noise control Energy parameters Measured versus Simulated T 30 at 8000 Hz Receiver 2 E Simulated gt lt Measured T 30 s T 30 s E 9 A ac m id R2 at 6 84 m R5 at 8 95 m Q O its R1 at10 12m R4 at 12 76 m 6 125 250 2000 4000 8000 o Distance Frequency Hertz Figure 11 0 2 Simulated and measured reverberation time T3 inside the Auditorium 21 at the Technical University of Denmark The graph on the left shows the values at all receivers for a given octave band while the graph on the right shows the values at all octave bands for a given receiver Press the Up and Down arrows to change frequency Press R to change receiver Calibrate Measurements So far recording and processing of impulse responses have been described without calibration taken into account In order to calculate the Sound Strength the system set up should be calibrated Two calibration methods are available from the Tools gt Calibrate measurements Menu in ODEON that follow the ISO 3382 1 standard Diffuse Field Calibration This should be carried out in a room with long reverberation time and diffuse fi
196. l results of the Multi Point calculation DL is given for the frequency bands 63 Hz to 8 kHz and DLzco is the A weighted Rate of Spatial Decay for the frequency bands 125 Hz to 4 kHz For DLz as well as DLzco the correlation coefficients are calculated If the correlation coefficients are low this may indicate bad locations of source and or receivers however it may also indicate a very low damping in the room the Spatial Decay Curve being almost horizontal The measuring points Receiver points and the source position are of course essential to the DL2 parameters and should follow ISO 14257 2001 As an example a path of receivers may be chosen in the following distances from the source using logarithmic increment 1 2 4 5 6 3 8 10 metres The positions should also follow the standard with respect to distance from floor and reflecting surfaces ODEON will use all the receivers defined in the receiver list In some cases the positions of the receivers will not combine with the receiver positions that should be used for the receiver path in the DL calculation In this case the following solution is recommended e Make a copy of the room using the File Copy files option e g copy a room called MyRoom to MyRoomDL2Path and load the new copy when prompted for during the copy process e Delete receivers that are not wanted in the receiver path e Define the receivers needed e Finally make the Multi Point response calculation with
197. ld be used with care Christensen Nielsen amp Rindel 2008 In the Material list in Odeon there is a button which will limit the range of absorption coefficients assigned to surfaces in a room to a selected range Solution if materials data are uncertain There is really not much to do about the uncertainty of material data if the room does not exist except taking the uncertainty of the materials into account in the design phase If the room does indeed exist and is being modelled in order to evaluate different possible changes it may be a good idea to tweak adjust uncertain materials until the simulated room acoustical parameters fits the measured ones as good as possible Absorption properties in a material library are often by users assumed to be without errors This is far from being the truth For high absorption coefficients and at high frequencies the values are probably quite reliable However low and even mid frequency frequency absorption data and absorption data for hard materials will often have a lack of precision Low frequency absorption At low frequencies the absorption coefficients measured in a reverberation chamber are with limited precision because e There are very few modes available in a reverberation chamber at lowest frequency bands e Low frequency absorption occurs partly due to the construction itself rather than its visible surface structure Often it may not be possible to reconstruct a complete building c
198. le and in the Mix convolved wave results into one wave file table in order to obtain the highest dynamic range If using the Streaming convolution option available from the main display in the Joblist ODEON will maximize the auralisation output level if changing input signal or BRIR from within this display you may press the Maximize Gain button to maximize the gain for the new setup Relative play back levels In some cases you ll be interested in obtaining correct relative levels e g for comparisons between different seats in a concert hall In this case you should remember to use the same recording level convolver level and mixer level in the samples to be compared it is a good idea to use the same input Signal file to make sure that levels are the same at this point If you wish to compare across different rooms you should also be careful to remember that source gains in the rooms corresponds If using the Streaming convolution option available from the main display in the Joblist ODEON will maximize the auralisation output level so if you wish to compare different setups you should make sure to set the Gain in the Streaming convolution display to the same value Absolute play back levels for headphone auralisation Setting the level to an absolute level so the subject presented to the auralisation sample experiences the same level as would have been the case in the real room is a bit tricky as it involves every part in the signal chain
199. lection based on surface size and absorption coefficients of the last reflecting surface Rindel 1993 Then the reflection is filtered convolved through 9 octave band filters Kaiser Bessel filters the ninth being extrapolated and finally the reflection is filtered convolved through two corresponding 80 directional filters one for each ear Head Related Transfer Functions creating a binaural impulse response for that reflection This process is carried out for each reflection received at the receiver point and superposing all the reflections a resulting Binaural Room Impulse Response BRIR for that particular receiver point is obtained The actual order in which the filtering is carried out in ODEON differs somewhat from the description above otherwise the calculation time would be astronomic but the resulting BRIR contains the full filtering with respect to octave band filtering in nine bands as well as directional filtering B format filters for external decoding and Surround filters for loudspeaker playback Calculation of these impulse responses are based on the Ambisonics technique which is covered in Gerzon 1992 Bamford 1995 Malham 2005 Furse Most of the ODEON specific steps involved in the generation of these filters are similar to those used for generation of BRIR s there are however two differences HRRF s are not used in this calculation which is aimed at loudspeaker representation where the listener will receive r
200. ll try to detect the shortest diffraction path from source to receiver in the following way 1 A ray is sent from source towards the receiver and the first surface hit is registered A 2 A ray is sent from receiver towards the source and the first surface hit is registered B A diffraction path should follow the path Source An Bm Receiver where An and Bm are the n and mt point on an edge of each of the surfaces The shortest path with full visibility between the three sub paths Source An An Bm and Bm Receiver is used for the further calculations In cases where there is another surface which blocks the paths between one of the sub paths ODEON will right or wrongly not detect a diffraction path n An B m An Bm An Bm R R R i R S S S S a b c d Diffraction paths and their detection a Odeon detects a 1 point diffraction path over a thin screen b Odeon detects a 2 point diffraction path over a wide barrier C Odeon fails to detect a diffraction path because the path from A B is obscured by a third surface in fact it s a 3 point diffraction path d Odeon detects a diffraction path over two thin barriers contribution from diffracted sound over a screen around a corner or around a book shelf is not likely to change results noticeably however in rooms such as an open plan office with very absorbing ceiling or in outdoor situations diffraction around edges may play an important role 6 8 Radiated rays fr
201. lly only the energy of the direct sound should remain after the truncation Unfortunately the truncation is quite approximate at low frequencies because the direct sound and the subsequent reflections overlap to each other due to the long wavelengths of the sound waves and the presence of phase shifts It is recommended that more than six source receiver combinations should be used for the free field calibration since it is very likely that the source does not produce a perfect omni directional pattern but a pattern full of lobes This means that in contrast to the diffuse field calibration the free field calibration is very sensitive to receiver positions around the source Similarly to the diffuse calibration you should select and load all impulse responses obtained for the same calibration at once so that ODEON will calibrate the level according to the average of the values Once a calibration has been performed it is important that any external level adjustment not set inside ODEON is set to same level during measurement as was the case during calibration This includes levels set for play back level if set to 100 it s easy to remember recording level if set to 100 it s easy to remember and microphone boost set inside Windows as well as gain factors set on external devices used in the measurements Or alternatively that the factors are corrected using the External adjustments gt Gain in the Program setup Measurement setup to compensa
202. lly the list is collapsed showing the active parameters visible ones To change the state of each parameter click the Expand tables button at the top right corner The window becomes like the one in the subsequent figure The visibility column appears where you can select which parameters will be displayed in the single multipoint response multi point response and grid response The selected parameters will be taken into consideration in the genetic material optimizer as well Click here to expand the list a EDT Reverberatio s 0 00 2 50 Reverberation time initial 10 dB x 2 T_15 Reverberatio s 0 00 2 50 Reverberation time 15 dB amp 3 T_20 Reverberatio s 0 00 2 50 Reverberation time 20 dB B 4 T_30 T 30 S 0 00 2 50 Reverberation time 30 dB E 6 Ts Centre time ms 0 00 200 00 Centre time or in other words gravity time in milliseconds 10 D_50 Formula 0 00 1 00 Definition 13 C_80 Formula dB 10 00 10 00 Clarity Music Type specific data for reverberation time EDT Start 0o dB Display XI parameter ISO 3382 2 B 2 Stop wO dB Room acoustic wide band parameters _ Number Name Type Unit Manual Min Grid Max Grid Measured Description O Z o Formula dB 10 00 10 00 Aweighted SPL 63 8000 Hz E Formula dB 10 00 10 00 Linear SPL 63 8000 Hz Formula dB 10 00 10 00 C weighted SPL 63 8000 Hz Formula dB 10 00 10 00 A weighted direct sound presure level 63 8000 Hz STI 0 00 1 00 Speech Transmision Index specfically S
203. loudspeaker axis is positive Nodes Vector Vector node Attributes X 0 993768018134063 Y 0 0694910287794796 Z 0 0871557402186753 155 Appendix D Modelling rooms in the ODEON Editor The ODEON Par modeling format language Geometry models can be made using the parametric modelling language which is built in to ODEON The model data are typed into a text file given the file extension Par using the modelling language described below You may use the supplied editor ODEONEdit to create and edit your text files The ODEON modelling format is not case sensitive so upper and lower case letters can be used as desired A simple modelling example At its simplest but not fastest a floor with the dimensions 4 x 4 metres can be defined as follows using the reserved keywords Pt and Surf in order to define points and surfaces FloorSurface Par HHH Pt Pt Pt Pt Surf 1 HHH NNR WHYS WZS A AOS AA SS esses A One may choose to model the room point by point and surface by surface as in the example above however for many geometries it will be an advantage to use parameters to describe basic dimensions in the rooms and to use high level statements to describe multiple points and surfaces in a fast and flexible way Before starting your first large modelling project it is a very good idea to read through chapter 3 or at least skim it it will pay off in the end Another way to learn about the modelling language is to s
204. m definition statements Unit CoordSys Debugging Facilities DebugIsOn Debug Line folding markers Line folding is a feature of the ODEON editor where a section of lines e g a part of the geometry can be collapsed folded into one line in the editor for a better overview This is the only functionality of the two keywords BlockBegin BlockEnd they are ignored by ODEON when a _par file is loaded The keywords are automatically generated when a room is generated by ODEONExtrusionModeller or when a room is imported in the dxf format 158 BeginBlock lt optional comment gt NumbOffSet 100 Pt 1 0 1 0 Pt 2 0 1 0 Pt 3 1 1 0 Pt 4 1 1 0 Surf 1 A surface 1 2 3 4 EndBlock lt optional comment gt Any code between a matching set of BlockBegin and BlockEnd s can be collopsed by clicking a small in the left side of the editor making it easier to handle large files in particular with many layers All blocks can be collapsed from the view menu in the ODEONEditor And Blocks can be nested that is contain Blocks within Blocks Defining constants Constants must follow the syntax Const lt Name gt lt Value gt where value is a mathematical expression which may be based on numbers or constants and variables that has already been defined Example 1 Const CeilingHeight 3 4 Example 2 Const FloorLevel 1 Const CeilingHeight FloorLevel 3 Example 3 Const FloorHeight 1 Const Length 6 Const Ceiling Height FloorLevel
205. mainly supposed to have an effect in the frequency domain the assumptions being that transducers are closely spaced and that signals are periodic Delaying a transducer means that a signal will be emitted later from that transducer than it would otherwise Example If transducer 1 is delayed by 0 ms and transducer 2 is delayed by 3 ms then at t 3 ms transducer 2 emits the same signal value as transducer 1 emitted at t 0 by having a delay the transducer looks back in time 10 3 Playing with level If the level is not the same for all the units in the array but if it increases gradually from bottom to top the beam of radiation becomes asymmetric An example is shown in fig 10 6 with a level increase of 2 dB per unit 101 PPESP SIT ISE Perea rit tas PSR PPP CST TTI S PPP eee tte pepe eet ie Perea rit Figure 10 5 The radiation in octave bands from 250 Hz to 8 kHz for a line array with 7 units The level increases by 2 dB per unit from bottom to top 10 4 Combining delay and level adjustments With the combination of delay and level adjustments it is possible to design a sound radiation that is asymmetric and directed off the horizontal axis see the example in fig 10 7 This can be used to create a very uniform sound level over an extended audience area 100 He Figure 10 6 The radiation in octave bands from 250 Hz to 8 kHz for a line array with 7 units The time delay is 0 1 ms per unit and the level increases
206. n A Room acoustic parameter list interface allows the user to edit existing parameters and create his her own ones This is a great feature for improving parameters of inventing new ones especially for research purposes Apart from these an improved calculation setup improved grid scaling echo detection curves dynamic diffusion curves and many other features were introduced You can read more online at http www odeon dk version 11 Features introduced with version 10 ODEON 10 is Unicode and utf 8 enabled allowing text in complex character sets to be saved with your projects whether this is in the text files such as the geometry files par or texts composed in the various comment fields in ODEON the material library is not fully Unicode enabled yet In particular Unicode allows text in Asian character sets Japanese Chinese and Korean to be saved from within the ODEON application It is also possible to save text which is a mix of texts in different character sets When upgrading to version 10 or later the order of materials in the Material list has been revised please see Chapter 4 Read more on the features introduced with version 10 at http www odeon dk version 10 Features introduced with version 9 1 With ODEON 9 1 is delivered a new type of Dongle which makes it easier for the user to install new upgrades This is more thoroughly described below in chapter 1 4 and at www odeon dk dongle update Version 9 1 has a fa
207. n Chapter 2 an architectural model needs to be greatly simplified for acoustic calculations In figure a typical cross section of a diffuser is shown The surface is not planar It has a pyramidal structure which needs to be simplified Compensation of this simplification is done by using an appropriate scattering coefficient For the depth of structure specified in the figure Depth of structure the scattering coefficient is found from the graph in figure Scattering coefficient according to depth of structure For example if the depth is 2d 10 mm the scattering coefficient will be 20 The same process can be applied for any rough surface not only a uniformly shaped diffuser In this case an average depth of structure should be used for figure 6 5 It is important to understand that even a few rough objects sparsely placed on a surface increase scattering coefficient a lot The default scattering coefficient in ODEON is 5 which corresponds to a diffuser with a depth of around 20 mm However most of the times the surface does not need to have an exact shape as in figure Depth of structure to provide a high scattering Even a window frame on a flat can introduce as much scattering as the uniform diffuser in figure Depth of structure Depth of structure A diffuser with a depth of structure edual to 2d This type of structure is simplified in ODEON by modelling it as a plane surface and assigning scattering coefficient according to Figure 6 5 T
208. n in the roulette wheel according to its fitness Parents with higher fitness value are more likely to be selected Random Parents are selected completely randomly Tournament A fixed number of individuals are selected randomly from the whole population This is a Tournament Field Only the best parents are kept from this field This process is repeated several times until a specified number of the best individuals is collected StochasticTournament The best parents from a Tournament Field are selected by selecting better fit individuals proportionally Tournament Field times Ellitist the top n percent of the population is chosen and re chosen Individuals per material A sufficient population of individuals is needed for a genetic evolution By default 2 individuals per material are generated For example if we use 10 materials in our search 20 individuals will be included in the procedure Other common values are 4 6 and so on The values always need to be multiplies of 2 Crossover Probability Determines which portions of two parents will be combined to create two new offspring Inversion Probability Controls the likelihood that inversion will occur in a child s chromosome e g whether a portion of the child s chromosome will be flipped The probability is between 0 never and 1 always Mutation Probability Controls the likelihood that each bit of a child s chromosome will be mutated flipped between 0 never
209. n of a Bounding Box hi TST He HEL egy fi Wiyata MRH ALANA ae LN TT BRR AAU ine BREN NSS Ny Kee lous versions Below we have gathered the most important developments introduced with the recent ODEON versions Please read more online at http www odeon dk development room acoustics software Features introduced with version 12 1 4 Features of prev A very important version brought simulations and measurements together A room impulse response measuring system was introduced that allows capturing of impulse responses with the 15 aid of a loudspeaker a microphone and a laptop The tool is available in all editions from Basics to Combined and practically offers simulations and measurements inside the same software package Moreover ODEON 12 was equipped with full support for ISO 3382 3 Open Plan Offices tools for grouping receivers in the multi point response two types of radiation in line and surface sources calculation of IACC and many more enhancements Read more about them online at http www odeon dk odeon12features Features introduced with version 11 Parallel processing was introduced with ODEON 11 providing faster calculations on multicore computers The parallel processing is used for multi point and grid responses where many receivers have to be calculated at a time In addition a Fibonacci Spiral distribution used in ray tracing allows a perfect distribution of source radiation patter
210. n of the first point and surface in the Box Using the same number but with negative sign defines the box and its counterpart mirrored in the XZ plane Y 0 A Box will take up several point and surface numbers which must all be unique lt Length gt Length is oriented in the X direction on the Figure lt Width gt Width is oriented in the Y direction on the Figure lt Height gt Height is oriented in the Z direction on the Figure lt T B N gt The 7 B N parameter specifies whether the Box should have a top and or a bottom The options are T B TB and N for none Insertion point The insertion point of the Box is always the centre of the floor bottom surface Connection points The four foot points in Box are stored in PlistA The four top points in Box are stored in PListB The Box example shown was generated with the following code BoxStatement par HHH const L6 const W 4 const H 2 7 Box 1 LWH TB Walls floor and ceiling HHH The Cylinder statement The Cylinder statement defines a cylinder shell with or without top and bottom The statement may typically be used for modelling cylindrical room or columns The Cylinder2 statement which creates a cylinder of the calotte type will usually be preferable for modelling cylindrical ceilings The syntax for Cylinder is 175 Cylinder lt Number gt lt NumberOfSurfaces gt lt Radius gt lt RevAngle gt lt Height gt lt T B N gt lt optional name gt lt
211. n the ODEON Rooms folder Specify transmission data for surface number 19 of type Transmission en Se Add or edit a material Maternal data List of transmission matertals in room Create clone of material for this wall Reduction indexes Frequency 50 100 200 400 an0 1600 3150 6300 Hz A izooe 3200 45008 soo 5600 4600 53 00 f Frequency 63 125 250 500 1000 2000 4000 6000 Hz R yao 2100 aoj 4800 saoj ssool soooja saog dB x ba Frequency a0 160 315 E30 1250 2500 S000 10000 Hz R 406 2700 nop so sob asco s20 sao Batch operation on reduction values Copy to all bands 0 10 E dE Increase level by from first band Copy centerbandsto 1 3 octave side bandes o o o o 7 Double sided wall seperate surfaces Description Light double wall Update double sided wall V Update selected wall upon exit Sugested wall to be updated wall number 18 Wall thickness 0 25 metres In the Specify transmission data interface you can insert transmission coefficients in dB for 1 3 octave bands If your data come only at full octave bands you can insert the full octave band values and copy the data to 1 3 side bands by pressing the Copy center bands to 1 3 octave side bands button The option Double sided wall allows having the same transmission coefficient for two surfaces of different absorption 143 Remember With the transmission tool ODEON is able to calculat
212. n uncompressed wave files having the well known wav extension Most if not all resolutions are supported so files edited by programs such as CoolEdit or Adobe Audition can be used without any conversion being required e 8 16 24 32 bit PCM e 32 bit IEEE Float e 8 16 24 32 bit PCM Extensible tells the number of significant bits e 32 bit IEEE Float Extensible includes some additional info which is not used for input by ODEON 60 To be compatible with the HRTF s supplied with ODEON the wave files should always be at a 44100 Hz sampling frequency The ODEON program comes with a few anechoic samples which are installed to the ODEON WaveSignals directory If you wish to extend the library of input signals you should put your new signal files here or in a subdirectory to this directory e g ODEON WaveSignals English voice or ODEON WaveSignals NoiseSignals A few audio CD s containing anechoic recordings are commercially available namely the Archimedes CD Music for Archimedes 1992 which contains some recordings of solo instruments and the Denon CD Anechoic Orchestral Music Recordings 1988 which contains semi anechoic stereo recordings of orchestral music The easiest way to transfer recordings on an audio CD into wave files on the hard disk on a computer is probably to use a software application known as a CD ripper This also ensures the transfer is without loss in quality Signals ripped from CD aud
213. nah Ma E aad m anan Er d T OE Aand Ai M Improved Source Receiver Positioning In ODEON 13 it is possible to change source or receiver position directly on the 3D Edit Source display using CTRL LMB While you rotate the geometry a grid is adjusted to the visible plane to help anchoring the source on its nodes In the next figure the grid is displayed for the XZ plane i y DASANAN ise THAN EL lt N iit a i ry J Nai SS Improved Tabulation Tool Making an array of sources of receivers is easier than ever Now all members of the array are displayed in the 3D view in real time before changes are confirmed 10 s _ IN Hi ie tT UES Le is AA BF VENI ie i AX TOX Aff i PD y H 4 KYA ZIYA AKNA Projection of x and y coordinates Horizontal grid x and y relative to horizontal plane WAS gt C Projected grid x and y relative to projection as defined by azimuth and elevation 4 00 Metres Nx 5j Second dimension Y 5 00 Metres Ny 2 4 Elevation angle angle 0 00 Keep relative source receiver height Additional drop height 1 00 Metres drops new items from san height of new items is set to the sami ea we Pi a Improved Source Directivity Editor The already known directivity polar plot editor in ODE
214. nce of impulsive noise during the measurement no matter if the impulse response was recorded directly using external impulses or obtained using the sweep method implemented in ODEON If in doubt it is a good idea to make an extra measurement while still in the field to see if it is consistent Try to make a measurement with a much longer sweep if using the sweep method Normally such errors will issue a WARNING message in the title bar of the measurement window stating the value of XI non linearity parameter being too high Remember ODEON helps you making a high quality measurement by displaying WARNING messages in the title bar of the measurement window Moreover when the signal to noise ratio is not sufficient ODEON will display asterisks instead of values for the various acoustic parameters In the Errors subfolder inside the Measurement folder of your ODEON installation there are examples of bad impulse responses due to distortion low signal to noise ratio and electromagnetic feedback While signal to noise ratio or electromagnetic feedback may be something easy to track in the measurement distortion does not provide clear signs and might let us accept an overall pure quality impulse response In the following examples of good and bad impulse responses are given 115 Warning messages Depending on the quality of the measurement ODEON may come with one of the warning messages in the title bar of the measurement window pr
215. ncies above the Nyquist frequency If this is the case you will notice that the sweep reproduced over the loudspeaker does not appear to be going monotonically from low to high frequencies If pronounced it is recommended to use a better e g external sound card Similar Setting microphone amplification If using the built in microphone amplifier it has a mixer slider which should be set to max manually by the user However it is likely also to have a dedicated microphone amp boost which can typically be set in the range 0 to 30 dB be sure to set it so a desirable response is achieved If using a different setting in your measurements than used in the calibration be sure to adjust the external adjustments in the measurement setup accordingly Disable system sounds in mixer This is important so that no disturbing sounds pop out during a measurement Setting up your sound card correctly Set up your sound card correctly for use with the ODEON measurement system Many soundcards are equipped with advanced features such as noise cancellation echo cancellation reverberation which may be useful when using internet video conferencing or listening to music However any such feature is highly undesirable when using the sound card for room acoustics measurements thus should be disabled For example many PCs come with the built in Realtek audio device You can access the global settings for this device within from the Realtek Audio Manage
216. ndent sound power level the equalization option in the Polar plot editor is used for entering the dB values Calibration Three different options are available in the Polar plot editor e If No calibration is selected ODEON will use the SPL dB values as entered in the polar plot table adding the equalization values entered This option is typically used if the source is a loudspeaker e If Calibrated source is selected ODEON will add the equalization if entered then shift the resulting SPL s of the source in order to obtain a reference sound power level of 0 dB re 10 Watt at the selected Calibration frequency band typically 1kHz This calibration type is typically used when a generic source with an adjustable level is needed for calculations such as the OMNI or SEMI directivity pattern e If Sensitivity calibration is selected the SPL s of the source will be shifted in order to obtain the SPL in dB on front axis of the source at the distance specified as Ref distance metres at the selected Calibration frequency band This calibration type is typically used if the free field sound pressure level measurements are available for a natural source Maximum dynamic range The minimum level will be Max level minus Maximum dynamic range If the range is large the display may be difficult to interpret view in the directivity viewer The source will usually have its max levels on its axis Text format Normally the text format should not be nece
217. neadnenendnssdnescaoantoadadeareseadinns 102 10 5 Usmo Ne cgual ZET cepporno N N S 103 10 6 Bringing the array mio the TOCIN seioscccscccecssceccasnsssosgasnasasacssaneeonsemsosasaaueanssonasenstanaisociacesonsentas 103 11 Measurement Systein nn cecssssscssnsssssesssnecsssecsssecsssecesssecsssecssnecessesesseeesneees 104 11 1 Terai al DICK o rO eeso anaE A SE SESA E E 104 11 2 PECL MONI G1 AEEA EE TO E E T O N A AE 106 11 3 Measurements ae FPO COSS HIP sorain en a Erer E E E aA E 106 11 4 IV ST EC U a E E E E E E 111 11 5 Editino mopulse RESPONSES sieneen aE a d E T eee ee 113 11 6 Making auralisations with Impulse Responses ssesseseeeeseeseerrerseesersresresresresrreenrenresreseeses 114 11 7 Examples of Impulse Responses Warnings ccees cess cesecseeseeeeeeseeesesseeesesesesesseeeeeeens 115 12 Genetic Material Optimizer nnn 125 12 1 General Steps for Material Calibration an saredseiiaucononstveanee earmpe seed soutaureen eased aumoubeaacpeneaunas 125 12 2 How Genetic ak OV Or Kopara a ire AEE E 128 12 3 Ferlota ae rie iii Za ON seirer isa E EENET 131 SCL eee 138 APPAIA Cd DULAT eosar E E E OEE EAEE OE ENEN 141 Appendix B Specify Transmission through walls essessesserseessessessesresrssrrsesresresresresrestesrnrenreneenreseesees 143 Appendix C Description of XML format for import of array loudspeaker data cece 146 Appendix D Modelling rooms in the ODEON EditOf esessesseeseesseesesesresrrsesresresresresresensr
218. ngle point Multi point and Grid response results if this predefined parameter is included and set to Visible in the Room Acoustics Parameter List for reliable results the reflection density Density reflections should be greater than 25 ms or so If multiple sources with similar source power are visible from all most receivers then the Number of Late rays may be decreased significantly If the room model contains strong decoupling effects or uneven distribution of the absorption it is desirable to increase the Number of late rays even if the calculated reflection density seems high enough It is recommended to experiment with different Number of late rays and study the decay curves in Single Point Responses If many surfaces are added to a model in between calculation it is recommended to re specify the Number of late rays too The ODEON algorithms are hybrid methods based on the Image source method and ray radiosity for the early part of the energy and ray radiosity for the later part of energy Therefore there is also a Number of early rays 88 Specialist settings Below the General settings are the Specialist settings which should only be adjusted if you are an ODEON expert or experimenting One exception is if you are importing rooms from earlier versions of ODEON then it will be a good idea to enable Screen diffraction to make sure diffraction around screens is included in future calculations Impulse response details Max
219. ngs in a room The 1000 Hz octave band is shown Hand claps is an effective way to obtain impulse responses in rooms In the next figure the broadband signal is shown Onset and truncation times are derived automatically but the signal to noise ratio is very poor for the selected impulse response and its 117 length is too short so that a stable noise floor cannot be derived Therefore some acoustic parameters cannot be calculated in this case Decay range less than 45 dB 24 48 dB at 250 Hz Measured response C Odeon13Combined Measurements Errors Handclap1_2_3_4 wav tojta Raw Impulse response broad band Raw decay curve broad band Decay curves broad band T 30 seconds Decay curves all bands Energy parameters Parameter curves tafata C Odeon13Combined Measurements Errors Handclap1_2_3_4 wav Raw decay curve broad band M E Measured M Onset time Truncation time l Noise floor i 7S Se E eee q 4 a 12 lt l e9 Time seconds The same hand clap recordings as in previous figure but squared so that the dynamic range is more visible Broadband signal is shown Impulse response with electromagnetic feedback When long cables are used in a measurement for the microphone as well as for the loudspeaker and the cables lie close to each other and parallel for some distance this can lead to electromagnetic feedback resulting in an early peak
220. nl gt lt ZMathExpression1I gt lt XMathExpression2 gt lt YMathExpression2 gt lt ZMathExpression2 gt NumberOfPoints lines each defining a point in the multi point sequence should follow the MPt statement lt Number gt A unique number from 1 to 2 147 483647 for identification of the first point in that multi point sequence lt NumberOfPoints gt The number of points defined by this multipoint statement if the number is 3 then 3 lines should follow each describing the coordinates of a point Example 1 defining point number 100 in x y z 1 1 1 and point number 101 in x y z 2 2 2 MPt 1002 1 0 1 0 1 0 2 0 2 0 2 0 As a special option for multi points it is possible to repeat a coordinate used in the previous point of that multipoint sequence or to repeat the coordinate while adding or subtracting a value from that point Example 2 defining point number 100 in x y z 1 1 1 and point number 101 in x y z 1 2 0 MPt 1002 1 1 1 l Defining a series of points using the CountPt statement The CountPt statement must follow the syntax CountPt lt FirstPointNo gt lt MaxCount gt lt X MathExpression gt lt Y MathExpression gt lt ZMathExpression gt Use the CountPt statement to define a series of points using a counter This statement makes use of the predefined counter PtCounter which will run from 0 to MaxCount 1 producing the points with 161 the numbers FirstPointNo to FirstPointNo MaxCou
221. nn refers to the row in the table Conv no The wave files created as results from the Mix convolved wave results into one wave file table will have the extension MixAuralnn Wav where nn refers to the row number in the table Mix No The output from surround auralisation follows rules similar to those of the binaural ones the impulse responses are stored in wave files following the WaveFormatExtensible format where a signal is available for each loudspeaker channel in the specified surround setup The impulse response has the extension SurRoundnn Wav where nn refers to the relevant job number The convolved files have the extensions ConvSurRoundAuralnn Wav where nn refers to the row in the table Conv no and vice versa for the mixed files These files should be playable using the Windows Media Player Publishing audible results on the Internet To publish calculated demonstration examples on the Internet it may be useful to convert the result wave files into compressed mp3 files or wav Mpeg Layer3 files as download times for wave files can be lengthy One minute of compressed stereo signal will depending on the compression rate take up approximately 1 MB If publishing examples make sure that copyrights are not violated You are free to publish examples which are calculated using the anechoic examples supplied with ODEON For the orchestra recordings the conditions as specified in the software license agreement 8 must be observ
222. not an integer it will be rounded to the nearest integer value lt CountTo gt Last value the counter takes The CountTo value must be greater or equal to the CountFrom value The For statement will take CountTo CountFrom 1 loops The CountTo value is considered an integer value if the number entered here is not an integer it will be rounded to the nearest integer value The following example will produce the points 1 to 5 with the X coordinates 5 10 15 20 and 25 metres while the counter MyCounter loops through the values 1 to 5 For MyCounter 1 5 Pt MyCounter MyCounter 5 0 0 end When using For End constructs it should be remembered that point and surface number must be unique This is easily obtained by incrementing the special variable NumbOffSet appropriately in each loop An example on this kind of numbering can be found in the sample file BoxColumnRoom Par where NumbOffSet is incremented by eight in each loop each time a new column which contains 4 surfaces and 8 points is created 168 Sample room files ForRotunde Par BoxColumnRoom Par Unit The Unit statement is used if you wish to model in a unit different from metres The unit used in the parametric file is by default assumed to be metres however if you prefer to model in another unit this is possible using the Unit statement Check the example below Example modelling in Inches Unit Inches You may choose your unit among the following predefined Met
223. nt 1 Use the PtCounter in the expression of the x y and z coordinates to create the desired differences between the count points Example defining 7 points on a circle with a radius of 10 at Z 0 metres CountPt 100 6 1 10 CosD PtCounter 360 6 10 SinD PtCounter 360 6 0 Note First and last point in this series of count points are equal redundant This will typically be desirable when using the CountPt statement along with RevSurf statement Defining a single surface using the Surf statement A Surf statement is divided into two lines and must follow the syntax Surf lt SurfaceNumber gt lt Optional Description gt lt ListOfPointNumbers gt The Surf statement is used to define a single surface in some situations with symmetry two surfaces The Surf statement is constructed from two lines one identifying the surface by a number and an optional name and another with a list of corner numbers lt SurfaceNumber gt A unique number from 1 to 2 147 483647 for identification of the surface Using the same number but with negative sign defines the surface and its counter part mirrored in the XZ plane Y 0 The surface number may be defined using mathematical expressions lt Optional Description gt A string displayed and printed for easy identification of the surface Could be something like Main floor lt ListOfPointNumbers gt Each surface may be bounded by between 3 and 500 corners which all lie in a plane
224. nt if Angular absorption is enabled in the Room Setup e Air absorption due to the length of the reflection path e Distance damping due to the distance travelled from the primary source to the receiver e Scattering loss frequency dependent due to the scattered energy which is handled by a scattering tree Scattering may occur because of surface roughness as specified by the scattering coefficients in the materials list or due to limited surface dimensions or edge diffraction Early scattering In short each time ODEON detects an image source IMS an inner loop of scatter rays not visualised in the 3D Investigate Rays display is started taking care of the scattered sound which is reflected from this image source surface Example If all scattering coefficients in a room is 0 5 then the specular energy of a first order IMS is multiplied 1 0 5 and the specular energy of a second order IMS is multiplied by 1 0 5 1 0 5 The scattering rays handle the rest of the energy The early scatter rays are handled in a way which is indeed inspired by the way in which ODEON simulates surface sources actually each time an image source is detected ODEON will simulate a surface source which will emit Number of early scatter rays from the image source surface The early scatter rays will be traced from the current reflection order and up to the transition order At each reflection point of the early scattering rays including the point
225. ntemor Exterior mode fol eats verte Peurtace List cd P Room materiai library Elmia RoundRobin2 detailed LS F r r Number Specification snena n eae ma a eA _ 2036 Tepoich in Sh ngerwebart 4 5 mm dick inpr gniert direkt auf Boden Ref M Hedd HA Muller Tascherbuch der Tedrischen Akusak Springer 1994 3005 0659 0 000 Normal AUDIENCE AUDIENCE 0 59 PUETA Satirvorhang 82 Rayi 20cm vor Wand 1 Sfache Faltung Ref M Hedi M A Miller Taschenbuch der Technischen Akustk Springer 1994 30004 0 60 9 000 Noma AUDIENCE AUDIENCE 33 aka edurdene Mner ataser tayijicm 30 mm dick Ref M Mech H A Muller Taschenbuch der Technechen Ahustik Soringer 1994 004 040 9 000 Moma AUDIENCE AUDIENCE 3 68 Ec 7i LERI gebundene Miner alaserpiatte 12 Rayliom J0mm dck mit 50mm ichtem Wandabstand morvtert Ref M Heck H A M ler Taschenbuch der Technischen Akustk Soringer 0004 0650 5000 Norma AUDIENCE NDINE 3 60 3 f 2040 Mresafacersiatce 20 Rayi cm Somm cick mit Simm Schtem Wandabstand sichtseitg mit mm Sperrholz abgedeckt Ref M Hedd H A Muller Taschenbuch der son osol ooo Moma AIOE ADDE 280 ae 2041 Holz 160m dick auf 4 0m Mottlatten Ref M Hedd H A Miller Taschenbuch der Technischen Akustk Springer 1994 i e004 055 0 009 camel AUDIENCE AUDIENCE 280 A 2042 Goskartorglatie 18 mm cick 16 gin 2 400 mm wor starrer Wand Ref M Hedd H A Midler Taschenbuch der Technischen Akustik Sp
226. ny provision of this Agreement shall be declared void or unenforceable the validity of this Agreement and all other provision shall not be affected ODEON product support and service is only provided to registered ODEON users with a valid support agreement Contents 1 ODEON TiS ANNU NOUNS ites satists cece teecresntetveres ese hcspecamnies coerulea eevee 8 1 1 Installing and running the program si scesscasevsncnsavecucsascanvansaaansencnstesaccanndecsarsaabeaancuseesszotanecunuyicds 8 eZ Upgrading from previous VELSIONS wise cssestasesendacactedasossansdscsaancagacswebiaaenvesesosoeesiassacenasieatnndsiaatecen 8 1 3 Anats nent vain ae Bah Ri hc Demeewereerrrerr errr E ree mre rerre reece rr rer creer eccrrer Terr rer 9 1 4 Features of previos Vel SIO S asec sce cacy sosaontse sa cengautemtic tesa a enie ciate nn 15 1 5 How to upgrade or update your current License eee eeseeeeceseceseeeseecsaeeeseeeeecsaeseseeesees 17 2 Modelling ROOnns 0 0 cccccccccccccccccssssessssesssseesssseessssessssessseeessescsssesssesssesssseeesseeessseessseses 20 21 Precalculated KOONS ipene arson ar E a E Ea aE 21 22 Guidelines Oi FOOT Modeling esenee T 23 23 poring a modeliron oket ela 6 eerie n ni n rere nee ene men err ne er een ee er 25 2 4 porne DAF aad Bs e e aen r ner Conte arene rr ne T E ee enone er 27 25 TRE ODL LONT eire on Mode Ils saeeenpe eee mertoernr rere A er rears 33 2 6 iyikere Bol ncerel atte EON epee E et eerie re en reer een no
227. o noise ratio Decay range less than 45 dB 20 09 dB at 4000 Hz Measured response C Errors Auditorium21_Insufficient_Signal2Noise_S2R2_2000ms wav Raw Impulse response at 1000Hz Raw decay curve at 1000Hz Decay curves at 1000Hz T 30 seconds Decay curves all bands Energy parameters Parameter curves Freque ac C Errors Auditorium21_Insufficient_Signal2Noise_S2R2_2000ms wav Raw decay curve at 1000Hz M E Measured M Onset time M Truncation time M Noise floor i i ooo TST gi np i ul Hl ij my SPL dB ren 0 1 Time seconds Impulse response of insufficient signal to noise ratio Same recording as in previous figure but squared Tip Every doubling of the sweep length results to a suppression of the noise floor by 3 dB As an example increasing a sweep signal from 2 sec to 64 sec will result to a 15 dB gain in the signal to noise ratio Impulse response between buildings This is an example of measurement outdoors between several buildings In this case distant strong spikes at high frequencies indicate the presence of prominent echoes between the buildings At low frequencies the effect is less prominent In addition the reverberation time is much longer at low frequencies 123 C Odeon12Combined Measurements IR_BetweenBuildings2 wav Raw decay curve at 4000Hz Jv E Measured M Noise floor Onset time Truncation time
228. of the Extrusion Modeller in the right Figure however a few tricks may be found in the description below If modelling a room in the XY Modelling plane then a chair may in effect be considered an extrusion which excludes the top and three of its sides In this example we will create a chair with the seat dimensions 0 4 x 0 4 and a back rest with the height of 0 4 Legs and other small details should be omitted To make things easier do the modelling around origin then move the chair to its final location when finished When modelling around the origin it becomes easier to read the dimensions of the seat of the chair and to use grid and snaps without the need to calculate dimensions of the seat e Set the snap size s to 0 4 metres e Click the 4 points in the seat of the chair Insert or Esc toggles point input on off e Change the z coordinate in the chair to 0 4 metres in the Surface editor to define seat height e Change dZ to 0 4 metres in the in the Surface editor in order to define the height of the back of the chair e Uncheck Top in the surface editor e Uncheck the 3 sides which are not the back of the chair in the Point editor 37 e Finish the surface by pressing the Insert or Esc key e Finally move the chair to the desired location using the left mouse button Using the circle tool and the mirror 100 Io 00 hoo o a Pal li hoo 0 00 0 00 1 00 1 00 SS 0o 2 00 2 00 Z 1 00 ji 00 In this example a t
229. ogram changes from previous versions etc Chapter 2 covers geometry modelling New releases of ODEON often include new facilities which can speed up the modelling process as well as tools for verification of geometries Some of the facilities are a plug in for Google Sketchup extensive support for import of CAD models in the axf as well as the 3ds 3D Studio format and a stand alone drawing program for modelling of so called extrusion models Tools for verification of room models are also covered in this chapter With improved support for 3 party tools for geometry modelling knowledge about ODEON parametric modelling language is becoming of less importance The advanced user will find that the description of the par format has been moved to appendix E The format features support for symmetric and semi symmetric rooms use of constants variables counters loops etc and can be useful when studying geometrical parameters such as angle of reflectors etc As all geometry also imported ones are stored in the par format it is possible to add features described in terms of parameters to an imported geometry if familiar with this format Chapter 3 deals with the materials to assign to the surfaces of the rooms absorption scattering and transparency coefficients as well as transmission data Special materials that may speed up the modelling process and how to manage the material library are also covered in this chapter Chapter 4 deals with the a
230. olX NumbOffSet NumbOffSet 8 comment hint setting NumbOffSet to Auto would do the same job Box I Column W ColumnW h n columns in the room MTranslate L NumColX 1 0 0 end end HHH Using hybrid statements and coordinate manipulation The following example demonstrates an example on how to use the hybrid statements Cylinder2 and Dome2 as well as the coordinate manipulations which are essential to the use of the hybrid statements This example is a rather complex one so the main parts of the file is explained below Line 3 7Defining constants Line 8 9 Inserting the cylindrical wall which needs a rotation of 90 around the Z axis Line 11 The foot points of the cylindrical wall which is temporarily stored in PlistA are stored in PlistO for later use definition of the floor Line 12 13 Inserting the dome shaped ceiling The Z rotation has already been set to 90 when the wall was created Line 14 18 Setting the coordinate manipulation for the ceiling and creating the ceiling Line 19 Resetting the coordinate manipulation to work in absolute coordinates Line 20 23 Creating Wall floor point Line 24 25 Defining floor using the cylinder points stored in Plist0 Line 28 29 Defining side walls using symmetric modelling Line 30 31 Defining back wall using the ceiling cylinder points which is still stored in PListB Dome2 and cylingder2 x 2 room par HH pe const H 5 i Const L 10 Const W 15 Const N 12 Const H
231. ollowing algorithms The frequency fi of the lowest sub band in an octave is defined as follows 1 f Fe x2 And the following frequencies fn sub bands are defined as 1 Ja far Where n 2 to n The 6 sub bands Ns 6 centered around 1000 Hz have the following centre frequencies 1059 46309435929526 1189 20711500272106 1334 83985417003436 Frequency of n sub band in 1000 Hz octave for Ns 6 Results of calculations should be for full octave bands for the octave bands 63 125 8000 Hz ISO suggests that center sub octave frequencies should be derived from 1000 Hz For an uneven number of bands e g 1 3 octave bands this gives us one band below one band at and one band above the center frequency of the Octave center frequency so the 3 sub octaves are in balance relative to the octave band however for an even number of sub band frequencies thing becomes unbalanced if using ISO frequencies there will be 2 sub bands below and 3 sub bands above the full octave band centre frequency or opposite 3 below 2 above Attributes Re Img The energy of the transducer is emitted in Ns Sub nodes The center frequencies of the Sub bands are centered around the full octave band centre frequency in order best to simulate the energy in each octave band If Ns is odd e g 5 then these frequencies will coincide with ISO centre frequencies Be aware though that the centre frequencies used will differ if an even number e g
232. om a source In ODEON Combined and Industrial version only three different kind of sources are available the point the line and the surface source Point Sources For Single Multi and Grid response calculations and for the 3D Investigate Rays and 3D Billiard displays rays are radiated in directions distributed as evenly as possible over a solid angle To obtain the even distribution ray directions are emitted in directions given by the Fibunacci spiral Using a spiral also ensures that there will be no preference towards or away from e g horizontal rays as 79 might be the case if rays were radiated from a number of e g horizontal rings For Quick Estimate and Global Estimate the radiated directions are chosen randomly allowing the calculation to be finished after any ray without getting a very uneven distribution of radiated directions Surface Sources and Line sources Industrial and Combined versions only For these source types points of origin for the rays and their initial directions are the same no matter the calculation type For each radiated ray a random point or origin is chosen at the line or surface source From this point a ray is send out in a random direction and two types of radiation can be chosen Lambert and Spherical When Lambert is chosen for the radiation the strength of each ray is proportional to the cosine of the angle between its direction and the normal Practically rays radiated parallel to the surface h
233. om material library will automatically change depending on 3 criteria 1 Materials from the surface list that are matching the new global material library will stay the same 2 Materials that are not in the global material library but where there is room for the original number from the surface list in the material library is added to the room material library with the same number 3 Materials that are not in the global material library and that has a number which conflicts with one from the new material library will get anew number at the end of the local room material library If not satisfied with the new global material library it can be replaced with another library under materials in the installed files 54 gt Auralisation Combined and Auditorium editions only Although much effort has been made to make it as easy as possible to use the auralisation capabilities available in ODEON its felt that a separate chapter is needed as this is where all the threads from room acoustics modelling signal processing wave signal files transducers psycho acoustics recording techniques etc meet In the description of auralisation techniques special words are frequently used please refer to appendix A Vocabulary for a short description In this chapter it is assumed that you have tried the basic auralization functions in the Quick start guide tor ODEON Auditorium or ODEON Combined The basis for auralisation in ODEON is ei
234. oma Audenetsbies DF oo r a 7 an m 100 Rough concrete Sobran 1973 geie m m anin 4S 61 3x04 005 0000 Noma STAIRS A os Sa pT am 1973 ea mna naarn n aan Aiarmal crame ctame once OCK p of 5 7y Mat 14307 Au ence tebies Aud21 Theatre seats empty hiip www see eduireferenca_nsterai tl 104 Concrete block with or without plaster painted Ref Dalenback CATT 63 tz 125 Hz 250 Hz 500z 1000 00H 4000Mz 8000Mz olw Cass a 105 crows concrete blocks methout surface fresh 400 800 kan Kratern 1964 0 04200 0 05500 0 06009 0 08400 0 06800 0 12400 0 12500 0 14400 0 10000 Not dessified 63Hz 125 He 250 Hz 500Hz 1000Hze 2000Hz 4000Hz 8000 Hz olw Class cao Gata wmasegned wal type Normal 0 00000 0 00000 0 00000 0 00000 0 00000 0 00000 0 00000 0 00000 0 00000 Not cassified Have a look at the jobs defined in the Job list SE The first two jobs are of particular interest since they have the Multi point response enabled The material optimization is done based on responses of multiple receivers and comparison to measurements Run both Job 1 and job 2 and see the results in the Measured versus Simulated tabsheet in the Multi point response results The difference between measurements and simulations for most of the parameters is significant 131 BD JobList Binaural mode Headphone Subject 021RestOdeg_diffuse wav feleTys Active sources for the selected job pa m P1 S01 Point source at x y 2 1 50 1 50 1 60
235. ometry which was exported from ODEON to the CAD program is to be imported into ODEON again this option should be on in order to import all 3D data In some cases it may be desirable to switch this option off as the DXF file may contain such entities which the modeller did not intended to be included in the 3D surface model to be imported the entities may have been modelled for other reasons e g as assisting lines in the modelling phase 3DSOLID REGION BODY recognized but not supported These entities are ac s solid modelling entities which are not directly supported by ODEON However solid modelling is probably the most powerful way of creating 3D surface models allowing the use of commands such as UNION SUBTRACT INTERSECT SLICE INTERFERE etc and with a few steps it may be possible to convert these entities into something which is understood by ODEON In Cadopia IntelliCAD Professional the 3DCONVERT command will convert above mentioned entities into entities recognized by ODEON Poly faces It is recommended to perform this operation on a copy of the CAD file rather on your original In 3DStudioMax select all the entities in the drawing using the ctril A shortcut then right click the mouse on the drawing and select the Convert to Polyfaces other options may also work It is recommended to perform this operation on a copy of the CAD file rather on your original Since AutoCAD 2000 the conversion process involves exporting to a 3ds f
236. ommon Loudspeaker Format was www cifgroup org developed and is maintained by the CLF group at www clfgroup org It is an open though secure file format for loudspeaker performance data and polar plots Loudspeaker manufacturers can use the CLF format to supply data to end users of professional acoustic computer programs CLF is defined in two parts a binary format for data distribution and a text based format used solely by the loudspeaker manufactures for data input and editing As a user of ODEON you should deal only with the binary distribution files having the extensions CF1 and CF2 In order to view all data in the CLF format you should download a free viewer from the CLF home page CF1 has a frequency resolution of 10 degrees 1 1 octave and cr2 has a frequency resolution of 5 degrees 1 3 octave If data are available in either format for a selected loudspeaker then the cr2 format should be preferred because of its angular resolution Currently ODEON does not make use of the higher frequency resolution of the cF2 format however in the future ODEON will make use of this extra information for calculation of the distance dependent directivity of loudspeaker arrays which are composed from multiple units which are added with phase The CLF Group is providing a set of free tools for data editing conversion from text to binary format and viewing binary data allowing loudspeaker manufacturers to create view and verify binary distribution fil
237. on Coefficient material 4043 Range Cocft Coeff 0 45 _ o 0 30000 0 30000 0 12000 0oe000 0 06000 0 06000 0 05000 0 0S000 2 17 07 04 inital boards 100 mm cavity filled with Optimized 1 102 Smooth concrete painted or glazed 0 01000 0 012000 0 01000 0 01000 0 02000 0 02000 0 02000 0 02000 2446 9 g 9 35 F Current obran 1973 g 03 w Low Limit 2 3004 Wooden floor on josts Inger sev 0 15000 0 15000 0 11000 0 10000 0 07000 0 05000 0 07000 0 07000 187 57 30 B 0 25 s 949 02 High Limit 3 10007 Sold wooden door Bobran 1973 0 14000 0 14000 0 10000 0 06000 0 08000 0 10000 0 10000 0 10000 37 9 2 015 J Auto scale 4 30 30 absorbent 0 30000 0 30000 0 30000 0 30000 0 30000 0 30000 0 30000 0 30000 74 a oi in i ie Be 0 05 pamen T 5 14307 a Aud21 Theatre 0 02000 0 03000 0 04000 0 05000 0 07000 0 03000 0 089900 0 09000 190 79 9 63 125 250 500 2000 8000 N ON p E E E E E S a Frequency Hz v ra Error as a function of Generation i 63 Hz amp 125 Hz 5 250 Hz W 500 Hz 0 1000 Hz 4 2000 Hz 4000 Hz gt lt 8000 Hz 0 Generation Individual Receiver independent ray tracing Tracing rays Job Remaining time The interface of the Genetic material optimizer before any calculation has started Later in this chapter an overview of the different parameters is presented 130 12 3 Perform a Material Optimization The room Auditorium 21 at the Technical University of
238. on for finding an appropriate list of surfaces to edit e Toggle between Global and Local Material Library switches between the Global material library and Local material library e Remove extreme absorption removes absorption coefficients below 0 05 and above 0 95 from the materials in the surface list This should be used if there is doubt whether the absorption is estimated correctly Because these extreme values in selected absorption coefficients can have an extreme effect on results as well Oppenheim amp Schafer 1989 e Show reference on the web connects to internet page for selected material in material library if link is available 4 4 Opening an existing room form a version earlier than version 10 for the first time The material library has been changed radically for Version 10 The numbering of materials in the Global material library has changed and a many redundant materials have been removed Therefore some automatic update mechanism of local material library has been implemented to make the transition as smooth as possible If your model already had a local room library from the old version and you open it in ODEON Version 10 or later Then the Global material library will be the new ODEON library and the local room library will still be exact the same as you created in the old version If your model did not have a local material library in the old version and you open it in ODEON Version 10 Then the numbers in the local ro
239. onstruction in a reverberation chamber and if reconstructing only a fraction of the wall in the reverberation chamber it will have different absorption properties because it becomes more or less stiff 94 Hard materials Hard materials such as concrete are often listed as being 1 or 2 absorbing It may sound like a difference of 0 5 or 1 is not a significant difference However if a room is dominated by this material or if one of the dimensions of the room is a change from 1 to 2 is a relative change of 100 Materials scattering coefficients The knowledge on scattering coefficients is currently rather limited Hopefully in the future the scattering coefficients will be available for some materials Meanwhile the best that can be done is to make some good guesses on the size of the scattering coefficients and to do some estimates on the effect of uncertainty Measurements Eventually the reference data which you may compare with simulated room acoustical parameters are not perfect We must accept some tolerances on the precision of the measured parameters Receiver position s Common etrors are e to base the room acoustic design on simulations in one or only few receiver positions e to place the receiver close to a surface e to use too short source receiver distance Source Receiver distance Point response calculations made in ODEON are to be compared with point response measurements and as such the ISO 3382 1
240. oom acoustic parameters The ODEON measurement system utilizes a modification of the algorithm by Lundeby for automatically detecting the noise floor and truncating the impulse response at an appropriate time Lundeby 1995 105 11 2 Equipment The ODEON measuring system is very flexible Only elementary equipment is needed for an initial measurement e A microphone e A loudspeaker e Sound card e Amplification e Lap top PC Even a built in sound card of a portable PC could be used for a simple measurement However better quality of the measuring equipment yields results of high accuracy Built in sound cards often use low quality filters which introduce distortion and additional background level in the recorded response USB audio interfaces of moderate prices indented for gaming or music recordings can work pretty well Power supply directly from the USB port is preferable for increased portability A flat frequency response microphone and loudspeaker can help to obtain enough signal to noise ratio for as many as possible octave bands If someone wishes to experiment with some other stimuli than the sweep signals used by ODEON such as gunshots balloon paper bag popping only a mobile recorder is necessary The impulse response can be recorded as a wav file and be post processed afterwards in ODEON Regarding the loudspeaker this should be as Omni directional as possible according to the specifications in the ISO standard 338
241. oordSysOffSet node FarFieldDistance 1 In metres 1 indicates that ODEON should make its own estimate of the far field distance For distances greater than FarFieldDistance ODEON will use a pre calculated balloon calculated for that distance This far field balloon is calculated and handled entirely by ODEON FarFieldResolution 2 Far field balloon resolution in degrees 1 2 3 5 and 10 degrees allowed NumberOfSubBands 6 Frequency resolution per octave band applied for phase summation in the calculations Can be 3 6 9 A resolution of 6 is suggested as a good compromise between calculation time and quality of results Nodes to ArraySource Position Orientation CoordSysOffset EQ Transducer there can be multiple Position Node Attributes to Position X 1 Y 1 Z 1 Uses the same coordinate system as in ODEON and as in the room to be imported into Defaults to 1 1 1 and can easily be changed from within ODEON Orientation node Attributes to Orientation Azimuth 180 around vertical axis counter clockwise is positive Elevation 90 up is positive and down is negative Rotation 10 180 counter clockwise rotation around loudspeaker axis is positive Nodes to Orientation Vector Vector node Attributes to Vector X 1 Y 0 Z 0 Is used for defining orientation need not be a unit vector Above ts the default orientation of an array or a loudspeaker right turned c
242. oordinate system assumed Vector is just another way to specify Azimuth and Elevation not needed if Azimuth and Elevation have been specified defaults to 1 0 0 READER should allow both WRITER should normally just use the one most comfortable 151 CoordSysOffSet node Offset of array coordinate system see ArrayCoordSys Attributes to CoordSysOffSet X 1 Y 0 Z 0 EQ node Overall equalization in dB of the array speaker one value for each octave band Attributes to EO Octave_31 0 Octave_63 0 Octave_125 0 Octave_250 0 Octave_500 0 Octave_1000 0 Octave_2000 0 Octave _4000 0 Octave_8000 0 Octave_16000 0 ODEON will only make use of 63 to 8000 Hz octaves other bands are ignored by ODEON Any band omitted defaults to 0 dB Transducer node there should be at least one transducer in an array typically there will be more Transducer all nodes below included Attributes Description A descriptive text Gain 5 Overall in dB per octave band defaults to 0 Balloon Omni S08 Balloon data are stored in a directory known to the reader e g C OQDEON9Combined DirFiles The balloon name may include a sub directory name it probably should e g Balloon LspManufacturerName SpeakerModel23 CF2 Balloons can be in CF1 CF2 and 508 formats Delay 0 012 Delay of that transducer in seconds InvertPhase false If phase is inverted 180 defaults to FALSE SampleRat
243. or menu entry inside the ODEON program The tools allow you to expand the set of source directivity pattern files available for point sources in ODEON The ODEON directivity pattern file Version 3 or later contains information on the sound levels for the eight frequency bands 63 Hz to 8 kHz in dB for each 10 azimuth and 10 elevation These files are binary and have the extension sos An example on a directivity pattern is the pattern stored in OMNI SO8 Entering a directivity plot using the Directivity polar plot editor The easiest way to enter a new directivity plot is to use the built in polar plot editor which allows building a directivity plot from a vertical and a horizontal polar plot Since ODEON 13 a new interface has been introduced for the polar plot editor Click Tools gt Directivity Patterns gt Creat directivity pattern in plot editor The following window appears Fae Toolbar Options Tools Window Help SBS TDRELET OGG CB KAHSSEEt 4 SR a Azimuth 63Hz Fixed 125 Hz Fixed 250 Hz Freed 00 ts Fed 1000 Hz Fixed 2000 Hz Fixed 4000 Hz Fixed 8000 Hz Fixed lt N o m a a w Preg E 50o m50 S09 15 0 dB at 1 metre 10 id a P oo C 0 00 e3 20 o a a ww 0 00 30 a a Py ow 0 00 40 o m a D ow O 0 06 50 o a a a ow C 00 wo 60 ial a 0 00 0 00 sr 70 a a ow fF 0 00 80 o w a P w 6 owo 4 90 a a a o C 0 00 w a a xo 0 00 1 al a 09 0 00 w a a ow 0 00 a a a ow 0 00 i
244. orithms can be applied safely The following theoretical formula proposed by Schroeder gives the frequency limit above which room acoustic calculations RT Va 2000 y where RT is the reverberation time in the room inversely proportional to absorption and V is the volume As an example in a concert hall with reverberation time 2 sec and volume 600 m5 the are most reliable Schroeder s frequency would be 115 Hz so the octave band 125 Hz can be considered reliable in terms of modal overlap Practically we can accept frequencies well below the Schroeder s limit especially in large halls Experience has shown that even frequencies lower by 50 from this limit can be trustful 6 1 What can we model in ODEON Taking into account the energy based high frequency limitations we assume that ODEON calculations are more trustful in moderate to large rooms As absorption becomes higher this size can be reduced Normal rooms which ODEON has been successfully used for are e Meeting Rooms e Cinemas e Classrooms e Theatres e Offices open plan offices e Open theatres e Auditoriums e Train stations e Concert halls e Airport terminals e Operas e Noise halls factories e Big rehearsal halls e Industrial spaces e Small scale city plans outdoor modelling e Cathedrals worship spaces e Restaurants e Shopping malls e Multi floor buildings with atriums 65 In the following sections an overview of the calculation methods in ODE
245. pt to discover the source code of the software or the creation of derivative work based on ODEON software is strictly prohibited In no way may you transfer assign rent sublicenses lease sell or otherwise dispose of any portion of the software on a temporary or permanent basis TRANSFER RESTRICTIONS The ODEON software is licensed only to the licensee In no way may you transfer assign rent sublicenses lease sell or otherwise transfer any portion of the license on a temporary or permanent basis this includes transfer between branches of international companies that are registered in different countries The license may only be transferred to anyone with the prior written consent of ODEON A S Any authorized transfer of the ODEON software shall be bound by the terms and conditions of this Agreement TERMINATION This license is effective until termination This license will terminate automatically without notice from ODEON A S if you fail to comply with any provision of this Agreement Upon termination you shall destroy the written materials and all copies of the ODEON software including archival copies if any NO LIABILITY FOR DAMAGES Neither Odeon A S nor anyone else who has been involved in the creation production distribution or delivery of this program shall be liable including without limitation damages for loss of profits business interruption loss of information incorrect results recovery of data or other pecuni
246. r The noise suppression and echo cancellation options for the recording device microphone should be deactivated Measure Impulse Response Wensure iipulse responses Comusondal sweep sos x The Measure Impulse Response SHIFT CTRL L window ae son El me can be accessed either from within the toolbar Impulse response length E gt Ay button or the Tools menu in the menu bar of ee d E ODEON You can specify the Sweep length A i e the duration of the sweep and the BA Asad m expected Impulse Response length which me defines the final length of the recorded oae 35 impulse response as well as the extra Gan a256 40 recording time after the end of the sweep It is adeno Meroen HAD recommended to make an initial measurement n with the default values and check the quality of the obtained impulse response before changing them You can pre listen the sweep signal by pressing the Play test signal button The level of the sweep signal can be adjusted from within ODEON using the vertical Gain slider Specific values can also be entered at the Gain box On the right most side of the window two indications for the output level are displayed The blue bar shows the current level captured by the microphone while the red bar shows the maximum peak value being captured so far You can pre listen the sweep signal by pressing the Play test signal button gt The Nyquist theorem states that sampling rate sho
247. r any copy thereof All rights are reserved by Odeon A S GRANT OF LICENSE In consideration of payment of the License fee Odeon A S grants each customer a private person or a company registered in a country a non exclusive right to use the ODEON software This Agreement grants you the right to make copies of the software for archival purposes and to copy the software on the hard disk of your computer The software may be loaded onto multiple computers but can only be fully operational on a single computer at any given time The computer operating the ODOEN software needs to have the hardware key delivered with the software attached to the USB port HARDWARE KEY Granting of a license is bound to ODEON software use with a hardware key The hardware key holds the license information and determines the version and edition of the Odeon software that can run when the hardware key is attached to the computer If a hardware key is defective during the warranty period first licensed year Odeon A S will replace the hardware key The customer will pay the hardware key fee for replacing the hardware key should it become defective after the warranty period The hardware key contains the name and country of the licensee which are displayed in the software on print outs from the software and on graphics exchanged from the software LICENSE RESTRICTIONS The modification adaptation translation reverse engineering de compiling disassembling attem
248. r better auralisation option than previous versions for presentation through 2 loudspeakers called Super stereo With Super stereo you can make 2 loudspeakers have a much more spacious sound with the right frequency response for this type of auralisation In Chapter 2 an intro on how to do this can be found As an example of use if you are both musician and acoustician you can make an auralisation of a recording of your own music played in a room you have modelled in ODEON so it sounds right on your own loudspeakers Some outdated directivity files such as TlkNorm so8 have been replaced by TlkNorm_Natural so8 which is smarter for auralisation use where there might be a risk of adding the overall frequency response twice to auralisation output once from the directivity pattern and once more from the auralisation signal which inherently includes the same source spectrum please see chapter 5 for further information With the _Natural version of the directivity files it is possible to obtain correct equalization of auralisation output while also achieving correct prediction of SPL Therefore the _Natural versions of the directivity files should be used when defining new sources The old versions of the files are kept in old_so8 a subdirectory to the Dirfiles directory If you wish to use the old directivities in old existing projects then open the Source receiver list and click the Repair broken directivity links button shortcut Ct L See Chapter 1
249. r room 52 e Normal is the default value which results in default handling of scattering and diffraction taking the Reflection Based Scattering method into account if it has been enabled in the Room setup e Exterior forces a surface to be handled as an exterior surface even if it was not detected as such by ODEON the result is that less diffraction is applied at the lowest frequencies where offset in the wall is not sufficient to result in low frequency diffraction e Fractional should be used for surfaces which are fractions of a bigger whole e g surfaces being part of a curved wall or a dome should not cause diffraction due to their individual area that is the individual surfaces do not provide any significant edge diffraction When setting the type to fractional the surface area used for calculating the Reflection Based Scattering Coefficient is determined from the box subscribing the room if the construction part which the fractional surface is part of is considerable smaller that the room box the scattering might be underestimated and a higher scattering coefficient should be assigned to the surface e Transmission is for walls which transmit sound to another room When this wall type has been selected it is possible to edit the transmission data to specify the reduction index transmission loss and to link the wall with another surface in case the wall is composed of two parallel surfaces with a distance between them and possibly w
250. random colouring at will using the R shortcut If the model contains layers it is also possible to show the layer colours using the Ctrl L shortcut Do note That 3DOpenGL may occasionally fail to display complicated surfaces including numerous holes correctly typically surfaces created by CAD software using solid modelling techniques and subtractions although the surfaces are perfectly legal with respect to ODEON In these rare cases you may assure yourself that the model is in fact correct by putting point sources at various test positions and conduct tests using the 3D Investigate Rays and 3D Billiard utilities al 3DGeometry debugger Overlapping and warped surfaces should be avoided in the room model specified in the geometry file but a certain amount of overlap and warp by default 50 mm is allowed without generating a warning By overlapping surfaces is meant surfaces which define a part of the same plane in space In the simple case this can be because the surfaces are simply duplicates another case could be a door which has been defined in the same plane as the wall in which it is mounted Overlapping 41 surfaces should be avoided because it will not be clear which absorption coefficient should be applied at a reflection in case of overlapping surfaces with different materials Warps can lead to holes in rooms at edges of joining surfaces with erroneous results as a consequence and the surfaces will not be well defined Using
251. ratio of the room as well as the size and number of surfaces in the geometry In short this means that ODEON will suggest more rays for very long room with many surfaces than for a basically cubic room with few surfaces The suggested number of rays will be sufficient for many rooms however in some cases more rays may be needed in order to obtain good results in particular in rooms with 1 Strong decoupling effects 2 Very uneven distribution of the absorption in the room Ad 1 If a dry room is coupled to a reverberant room then more rays may be needed in order to estimate the coupling effect well An example could be a foyer or a corridor coupled to a classroom If the room where the receiver is located is only coupled to the room where the source is located through a small opening then more rays are also needed Ad 2 In some rooms the reverberant field in the x y and z dimensions may be very different An example of this could be a room where all absorption is located on the ceiling while all other surfaces are hard Another example could be an open air theatre where the ceiling is modelled as 100 absorbing In particular if surfaces are all orthogonal while having different materials in the x y and z dimensions of the room and if low scattering properties on the surfaces are used then more rays should be used 93 More rays needed There are no way of telling if more rays are needed for a certain calculation but to get an
252. rce should be a perfect Dirac impulse too arriving at a time equivalent to the distance between the source and the receiver However in reality the direct sound and all the other reflections are not perfect Dirac impulses In fact each reflection consists of an onset and some decay As the frequency gets lower the decay of a reflection may overlap with the onset of a subsequent reflection In the derivation of many of the ISO 3382 parameters it is vital to capture the amount of early energy correctly ODEON uses advanced algorithms for successfully detecting the energy from the direct sound and discriminating it from the following reflections For every impulse response the correct onset time is indicated in the display According to the ISO 3382 standard ISO 3382 1 2008 the onset of the direct sound and of the whole impulse response should be counted at least 20 dB below the peak level of the direct sound This dB value is called Trigger level In many cases a lower value for the Trigger level may be desirable and this is the reason why this is an adjustable parameter in the ODEON measuring system Detection of Noise Floor Most of the impulse response recordings whether recorded directly or obtained using the sweep method come with a noise tail due to the ambient background noise and noise of the transmission line involved PC sound card cables and microphone This noise tail should be excluded before deriving the decay curve and the ISO 3382 r
253. re tT rrr creer 39 2 So aerae a e sxe ealc Bi Lat E E ens ee enn ne E E E ee nna Te ee oo ener ere 41 3 Sources and Receivers isccsccicccccees sipsusnssesier a riantesesacceeeestuasincerienserp signerede 43 3 1 TSC ON rcs ance E E 43 3 2 PN ee SOU CC e E E tans deepen A A E N A E 43 3 3 Common loudspeaker format CF1 and CF2 16S i secsastassnssuussunstsccessndunesnescesdarandvaiennesiteeneass 44 3 4 Creating new directivity patterns in the ODEON So8 format eee eee eeseeeeeeeeeeeeens 45 SLE nee ee 49 4 1 Material Library Right side of Material List WindOW eee eeeeeeeeereceseeeeeeeseeseaeeeseeenees 49 4 2 Surface List Left side of Material List window ccccccccesssseceeesssneeeeceseseeeeeesesseeeeeessenaeeees 51 4 3 Manage material library and material list eee eeceseeseeceeeeseeeseeceseseseeeseecsaeeeaeeenees 53 4 4 Opening an existing room form a version earlier than version 10 for the first time 54 Syme eT 1 0c a 55 5 1 Miil istening to Binaural Room Impulse Responses eee eeseeseeseeeseeeseeeseeseaeeeeeesees 55 52 Making offline auralisations creating WAV files oe eee cere ceseeeeeeceseceseeeseessaeseseeenees 62 6 Calculation Principles ccc ccssecsssscsssssssssseesssessseesseessseessseessn 65 6 1 sadavclarersiele cou colere a Rigi ODEON prem enee errrren rrr ter tr tenn rrr errr AA 65 6 2 Gopala orar vate iq 6 lel eee nrrere cr SP rre tr ee E rrr ent Pr veer era rrr ar errr ere erat
254. rectiviy oses Selected bard bd Rightoten _Battonr gt Top Back Front Tops Len eom Geineasas Mains dneni rge ani p Apply caliration Freqercy 63 128 250 s00 1000 2000 4000 300 Hz as Total Powes Spectum oo aco SCO ome ame ows aws ame a Omaisen 100 nases Create file Verhoal am Natural source Improved Reflector Coverage Interace Attractive visualization features have been added in the reflector coverage display Rays from source receiver to reflectors audience can be displayed with varying density and colors according to layers 11 Miina ACIS Ke ACT SS VATA FIZ D A Le ri IE 174 a i i N i ir j i ai A P kas 4 z ae iy ba rr i he a e Se n y Se et ee i gt e e e p Tt i Aw A ety A s S LD Teh N Zn ay hi T N N PERRI TA S Ny rh A amp iy 4 PR ANAA A NAN Be j iN ee RIES Sy a AG Yi MN Be HIN SIY SAAN SON y S ANAR AON Ss N BIN ellis CH NANN h A ya Material Optimization Tool Often acoustic refurbishment is needed in an existing room that suffers from bad acoustics The room acoustician is asked to make a virtual model of the existing room Materials are not perfectly known for all surfaces so calibration or fine tuning is needed in order to ma
255. reflection order Job calculations only Max reflection order is a stop criterion which determines how many times a ray can be reflected Under normal conditions it should be as large as possible In that case the Impulse response length will be the actual stop criterion and Max reflection order is only taken into account when stopping rays that has been trapped between two very narrow surfaces This number may be decreased if you are only interested in the very early reflections The Maximum allowed reflection order is 2000 Impulse response resolution Job calculations and Global Estimate The Impulse response resolution is the width of the time steps in the Impulse response histogram in which the energy of the reflections are collected during a point response calculation The histogram is used for calculation of reverberation parameters such as EDT and Tso A resolution of approximately 3 ms is recommended Early reflections Transition Order Job calculations only The Transition order TO determines at which reflection order ODEON changes from the early image source method to the late ray radiosity method In special cases you may want to alter the parameters to conduct special investigations E g to investigate a high order echo problem you may want to increase the Transition order to get the reflections displayed in the Single point responselreflectogram display Auditorium and Combined editions only However this will rarely result
256. remenrt was too short in this case make a new measurement with a longer impulse response length e The S N was too poor i e the decay disappears into the noise floor in this case increase the level of loudspeaker and or sweep length In the following impulse response shown both in pressure and squared mode the signal to noise ratio is not sufficient According to the ISO standard 3382 the measured impulse response should be valid at least 10 dB below the lowest decay level required to derive the decay parameter For example if T30 is the parameter corresponding to the lowest decay level the dacay range should reach 10 dB below 35 dB resulting to a total of 45 dB decay range Results may be fairly reliable as ODEON tries to compensate for the noise floor In the following example the decay range is less than 45 dB at 1000 Hz and therefore Tso cannot be reliably derived 122 Decay range less than 45 dB 20 09 dB at 4000 Hz Measured response TCA Errors Auditorium21 Insufficient SignaNowe SR 2000ms wav Raw Impulse response at 1000Hz Raw decay curve at 1000Hz Decay curves at 1000Hz T 30 seconds Decay curves all bands Energy parameters Parameter curves C Errors Auditorium21_Insufficient_Signal2Noise_S2R2_2000ms wav Ray Impulse response at 1000Hz 7 Ee SEES eae E e E E E E i ooo aae alee Op wee E e E S OL mee Time seconds Impulse response of insufficient signal t
257. res Centimetres Millimetres Inches Feets and Yards Or if you need a different unit simply type the scaling factor from your unit into metres e g Unit Inches corresponds to Unit 0 0254 The Unit statement may be used more than once in the same par file HHH Unit Inches e g imported model data in inches Unit 0 57634 model data measured on a paper drawing which appeared in an odd unit Unit Metres model data appended in the ODEON editor modelling environment sit is most practical to use metres as the unit when modelling in the ODEON senvironment then coordinate values will be the same in the Editor as inside the 3DView in ODEON HHH CoordSys statement The CoordSys statement is used if you wish to redefine the orientation or the coordinate system in which the geometry was modelled The statement is typically used if the geometry was by accident modelled in an orientation different from the one assumed by ODEON or if it was imported from a CAD drawing where the orientation may also be different To obtain the easiest operation inside ODEON the following orientation should be used using a concert hall as the example e X axis pointing towards the audience e Y axis pointing to the right as seen from the audience e Z axis pointing upwards The syntax is CoordSys lt X gt lt Y gt lt Z gt 169 where X Y Z indicate which axis should be used as the x y and z axis inside ODEON X Y and Z may al
258. ring becomes zero initial investigations suggests that edge scattering can be assumed to zero when the distance to the edge is greater than approximately one wave length therefore we define the edge scattering coefficient as O for d x COs 0 C edge Ff i d xcosx f C 0 50 8s R for d xcos0 lt C edge As can be seen scattering caused by diffraction is a function of a number of parameters some of which are not known before the actual calculation takes place An example is that oblique angles of incidence lead to increased scattering whereas parallel walls lead to low scattering and sometimes flutter echoes Another example is indicated by the characteristic distance a if source or receiver is close to a surface this surface may provide a specular reflection even if its small on the other hand if far away it will only provide scattered sound sa 1 fw Log frequency Energy reflected from a free suspended surface given the dimensions I w At high frequencies the surface reflects energy specularily red at low frequencies energy is assumed to be scattered blue fw is the upper specular cut off frequency defined by the shortest dimension of the surface fl is the lower cutoff frequency which is defined by the length of the surface 795 Boundary walls and interior margin As long as surfaces are truly freely suspended surfaces they will act as effective diffusers down to infinitely low frequencies For surfaces
259. ringer 1994 5 30004 0469 9 000 Norma AUDIENCE AUDIENCE 0 28 Q a 2043 Goskariorgiatte 38 mm cick bgyin 2 400 men wor starrer Wand hinterlegt mit 30 mm Mineralfasermatte 1 05 kg m 2 7 5 Rayioe Ref M Heck H A Miller m A EEA A AAA rE 2044 Doppelferster ssen Ref M Heck H A Muller Taschenbuch der Technischen Alustk LE 1934 aam oe si aes p so akai z5 Aaii AAA aA sem Lochdurchmesser mien thet 12Rayl Mneraifasermatte 1 05 kg m 2 7 Rayt 30004 0 68 gee Normal a AUDIENCE vs P Metalipareele aus 0 5 mm Alublech 85 ren bret free Schitzbreite maden Panelen 15mm 164mm Abstand vor starrer Wand hinterlegt mit 20 mm Mneraifaserplatte 2 5 004 0 6 9 000 Normal AUDIENCE AUDIENCE 2 00 Pitre Publ auf Holzstuhien 2 Personen pro m Ref M Heck H A Miller Taschenbuch der Technischen Akustk Springer 1994 3004 0 550 9 000 Normal AUDIENCE AUDIENCE 0 28 Publikum auf Holzstuhien 1 Persan prom 2 Ref M Hedd HLA Miller Taschenbuch der Technischen Akustik Springer 1994 30004 0 650 9 000 Normal AUDIENCE AUDIENCE 3 08 2049 Pubikum auf maflag genolsterter Gestuhi 0 85 m x 0 63 m Ref M Hedd HLA Miler Taschenbuch der Techrischen Akustik Springer 1994 0004 0 680 9 009 Noma AUDIENCE AUDIENCE 0 27 Phin Pudim auf malleg gepoistertem Gesti 0 90 x 0 55 m Ref M Heck HA Miler Taschenbuch der Tectrischen Akustik Springer 1994 70 004 0 650 0 000 Normal AUDIENCE AUDIENCE 2 16 a 2051 Gestuhi unbesetzt malig gepolstert 0 90 m x 0 55 m Ref M He
260. rocess Syntax for the Layer statement Layer lt Layer name in quotes gt lt R intensity gt lt G intensity gt lt B intensity gt or as another option Layer lt Layer name in quotes gt lt LayerColour gt lt Layer name in quotes gt A descriptional name which must start and end with a quote sign lt R intensity gt lt G intensity gt lt B intensity gt Three floating point values between 0 and 1 which together is describing the colour of the layer as a Red Green Blue intensity If using the Layer command Shift Ctrl L from within the ODEON editor the colour intensities are set by clicking the desired colour in a dialog box Do note that it s not advisable to choose a greyish colour as it may not be visible in ODEON 172 lt LayerColour gt As another option the colour of the layer can be described using one of the predefined colours Black Blue Cream Fuchsia Gray Green Lime Maroon Navy Olive Purple SkyBlue Teal MoneyGreen or White The LayerColours par example demonstrates the different colours The LayerStatement par example shows how to create a geometry on three different layers Selected surfaces can be selected for display in the 3DView 3DOpenGL and the Materials list LayerStatementRoom Par HH const L 40 const W 30 const H 3 const NumColX 4 const NumColY 3 const ColumnW 0 3 MTranslate 1 2 00 Layer Walls 1 000 0 502 0 000 orange colour Box 1 lw h tb walls in the room MPop
261. rs Parameter curves Freque Manual onset time o 3 pulseResponses S1R2_sweep4000 wav Onset snap window Alt O rve at 1000Hz Manual truncation time T Select complete signal Alt A E Measured Restore automatic truncation times Onset time Truncation time Noise floor Restore automatic onset and truncation times R Save impulse response Play Impulse Response file P Auralisation A Next frequency band Up Previous frequency band Down Copy to clipboard Ctrl C 0 1 2 3 Time seconds Click on the Measured Response menu to see all available tools for editing and other functions 113 Cropping To crop an impulse response Click the left mouse button hold down and drag a rectangle around the part you want to crop Release the mouse button ODEON now zooms the cropped part to occupy the whole window Choose Measured Response gt Crop Impulse Response C shortcut Now the part will be saved as a new WAV file Odeon asks you to choose a file name and confirm By default the name of the original file is suggested but be careful to rename it first if you want to keep the original file Manual onset time In most cases the automatic onset time by ODEON is totally acceptable However since ODEON 13 it is possible to decide on your own onset time Choose Measured Response gt Manual onset time O shortcut The mouse cursor becomes a cross symbol when hovering on the impulse response Click
262. rs are for the expert user of Genetic Algorithms As a beginner you can leave the settings at their default values Otherwise you can read the green box with the title Expert s parameters Editable list of materials The list of materials as they have been assigned in the material list are displayed here The materials can be edited and their absorption spectrum is illustrated directly on the right side If absorption or measurement data are not well defined or measured for specific bands you can exclude them from the search process by unchecking the associated column header The allowed variation for each material during the optimization process is defined by the Search Range value When this is 0 it means that the material will not change at all while when it is 100 any absorption coefficient value can be assigned When we like to exclude a material from the search process to reduce the search space of the algorithm and to increase the speed all we have to do is to make the Search Range value 0 When calculations start such a material stays inactive 134 Warning Assigning Search Range 100 might be attractive because it allows any adjustment for a material and leads to a very good matching between simulations and measurements However the optimized material might be unrealistically different than the original material Imagine beginning your search with a carpet material on the floor and ending with a variation of glass Extra func
263. rt s parameters Then they start to evolve the next generation From now on the generation of new individuals is not random This procedure will continue until you press the Stop Calculation button During the whole evolution process all graphs in the interface are updated constantly The following picture shows a screenshot after about 7 generations in a calculation You can find the calibrated room in the Rooms folder with the name Auditorium21 at DTU_Calibrated oP Genetic material optimizer lt o Active jobs wti ere ee Best Fittina Last error decrease 2 No description n K oD Ong Fitness ra No desorption sei et V E Best Fitness 4No description 2S SS No deseriotion Evolution Method No descr S Fia dann om Inriduals per material a S 8 No description s 9 No description a 10 No description 11 No descrpban a 12 No description 5 Z i Stop Catadeton aaa o e ae neta aaa 63 125 250 500 1000 2000 4000 8000 Frequency Hertz Frequency Hertz Material Specification 63Hz 125Hz lt 250Hz 500Hz 1000Hz 2000Hz 4000Hz 8000 Hz Initial Abs Absorption Coefficient material 51 Coeff 0 Backwall in aud 21 frst guess 50 0 13000 0 27600 0 37000 0 14500 0 15000 0 1740 0 22600 0 47100 s488 a0 PEETER inu ui 1 10007 Sold wooden door Sobran 1973 0 13100 0 21900 0 09600 0 04600 0 07700 0 09900 0 05800 0 13500 123 03 S S Low Limit 2 3004 Wooden foor on joists Ingersiev 0
264. rtion as described in the previous paragraph e WARNING max XI 158 16 at 8000 Hz Measured response C Odeon13Combined Measurements Errors HagiaSophieS1R6 wav o e Raw Impulse response at 8000Hz Raw decay curve at8000Hz Decay curves at 8000Hz T 30 16 52seconds Decay curves all bands Energy parameters Parameter curves Frega C Odeon13Combined Measurements Errors HagiaSophieS1R6 wav Ray Impulse response at 8000Hz p x E 3 0 1 2 3 5 6 7 8 9 Time seconds Impulse response with decay of high nonlinearity 121 he WARNING max XI 158 16 at 8000 Hz Measured response C Odeon13Combined Measurements Errors HagiaSophieS1R6 wav kojas Parameter curves Freq gt C Odeon13Combined Measurements Errors HagiaSophieS1R6 wav Raw decay curve at 8000Hz E Measured Onset time Truncation time Noise floor SPL dB 0 1 2 3 4 5 6 7 8 9 Time seconds Impulse response with decay of high nonlinearity Same recording as in previous figure but squared In the squared form a normal exponential decay should look like a straight line Insufficient Signal to Noise ratio In order to derive Tz at least 35 dB of valid decay must be present for each octave Reasons for not obtaining a valid decay curve might be e The captured impulse response may be to short because the impulse response length used during the measu
265. rve at 1000Hz Decay curves at 1000Hz T 30 seconds Decay curves all bands Energy parameters E Parameter curves Freque mE C Errors Wrong_onset_ElectroMagneticF eedback_Corridor_r10 wav Raw decay curve at 1000Hz E Measured M Onset time M Truncation time SPL dB 0 1 2 3 Time seconds Electromagnetic feedback in an impulse response recording Same file as in previous figure but squared for better visualizing the dynamic range Impulse response with harmonic distortion Harmonic distortion can occur at excessive high gains in some loudspeakers It can be a significant problem not easy to track and it can pollute the entire impulse response However part of distortion is visible at the beginning of the impulse response as a built up before arrival of direct sound Always inspect an impulse response for distortion and repeat the measurement if in doupt of the quality The following figures give examples of harmonic distortion The following common enhancements in sound cards can lead to distortion products Be sure to deactivate them e Echo cancellation e Automatic gain control e EER 27 14 dB at 1000 Hz use longer sweep Measured response CAC C Odeon13Combined Measurements Errors DistortionOfficeS1R4 wav Ray Impulse response at 1000Hz Pdea aaa aaee O i of jf x anD OO a d W i o o p H i Time seconds Harmonic distortion visible as a built up
266. s which is essential to the use of hybrid statements Box Cylinder etc is implemented as the following functions MTranslate lt TranslateX gt lt TranslateY gt lt TranslateZ gt MRotateX lt Rotation angle gt MRotateY lt Rotation angle gt MRotateZ lt Rotation angle gt MScale lt ScaleX gt lt ScaleY gt lt ScaleZ gt MPop MReset The manipulations carried out by the M family are cumulative This means that you can specify more than one operation to be carried out e g first rotate 90 around the Z axis then rotate 90 around the Y axis and finally translate 10 metres upwards The following example shows these operations carried out on a cylinder shell the Cylinder statement is described later Manipulating a cylinder Par HHH MRotateZ 90 MRotateY 90 Y MTranslate 0 0 10 Cylinder 1 20 5 180 10 TB Cylindrical ceiling HHH The transformation commands to be carried out must always be stated before the points geometry on which they should work is created To reset all previous coordinate manipulations use the MReset command To reset cancel the most resent manipulation MTranslate in the example above use the MPop command which will pop the operation of the matrix stack 171 Hints The order in which the coordinate manipulations are carried out is important usually but not necessarily always the MScale commands should come first then the MRotate commands and finally the MTranslate commands If you
267. s a well known problem among room acoustic consultants So far the task had to be performed manually in an iterative process of trial and error The reason for requiring such an optimization process is that materials in an existing room are not completely known For example we might be able to know that the wall is made out of concrete but we probably cannot know the actual type of concrete and the corresponding absorption data Usually 5 to 15 materials are present in a room and most of them can be known So the optimization process is limited to adjustment of a few materials on the basis of how different parameters agree between simulations and measurements Common parameters to evaluate in a calibration procedure are Early Decay Time EDT Reverberation Time Tso Centre Time Ts and Clarity Cso Cso Reverberation time and centre time or Clarity are complementary parameters meaning when the first increases the second decreases and vice versa Among the numerous types of search algorithms the ODEON optimization tool uses Genetic Algorithms which are proven to be efficient for searching in multi dimensional search spaces multi variable problems 12 1 General Steps for Material Calibration So far any optimization process had to be done manually With the Genetic Material optimizer tool in ODEON the largest part of the process is done automatically but in terms of setting up the calculations it remains the same The general steps for a calibration
268. s can be used for exporting and importing simple array types as they are defined inside ODEON the format being compact it may in some cases even be used for typing data manually into a text file If Domain Frequency then beeming filters are entered as a number of Complex numbers per octave band We suggest that NumberOfSubBands an attribute of the ArraySource is set to 6 Attributes DomainData has no attributes Nodes to DomainData EQ or Octave_31 Octave_63 Octave_125 Octave_250 Octave_500 Octave_1000 Octave_2000 Octave_4000 Octave_8000 Octave_16000 EQ node EQ node is only present if Domain FullOctave Attribute of Transducer Attributes to EO Octave_31 0 Octave_63 0 Octave_125 0 Octave_250 0 Octave_500 0 Octave_1000 0 Octave_2000 0 Octave _4000 0 Octave_8000 0 Octave_16000 0 ODEON will only make use of 63 to 8000 Hz octaves other bands are ignored by ODEON Any band omitted defaults to 0 dB Octave_31 Octave _63 Octave _125 Octave_250 nodes 153 These nodes are only present if Domain Frequency then a number of Octave nodes are defined each of these Octave nodes have Sub nodes Nodes Sub Sub Each Octave band has the centre frequency fc defined according to ISO 31 5 63 125 250 500 1000 4000 8000 and 16000 Hz For each of the octaves there are Ns number of Sub bands nodes of the Sub type the center frequency of the sub bands are defined from the f
269. s eXtending the data being exported and imported when needed by adding new attributes to nodes and by adding new nodes e Parsers for writing and reading the format are available in most or all modern programming environments e Even if a parser is not available to the persons exporting data e g Loudspeaker Manufacturer it is fairly easy to write XML formatted data the import part which is more complex is left to ODEON or similar programs The layout of the XML file To understand and learn the format the XML files installed with ODEON e g in the C QDEON10Combined ArrayXML should be studied This document explains details of the format but most of the format is best understood from reading through the sample files The node or tree structure of XML files is complicated but best understood by studying the sample files and look in this documents when more details are needed See figures on next page to get a picture of the tree structure The following XML sample files are installed with ODEON Monopol xml Dipol xml Quadropol xml Octopol xml Dipol_Domain_Frequency xml The names of the first four files are almost self explaining those are the classic textbook examples built from 1 2 4 and 8 sources These sources utilize the Omni so8 directivity pattern for all their transducers The four sources are all conventional arrays which have been assembled inside the ODEON Array source editor and exported from there Only
270. s feature Remember If keeping an earlier version be careful not to mix the use of old and new versions although we do strive to maintain forward compatibility we cannot guarantee that a room which has been loaded into a new version of ODEON will also load into an older version without problems Upgrading from version 6 and earlier If you upgrade from a version earlier than 6 then we do recommend that you read carefully through the manual as if you were a newcomer to ODEON There are a substantial differences between the early versions of ODEON and the ODEON software as it is today modelling has been made easier calculation principles has been enhanced and a huge amount of new features has been added Project files The only project file from versions earlier than 3 0 being fully compatible is the surface file sur The rest of the project files are no longer valid And even though the sur format is still valid it is not recommended to model rooms in this format The par format is a much more efficient format Upgrading from ODEON 4 5 and 6 to ODEON 7 and later If you are having problems loading a room which was created and worked fine in one of the above listed versions of ODEON this is probably due to a change that has been made to the surface numbering mechanism applied in ODEON The numbering mechanism has been changed slightly in order to avoid a conflict which appeared when using symmetric surfaces along with mod
271. s more visible Similarly to the Raw Impulse Response both the onset and truncation times are displayed In addition a horizontal solid blue line indicates the noise floor in the impulse response see Fig 12 2 When the noise floor of the response has a high degree of fluctuation this solid line becomes dashed indicating that the noise floor is far from being flat so that the estimation is not trustful Decay Curves On this display the pure part of the impulse response between the onset and truncation times is processed further Three types of curves are displayed in the graph for each octave band e E Measured This is the energy impulse response Raw Decay Curve between the onset and truncation times with a resolution specified either in the Room setup or the Measurement setup The display is fully compatible with the Decay Curves displayed in the simulated Point responses in the Single Point display in ODEON Auditorium and Combined e E Integrated This is the result of the backwards Schroeder s integration of the energy impulse response between the onset and truncation times e E Corrected This is the E Integrated curve with compensation for the lack of energy at the end of the response which is caused by the backwards Schroeder s integration Decay Curves all bands Here the corrected energy decay curves E Corrected are displayed for all bands on the same graph You can enable disable the different octave bands in the ri
272. scattering is in the following assumed to be scattering appearing due to random surface roughness 0 015 This type of scattering gives rise to S 0 TPR 0 06 scattering which increase with 8 5 0 25 frequency In Figure 6 3 typical 04 T q y 8 yp 5 08 frequency functions are shown In i x 0 9 ODEON these functions are used in 0 1 u the following way Specify scattering 0 iee m 63 125 250 500 1000 2000 4000 8000 coefficients for the middle frequency Frequency Hz around 700 Hz average of 500 1000 Hz bands in the materials list then ODEON expands these coefficients Figure 0 1 Frequency functions for materials with different into values for each octave band surface roughness The legend of each scattering coefficient curve denotes the scattering coefficient at 707 Hz using interpolation or extrapolation At present it has not been investigated in depth which scattering coefficient at the mid frequency 707 Hz should be used for various materials However initial investigations indicate that the following magnitudes may be sound 0 1 Table 1 Suggested scattering coefficients to use for various materials The given values are for the middle frequency at around 700 Hz to be assigned to surfaces in ODEON Suggestions may be subject to changes as more knowledge on the subject is obtained As has been explained i
273. se halls etc are dealt with in Beranek 1996 and Barron 1993 where different halls around the world are presented along with judgement of their acoustics and guidelines for design Short guidelines on which values to expect for Clarity and G in concert halls based on some simple design parameters as width height floor slope etc are given in Gade 1997 Some recommended values for room acoustical parameters are given in the following table Objective parameter Symbol Recommended Subj limen symphonic music 1 7 2 3 seconds Level rel 10 m free field Early Lateral Energy LFso gt 0 25 5 Fraction Early Support gt 13 dB E Total Support gt 12 dB E Recommended values for objective room acoustical parameters ISO 3382 1 2009 in large music rooms with audience according to Gade Gade 2003 Subjective limen as given by Bork Bork 2000 and Bradley Bradley 1986 Early Decay Time and Reverberation time The energies of all the reflections received at the receiver point are collected in histograms with class interval specified in the Room setup Impulse response resolution After completion of the response calculation early decay time and the reverberation time are calculated according to ISO 3382 1 2009 e The Reverberation time Tso is calculated from the slope of the backward integrated octave band curves The slope of the decay curve is determined from the slope of the best fit linear regression line
274. shortcut Alt A The anechoic recordings have been created by Helsinki University of Technology For use of the anechoic orchestra recordings delivered with the software please see in the beginning of the manual at the End user license agreement 64 6 Calculation Principles ODEON is an energy based room acoustic modeller meaning that sound waves are represented by rays A ray can be understood as a straight line connecting a source and a receiver Reflections are represented by image and secondary sources where still the sound to the receiver can be seen as a line ray These simplifications make it possible to calculate the acoustic response in large spaces while maintaining a short computation time However phase information is not included in the propagation of sound All calculations are performed in the energy domain and not in pressure domain Hence the acoustic differential equations cannot be solved directly as it happens in Finite Element Methods FEM or Boundary Element Methods BEM Calculations in room acoustics use mostly geometrical tasks and they are valid for spaces and frequencies where high modal overlap occurs that is when modes standing waves are so many that the sound field becomes smooth Single modes are prominent at low frequencies in small and reverberant rooms Fortunately a moderate sized space like a classroom is considered already a big volume where there is high modal overlap so that room acoustic alg
275. sing the specified offset on the numbers of surfaces this is explained later Example on the use of NumbOffSet creating surface 101 containing the points 101 to 104 and surface 201 containing the points 201 to 204 NumbOffSet Par HHH NumbOffSet 100 Pt 1 0 1 0 Pt 2 0 1 0 Pt 3 1 1 0 Pt 4 1 1 0 Surf 1 A surface 1 2 3 4 NumbOffSet NumbOffSet 100 Pt 1 0 1 1 Pt 2 0 1 1 Pt 3 1 1 1 Pt 4 1 1 1 Surf 1 Another surface 1 2 3 4 HHH Example creating point 1 4 and surface 1 setting NumbOffSet to Auto then creating Point 5 8 and surface 5 HHH Pt 1 0 1 0 Pt 2 0 1 0 Pt 3 1 1 0 Pt 4 1 1 0 Surf 1 A surface 1 2 3 4 NumbOffSet Auto Pt 1 0 1 1 Pt 2 0 1 1 Pt 3 1 1 1 Pt 4 1 1 1 Surf 1 Another surface 1 2 3 4 HHH 160 Defining a point using the Pt statement Use the Pt statement to define a single point The syntax must be as follows Pt lt Point Number gt lt XMathExpression gt lt Y MathExpression gt lt ZMathExpression gt Example defining point number 100 in x y z 1 1 1 Pt 100 1 1 1 Hint Point number and coordinates can be written using mathematical expressions allowing greater flexibility and reusability Parametric modelling defining multiple points Use the MPt statement to define a series of points which is typically used in connection with the ElevSurf or ElevSurf2 Statement The syntax must be as follows MPt lt Number gt lt NumberOfPoints gt lt XMathExpressionl gt lt YMathExpressio
276. sion dx dY or dz select the surface in the Surface editor table where it can also be specified whether the extrusion surface should have a bottom and a top and a Description may be entered The drawing depth and extrusion for each extrusion surface is displayed graphically at the bottom of the application window Editing or correcting an extrusion surface In order to make corrections to an extrusion surface select it in the Surface editor table and bring it into edit mode using the Insert or Esc shortcut Once in edit mode it is possible to change coordinates of the points insert or delete points and to move the surface using the mouse operations listed below It is also possible to enter the precise coordinates of points in the Point editor table which list the point in the selected surface so it is an option to draw a sketch using the mouse and then fine tune the coordinates afterwards in the Point editor table Operation on surface Mouse operation Create a new point in selected surface LeftClick mouse Select the point in selected surface which is closest to Mouse pointer Ctrl LeftClick mouse Move closest point in selected surface Ctrl Alt LeftClick mouse Move selected surface Shift LeftClick mouse Move selected surface when its not in edit mode LeftClick mouse Manipulating the viewport The viewport can be manipulated using the shortcuts listed below It is possible to make changes to the view while drawing a surface Vie
277. so have a sign to indicate that the axis should point in the opposite direction Example 1 the default orientation which is assumed by ODEON if the CoordSys statement is not used in the geometry file CoordSys X Y Z Example 2 changing the direction of the X axis CoordSys XYZ Example 3 Swapping the X and the Z axis CoordSys Z Y X Example 4 The CoordSys statement may be used more than once in the same par file HHH CoordSys X Y Z se g if the X axis was inverted in imported model data sresetting the coordinate system to the default CoordSys X Y Z model data appended in the ODEON editor modelling environment sit is most practical if using the default Coordinate system when modelling in the ODEON environment then coordinate system will have the same orientation in the ODEON editor as well as in the 3BDView inside ODEON HHH User coordinate system The UCS command is mostly there for compatibility with previous versions of ODEON the coordinate manipulation functions MTranslate MRotateX MRotateY MRotateZ MScale MPop and MReset included from version 4 21 allow far more flexibility For your own sanity it is not adviceable to mix the UCS method with the M family method The UCS command must follow the syntax UCS lt TranslateX gt lt Translate Y gt lt TranslateZ gt lt RotateZ gt The UCS command is used to create a User Coordinate System with its own X Y and Z translation It also allows a rotation
278. sources and surface sources The last two types are only included in the Industrial and Combined editions On the other hand the receiver has a very simple form being just a point mono microphone For auralisation see Chapter 5 the exact direction of the head is needed but this is something adjusted in the JobList 3 1 Generic point sources Generic point sources such as the OMNI or SEMI directional sources are typically used for calculating the frequency response and parameters characterizing the room acoustics Typically for the generic source is that it is defined mathematically 3 2 Natural point sources With natural sources we refer to sources such as human voice an acoustical instrument or similar Natural sources are typically used for auralisations and or calculation of acoustical parameters that depends on a specific sound power of a natural source E g speech intelligibility of a person or sound pressure level by a smaller machine A recorded signal for auralisation is associated with the directivity pattern of the actual source present during the recording In an auralisation the directivity pattern for natural sources in ODEON is used together with the recorded frequency spectrum of e g a voice and if not handled correctly this will Signal from anechoic result in auralisation where the overall Directivity with spectrum recording with of natural source spectrum of source frequency response is included not once but
279. ssary to use as most common sources are defined already in ODEON or available on www clfgroup org The data entered in the text format should be in relative calibration across frequencies but need not be in any absolute calibration The absolute calibration is an option from within the program as described under the chapter below Creating a new directivity pattern using a text file as input The first non comment line of the input file indicates whether the data is for e POLAR set containing horizontal and vertical polar plots for sources where only a horizontal and a vertical plot are known e g a loudspeaker e FULL set for complex sources where directivity data is known for each 10 Azimuth and 10 Elevation e SYMMETRIC set for symmetric sources e g a trumpet The second non comment line in the input file indicates whether the data represents a Natural source such as a musical instrument or a person speaking or if it represent a source such as a loudspeaker which does not include the frequency response of the source signal The purpose of this Boolean is to make auralisation correct for natural sources set NATURAL to e TRUE for sources such as musical instruments 46 e FALSE for sources such as loudspeakers Each of the subsequent lines of the input file should contain sound pressure levels in dB for a complete 180 of elevation from the forward axis to the backward axis The resolution must be 10 hence each l
280. stO 10 gt 13 15 A point list can be referenced in the following way adding point 1 before and 2 after the list in this example Surf Test_surface 1 PListO 2 To reset the list use the statement list 0 used in this example ResetPList0 Multi Surface MSurf The multi surface MSurf is essentially just a variant of the Surf statement Instead of typing one header line e g Surf 1 A surface name for each surface the header can be shared by multiple surfaces MSurf lt SurfaceNumber gt lt NumberOfSurfaces gt lt Optional Description gt lt ListOfPointNumbersI gt lt ListOfPointNumbers2 gt lt ListOfPointNumbers3 gt ores lt NumberOfSurfaces gt lines with lists of points describing each surface lt SurfaceNumber gt A unique number from 1 to 2 147 483647 for identification of the surface Using the same number but with a negative sign defines the surface and its counter part mirrored in the XZ plane Y 0 The surface number may be defined using mathematical expressions lt NumberOfSurfaces gt The number of surfaces in the MSurf lt Optional Description gt A string displayed and printed for easy identification of the surfaces Could be something like Main floor Example on multi surface containing 5 sub surfaces MSurf 1 5 Steps on a stair 5544 gt 5534 S112 gt 5122 5111 gt 5101 521255222 S211 gt 5201 5312 gt 5322 5311 gt 5301 5412 gt 5422 S411 gt 5401 5512 gt 5522 163 Elevation surface Use The ElevSurf
281. sults in a reflected direction which is the geometrical average of the specular direction and a random scattered direction Note Scattering is a 3D phenomenon but here shown in 2D 1 Snell s law is the law of Billiard saying that the reflected angle equals the angle of incidence 71 6 5 The Reflection Based Scattering coefficient When the reflection based scattering coefficient is activated in the room setup ODEON will do its best in estimating the scattering introduced due to diffraction whether it occurs due to the limited size of surfaces or as edge diffraction When the method is activated the user specified scattering coefficients assigned to the surfaces should only include scattering which occur due to surface roughness Diffraction phenomena are handled by ODEON The Reflection based scattering method combines scattering caused by diffraction due to typical surface dimensions angle of incidence incident path length and edge diffraction with surface scattering Each of the two scattering effects is modelled as frequency dependent functions The benefits are two fold e Separating the user specified surface scattering coefficient from the room geometry makes it easier for the user to make good estimates of the coefficients that will be in better agreement with the ones that can be measured In many cases a scattering coefficient of say 5 for all smooth surfaces may be sufficient e Scattering due to diffraction is distance and angle
282. sus orchestra etc The Offline auralisation offers greater flexibility than the real time auralisation allowing full control over which signal to pass through which of the 300 channels available and assigning individual level and delay to each channel If for some reason you need the auralisation output as a wave file it is also the offline auralisation which should be used Warning The auralization toggle button is disabled if Create binaural impulse response file Jnn or Create impulse response SurRoundnn is disabled in the Auralization Setup Ie SHIFT CTRL A 62 Single channel simulation First try to create a one channel J JobList Binaural mode Headphone Subject_021ResSdeg_diffuse wav 1 1 1 Convolve Binaural RIR with Signal file simulation OF a person speaking from a Convine Enabled Signal Sub Path Signalfile Time Channel Calib Job no Recewer Jobdescr source position Open the room Elmia 4 easi 2 2 W hra ni a Average none RoundRobin2 detailed par which is stored in 3 E Vioin Bach Average none a E Violin Boccherini Menu Average none the ODEON Rooms folder see section s E Von Mendelsohn Vk jem ane li i 1 6 E ee Average none 2 1 In the auralisation display select eas i R v nnne Averane fnane the Conv no 1 row and select the Voice Sabine Short file in the signal file field this is an anechoic recording of voice stored in a Windows Wa
283. t definition early 0 50 ms to total energy ratio Cso dB Clarity early 0 80 ms to late 80 energy ratio 1 dB abs G dB Sound level related to omni directional free field radiation at 10 m 1 dB abs distance LF Early lateral 5 80 ms energy ratio cos lateral angle STI RASTI Speech Transmission Index 0 03 abs Room acoustical parameters and their subjective limen as given by Bork Bork 2000 and Bradley Bradley 1986 Centre time time of first moment of impulse response or gravity time Example 1 If the real G value is 1 dB and the simulated is 1 9 dB then the difference is not noticeable Example 2 If the real LF value is 12 and the simulated is 17 the difference is just noticeable Note When comparing measured parameters to the ones simulated it should be kept in mind that the measured parameters are not necessarily the true ones as there are also uncertainties on the measured result These errors are due to limited tolerances in the measuring equipment as well a limited precision in the algorithms used for deriving the parameters from the measured impulse response or similar errors if results are not based on an impulse response measuring method There may also be errors due to imprecise source and receiver positions 92 9 1 Sources of error There are many sources of error in a room acoustical simulation leading to results which are less than perfect within one subjective lim
284. t F1 Exercise Importing a dxf file and changing its origo Try importing the file Elmia RoundRobin2 detailed DXF which is located in the rooms directory in your ODEON installation To make the operation of ODEON as smooth as possible it is desirable to move the Origo of this geometry Once the geometry has been imported this change may be made as follows Investigate the coordinates of the front edge of the stage 1 Turn on the modelling options in the 3DView shortcut M and move the mouse in order to investigate the corner s coordinates 2 If pressing the Ctrl key while Left clicking the mouse then the data for the closest corner is copied to the clipboard the data text can be pasted into ODEON s editor using the Ctrl V shortcut 31 Pasted corner data from the 3D View Pt 248 10 500 5 90000 24 00000 for a left point on the stage Pt 47 10 500 5 90000 24 00000 for a right point on the stage We may want to locate Origo at the front of the stage This can be done using the Mtranslate statement in the geometry file in order to move the mid point average of the two points above of the stage to 0 0 0 open the par file clicking the ODEON editor icon then just after the sign type MTranslate 10 5410 5 2 5 94 5 9 2 244 24 2 At the end of the file just before the sign type MReset in order to make the coordinate system neutral this is desirable when adding new surfaces to the g
285. t be a good solution Apart from the three constantly updated curves two more are displayed fixed e The Low limit specifies the lower values the corresponding material can be assigned per octave band during the optimization process It directly depends on the Search Range value e The Upper limit specifies the higher values the corresponding material can be assigned per octave band during the optimization process It directly depends on the Search Range value Tip For a direct comparison of different materials uncheck the Auto scale option Then 0 becomes the minimum value and 1 becomes the maximum one Run a calculation Now it is time to run a first calculation Press the Start Calculation button ODEON starts a preliminary run during which the original fitness values are evaluated red bars in Best Fitting graph Depending on the problem this step can be extremely fast almost no noticeable 135 Just after the preliminary run the main Genetic Algorithm starts with the O generation that involves generation of completely random individuals list of materials However this random search is always restricted between the low and the upper limits specified by the search range for each material For example a material that has been assigned a search range of 5 it will remain almost as the original even during the random 0 generation After the 0 generation is finished parents are selected with one of the methods described in the Expe
286. t file to sales odeon dk you should write the user name you wish in the email Requested upgrade relevant if edition is Basics Industrial or Auditorium In case your current edition of ODEON is ODEON Basics it is Default setting is None possible to request ODEON Industrial ODEON Auditorium or ODEON Combined In case your current edition is ODEON Industrial or ODEON Auditorium it is possible to request ODEON Combined It is not possible to upgrade from ODEON Industrial to ODEON Auditorium Time license relevant if the current version is time limited No update to time license This is the default setting Update time limitation If current license is time limited then it is possible to update the time license i e to request additional run time by entering 18 Remove time limitation the desired number of hour in the box on the right Common case for a time limited license is when you purchase a full license that it is always fed with a few hours for a start until delivery of dongle and complete payment is fulfilled When payment has been executed we automatically send you an update file that removes time limitation If however a few extra hours are needed before payment is completed you can request extra time with this procedure If the current license is time limited it is possible to request the time limitation to be removed 19 2 Modelling Rooms Creating new room models is probably the most time consuming task
287. t one of the cells that have to do with the Job like Job no Receiver Job descr in order to listen to the convolution result 63 Extended Auralisation examples with complete orchestras ul xX Since ODEON 10 a number of auralisation setups are available where entire orchestras have been modelled for auralisation in a concert hall and in an opera The installation includes music pieces of Beethoven 21 channels Bruckner 47 channels Mahler 39 channels and Mozart 11 channels To listen to these auralisations e Open one of the rooms in the Orchestra folder e g C Rooms Orchestra ComptonVerneyOpera par e Open the JobList shortcut Ctri Shift J e Calculate all jobs in the JobList shortcut Alt A e Once the calculations have finished in the Auralisation Display play the various tracks shortcut Alt s These auralisation samples include many setups that tells ODEON how jobs and wave files should be combined into a few final mixes It is likely that you will want to listen to auralisation in other receiver positions than the predefined ones The fastest way to do this e Make a copy of the file set using the File gt Copy Files menu entry just include room files when prompted e Enter the Source receiver list Shortcut Shift ctri s and change the coordinates of the predefined receiver double click the receiver in the receiver list e Open the JobList shortcut Ctrl Shift J e Calculate all jobs in the JobList
288. t the decay of low frequencies is captured within the recording of the whole sweep response and harmonic distortion is suppressed An extra decay period of a few seconds is recorded after the end of the sweep so that the remaining decay for middle and high frequencies is captured Octave Band Filtering The broadband impulse response obtained by the sweep method is filtered in octave bands between 63 Hz and 8 kHz using 2 4 order Butterworth filters according to the IEC 61260 1995 08 standard These analog Butterworth filters are implemented by digital infinite impulse response filters IIR filters which introduce unwanted transients in the beginning of the response Although the length of the impulse response of the filters is infinite a finite effective length is used so 99 9 of the energy of the filtered response is included The length of the transients in the filtered response is taken equal to the effective length of the filter Moreover reversed filtering is applied for decay analysis so that all the transients appear at the tail of the impulse response instead of the beginning and the filter phase distortion is suppressed ODEON automatically excludes the transient tail when processing the impulse response Detection of Direct Sound An ideal impulse response according to geometrical acoustics Kuttruff 1973 would consist of an ensemble of Dirac impulses with appropriate delays and strengths The ideal direct sound from a sou
289. tch simulations with measurements ODEON 13 has been equipped with an optimization tool based on genetic algorithms An initial estimate of materials is required Then the algorithm searches for the best set of absorption coefficients by a genetic evolution process that makes simulations match with measurements Eros Crossover method Geneexchange 0000000110011000110011000 Best Fitting Last error decrease 00010000 110100000111000 F z Frequency optimization IndwdusiBands Score 11 4431523797763 a o Orig Fitness 26 599 51 68 a an 31 66 v IB Best Fitness 21232 19 376 Avg Evoluton Method E Fitness 33 7333849575143 2 948 Seconds innaduals per material 12 Fitvess 11 4431523797768 cy 1 406 Crossover_Probabilty 93 9 Fitness 77 0416722127369 1507 1238 0695 o993 0977 lt Inversion Probab ty 30 0 0916 0 934 0 681 Average error in JND s oer NW fF AONI Mutation_Probabiity 19 Eltist percent 50 0 Copy range to other materiais 63 125 250 500 1000 2000 4000 8000 Frequency Hertz 63 125 250 500 1000 2000 4000 8000 Frequency Hertz 63 Hz 125 Hz 250 Hz 500 Hz 1000 Hz 2000 Hz 4000 Hz 8000 Hz l l 0 45 _ Plasterboard on frame 13 mm 0 43500 0 42500 0 19200 0 04300 0 05000 0 09500 0 05900 0 08100 2 a4 Initial boards 100 mm cavity filed with Optimized 1 102 Smooth concrete panted or gared 0 02500 0 02900 0 01300 0 0270
290. te for that Two step calibration Often the foregoing methods of calibration can be impractical for several reasons measurement can be performed much later than calibration and at a distant location In none of these cases can be guaranteed that the equipment settings remain unchanged The two step calibration tool added in ODEON 13 removes the requirement of fixed settings in the equipment The procedure adds a second step to the normal diffuse chamber or free field calibration methods in order to capture the spectral properties of the omni directional source used for the measurement Having these properties allows the user to change the settings in the equipment as long as the source remains the same The tool is available at Tools gt Measurement calibration gt Diffuse calibration gt Two step for a calibration when the main step is in a diffuse chamber and at Tools gt Measurement calibration gt Free field calibration gt Two step when the main step is a free field calibration 11 4 Measurement set up In the Options gt Program setup gt Measurement setup important parameters can be adjusted concerning the measurement signal the post processing of an impulse response the calibration and your sound equipment Parameters for detection of onset time The Noise floor window length specifies the time interval used to detect the noise floor before the onset time of the impulse response When a sound file contains more than one impulse responses in
291. th the installation of ODEON whether a value of 10 or 20 centimetres is chosen may not be critical but for rooms with a very jJumpy boundary it should be considered to specify this parameter Key diffraction frequency Default is 707 Hz in order to obtain the best result in the mid frequency range for speech and music This is the frequency at which diffraction is calculated for the ray tracing part of calculations All other parts of point response calculations take into account frequency dependent scattering Only in special cases where the focus is on another frequency range should this frequency be changed Scatter coefficients gt x xx handled as uniform scatter From published material on measured scattering coefficients there seems to be a general trend that modest scattering tends to be area based whereas high scattering is better represented by uniform scatter In ODEON small scattering coefficients i e below x xx is handled either with the Lambert Oblique or Lambert algorithms whereas scattering coefficients above x xx are handled using uniform scattering Scattering coefficients in the context above are the Reflection Based Scattering coefficients if that option is activated a typical result of this algorithm is that reflections close to an edge will be handled with uniform scattering which is desirable Through empirical studies we have found that x xx 0 5 yields best results If using a value of 0 then all reflections are handled
292. the Room setup so that the measured energy response can be compared side by side to the simulated energy response obtained in the Single Point Response External adjustments The overall Gain for the sweep signal playback can be adjusted here 11 5 Editing Impulse Responses Three functions are available for editing an impulse response Cropping manual onset time and manual truncation time All these options are available in the first two tabsheets Raw impulse response and Raw Decay curve Tip The Raw Decay curve display is more convenient for editing an impulse response as all values have been converted to dB leading to a better view of the dynamic range In addition all reflections become positive same phase and they are displayed in the upper part making it easier to decide what is the beginning and the end of the response In the next figure it is shown how to access the editing functions from the dynamic menu on the main menu bar See below for analytic descriptions of the editing functions Measured response Options Tools Window Help Add measured Parameters to Multi point job hs R H 4 y E Bey lie OF Restricted en E i nocan Add measured Parameters to Multi point job and close Ctrl Ins Invert zoom z ae aT F onse C Measurements Auditorium21_ImpulseResponses S1R2_sweep4000 wav bo la Em rop impulse Response emp p at 1000Hz T 30 1 89 seconds Decay curves all bands Energy paramete
293. the CAD program Using the ELEV command at least this is true in IntelliCAD and AutoCAD makes it possible to convert parts of a flat line drawing into a 3D drawing typically a 2D floor plan can be converted into a set of vertical walls Use the CHANGE command in order to change elevation and height of these entities from within the CAD program 3DPOLY As an option it is possible to import 3DPOLY 3D polylines as if they were surfaces when these lines are closed polygons When ODEON exports surfaces containing more than four points these surfaces are exported as 3DPOLY lines 3DPOLY lines will not respond to the HIDE or the RENDER commands when imported into e g AutoCAD however it is possible to convert POLYLINE s into REGION entities which are visualized correctly as surfaces in some CAD programs if the 3DPoly s are not plane this may not work In some cases it may be desirable to switch this import option off when importing to ODEON as the DXF file may contain such entities which the modeller did not intend to be included in the 3D surface model to be imported the entities may have been modelled for other reasons e g as assisting lines in the modelling phase POLYLINE when the POLYLINE is closed and the elevation height is set to zero This entity is not really a true surface however in some cases it may be used by some CAD programs including AutoCAD in order to bypass the limitation of maximum four points in a surface If a ge
294. the appropriate point source activated in the particular job 86 ISO 3382 3 Open plan offices parameters For these calculations the source is required to be assigned the S03382 3_OMNI SO8 directivity file The following parameters are supported Spatial distribution of STI This curve shows how the decrease of speech transmission index as a function of the distance from the source Distraction distance rp Distance from the speaker where the speech transmission index falls below 0 50 Above the distraction distance concentration and privacy start to improve dramatically Privacy distance rp Distance from the speaker where the speech transmission index falls below 0 20 Above the privacy distance concentration and privacy are experienced very much the same as between separate office rooms Spatial decay rate of A weighted SPL of speech Dzs This curve shows the decrease of the A weighted sound pressure level per distance doubling from the sound source which emits noise with the sound power spectrum of normal speech A weighted speech at 4 meters Nominal A weighted sound pressure level of normal speech at a distance of 4 m from the sound source The levels are summed up over the octave bands from 125 to 8000 Hz A weighted background noise The quantity shows the level of background noise summed up over the octave bands from 125 to 8000 Hz Fy Edit Room Acoustic Parameters in ODEON In the latest ODEON version the
295. the average deviation of a number of room acoustic parameters is calculated The parameters are normalized to their JND Just Noticeable Deference ISO 3382 1 e g 5 for reverberation time and 1 dB for Cso so it is possible to merge different parameters into one fitness number If the difference between measured and simulated parameter is less than 1 JND e g less than 1 dB for Cso this is fairly accurate as it is not possible for the human receiver to perceive the difference subjectively The fitness function used is given by the following formula K I gt Par loin Par Iya JND K I where is the fitness value error between simulated and measured value in JNDs and Par ay Par leas represent the simulated and measured acoustic parameter i for the source receiver combination k K is the total number of source receiver combinations while I is the total number of used acoustic parameters Table 1 Terms used in GA and their interpretation to the acoustic optimization problem Analogue to material optimization Absorption coefficient of a material for a specific band Set of genes materials that characterize an individual Individual A candidate solution that consists of a list of genes materials associated to a chromosome Population Ensemble of Individuals all different material combinations for one generation Generation A stage in the evolution process corresponding to a population
296. the parametric format described in Appendix E E g when modelling geometric shapes such as cylinders and domes or when combining 2 extrusions in different planes or when it is appropriate to describe parts of a geometry using parameters The par file can be edited in the ODEON text editor ODEONEdit The 3DView available in ODEON is a useful tool when investigating or modifying an already existing file load the par file into ODEON study the room in the 3Dview please see the help text shortcut F1 available from within this display then make the changes in the editor which can be opened from within ODEON Using the Extrusion modeller eee Izo oo le ls o0 hooo h2 haoo 16 00 Modeling plane l YZ Vertical cross section E oa RLE ISLES Le EE LELTE TTEA T LARIAN EIR UPd BSN EPIN TEE Re RE MORE ene WAL ISIE et BONE Le EIES Se SIE Meee 8 00 D XY Horizontal Z Vertical length 7 00 l 700 Grid and snap options E V Snap to grid aa et Lock equal points Grid spacing 20 200 Metres 6 00 4 6 00 Snap size 025 0 25 Metres eras Vivi als ean sia i Tae a alia OTe wide ae A Si CO ITS lO ee SO CL 7 a ga A ae a A a ae re Ce aI oy aed Ean Ot Ae Batak Sikes A e F Snap to existing coordinates V Lock H and snap Layers Use Layers _ Move all objects on layer _ 5 00 5 00 Layer Current Name Active Cok Blue layer X i EAEE ee E PUA OR a
297. the tree structure is described below and should be compared with the XML file to keep on track See also figure E1 and E2 Main node all nodes below included ArraySource Main node in the XML file which holds the entire array Attributes to ArraySource ODEONArrayVersion 10 0 If this number is higher than the version number known by the READER then the reader should not accept to read on This version number is the version number of ODEON when the format was last revised That is to say if the READER is ODEON 10 0 or another program aware of ODEONArrayVersion 10 0 then it should not accept data written in ODEONArrayVersion 10 2 This on the other hand implies that if your WRITER does not use any specifications added in a later ODEONArrayVersion then you might consider using the lowest version number possible to make it compatible with READERS not knowing of the higher version READER should assume no more than 4 decimals in the version number Gain 8 overall in dB per octave band defaults to 0 150 Delay 0 007 Delay in seconds of the entire array ArrayCoordSys Absolute Can be Rel hanging Rel standing or Absolute Most programs will probably export as Absolute ODEON has the two other options in order to allow automatic aligning and sort of transducers relative to the upper rel hanging or lower rel standing transducer The origo of the transducers can be offset using the node C
298. ther Binaural Room Impulse Responses BRIR s OF Surround Sound Impulse Responses BFormat is also available for the advanced user which can be calculated as part of the Single Point Response in the Job List if the Auralisation Setup gt Create binaural filters Or Auralisation Setup gt Create 2D Surround Impulse Response option is turned on In this section only the binaural simulation is covered but most of the points also go for surround Auralisation In short terms the BRIR is a two channel filter through which a mono signal passes from the sound source s to the left and the right ear entrance of the listener receiver Using convolution techniques to convolve a mono signal with the BRIR a binaural signal is obtained which when played back over headphones should give the same listening experience as would be obtained in the real room Mixing such binaural signals created with different source positions and signals but with the same receiver position and orientation multi channel simulations are possible e g simulating a stereo setup background noise versus loudspeaker announcements or singer versus orchestra 5 1 lwiListening to Binaural Room Impulse Responses As mentioned above the basis of binaural auralisation in ODEON is the Binaural Room Impulse Response BRIR s which are calculated as a part of the Single Point Response In order for ODEON to calculate the BRIR s you should make sure that the Create binaural impulse respons
299. tic distance a than was the case in the ray tracing process where drefl was assumed to be equal to dinc So which directivity to assign to the secondary sources We propose a directivity pattern which we will call Oblique Lambert Reusing the concept of Vector based scattering an orientation of our Lambert sources can be obtained taking the Reflection Based Scattering coefficient into account If scattering is zero then the orientation of the Oblique Lambert source is found by Snell s Law if the scattering coefficient is one then the orientation is that of the traditional Lambert source and finally for all cases in between the orientation is determined by the vector found using the Vector Based Scattering method Shadow zone Oblique angle Figure 0 2 Traditional Lambert directivity at the top and Oblique Lambert at the bottom Oblique Lambert produces a shadow zone where no sound is reflected The shadow zone is small if scattering is high or if the incident direction is perpendicular to the wall On the other hand if scattering is low and the incident direction is oblique then the shadow zone becomes large 77 If Oblique Lambert was implemented as described without any further steps this would lead to an energy loss because part of the Lambert balloon is radiating energy out of the room In order to compensate for this the directivity pattern has to be scaled with a factor which accounts for the lost energy If the angle is zero
300. tical dashed purple line indicates the onset time of the impulse response ODEON detects the strongest peak in a recording whether this is a typical impulse response or any sound file If the recording consists of a series of impulse responses e g hand claps the onset time of the strongest impulse response will be detected A vertical dotted red line unique to each octave band indicates the truncation at the noise floor Whenever there is not sufficient signal to noise ratio in the impulse response recording the truncation time cannot be determined safely Instead of the truncation time the estimated end of the response is displayed again by a vertical dotted red line You can zoom in by clicking the left mouse button and dragging towards the lower right corner of the graph In that way a zoom in rectangle is specified between certain values of the horizontal and the vertical axes Releasing the mouse button the rectangular area is magnified to fit the whole window Remember Navigate though octave bands by pressing the Up and Down arrows at the keyboard or by clicking next frequency band previous frequency band in the Measured Response menu that should be already activated in the top menu bar Raw Decay Curve This is a display of the squared pressure impulse response i e energy impulse response together with the onset and truncation times In this graph the values of the squared pressure are in dB ref 20 uPa which makes detail
301. till ODEON is a high frequency model so surfaces should be kept reasonably large Avoid using more surfaces than needed in order to mimic the geometry Modelling a lot of very small surfaces to achieve high geometrical fidelity will not improve quality of results but it will increase calculation time It is difficult to put concrete limits on the size of surfaces which should be used there will always be a need for small surfaces to fill in awkward corners of the geometry Remember A rule of thumb may be to keep surface dimensions larger than one wavelength at the mid frequencies one wavelength at 1000 Hz is approximately 0 34 metres 23 Often you will be in a situation where a few surfaces should be kept small down to 0 20 metres or even 0 10 metres You can just keep these surfaces if their number is low relative to the total number of surfaces Tip It should be possible to model most concert halls with a surface count of say in between 100 and 2000 surfaces Curved surfaces All surfaces in ODEON must be almost planar so curved surfaces have to be approximated by dividing them into plane sections The question of how finely to subdivide depends on the type of curved surface and how important the surface is Convex curves naturally disperse sound energy so if the surface is in an exposed position e g the end of a balcony near the stage one should avoid for example simply replacing a quarter circle with a single plane at 45 w
302. tions for materials Two buttons allow more flexibility when editing the list of materials Restore coefficients The button restores the absorption coefficients in all octave bands for the selected material It is useful when you have started modifying the coefficients for a material and want to obtain the original coefficients Copy range to other materials If you have modified the Search Range of a material default value is 50 you can assign the same value to all materials by pressing this button Original and best error displays The Best Fitting graph presents the original error between simulations and measurements versus the best value so far The error is represented in JNDs The Last Error Decrease shows the time elapsed since the last update of the best value so far at the Best Fitting graph Live display of error Here the error for every individual solution per generation is shown Normally the error fluctuates highly in the beginning of an optimization process and becomes smoother as soon as the genetic algorithm converges Absorption coefficients Three different lines of absorption coefficients are updated constantly during a calculation e The Initial line corresponds to the absorption coefficients of the original material e The Optimized line shows the absorption coefficient that provides the best results so far e The Current line corresponds to the absorption coefficient of the running solution which may or may no
303. to include small geometrical details at the first attempt If there are some large surfaces which are basically plane but contain complex geometrical features e g a coffered ceiling model them at first as simple planes Then first when this room has been made watertight make the necessary alterations to the geometry file The simplified version can also be used in the prediction exercises to give some idea of the effect of the feature in question Examples on parametric modeling This section will give some short examples on the modelling of rooms using the parametric modelling language of ODEON The options in this modelling format are many ranging from typing the model number by number to dedicated programming This section will try to give an idea on how to use the language and its keywords In the default room directory created at the installation of ODEON you may find several other examples on the Par format 181 Four ways to model a box These examples show four ways to model a box shaped room using plain numbers using constants using constants plus symmetric modelling and using the Box statement along with the MTranslate statement In each example the dimensions of the room are W L H 4 6 2 7 Below the box shaped room is modelled using plain decimal numbers Parametric sample BoxFromPureNumbers par HHH Pt 1 0 2 0 Pt 2 0 2 0 Pt 3 6 2 0 Pt 4 6 2 0 ceiling points Pt 11 0 2 ZF Pt 12 0 2 2 7 Pt 13 6 2 27
304. ts a number of CAD entities which can be exported from these programs and imported directly by ODEON without any extra effort Depending on the modelling program used and indeed how it was used different approaches may need to be taken in order to ensure that all or most of the drawing data are exported to the dxf file in a form which can be understood by ODEON If ODEON encounter entities in the import process which ODEON recognizes but doesn t support then ODEON will notify about it The modelling programs should be 3D modelling programs Programs such as AutoCAD LT only have limited support for 3D modelling and are not recommended Programs such as Rhinoceros AutoCAD IntelliCAD and 3DStudiomax are true 3D modelling programs and have been reported to be suited for the purpose Other programs may work as well but in any case you may have to experiment in order to find the optimum way to export and import the geometries from the programs About CAD drawings Room models to be used by ODEON must be surface models defined from plane surfaces no matter if the models are created in a CAD program or if they are modelled in the ODEON environment e g using the ODEON par format Once a model has been successfully imported by ODEON it is important to perform a thorough check geometries which look fine in the drawing program may still contain serious errors such as repeated misplaced or missing surfaces CAD entities supported by
305. tten as an ASCII text file having the file extension Par the old ODEON Sur file format is also allowed though not described in this manual You can choose to create the geometry file either by typing the model data directly into a text file in the supplied text editor ODEON text Editor using the format described in Appendix D using SketchUp the ODEON Extrusion Modeler described in section 3 2 or a third party CAD program e g IntelliCAD Autodesk Revit AutoCAD 3DStudioMax MicroStation or Rhino which is capable of creating 3D surface models and exporting these as dxf files as described in section 2 4 Finally you may combine the different modelling methods import a CAD model from a CAD program and extend or correct it using tools which come with ODEON Appendix F gives an overview of which file formats can be used with ODEON Remember No matter which approach you choose for modelling always check the validity of the models The room model must form a almost closed enclosure It should also be almost free from warped twisted duplicate or overlapping surfaces ODEON has several tools for checking models for such problems The tools are presented in section 2 6 It is suggested that you always use these tools when working on models of some complexity 20 2 1 Pre calculated Rooms Before making any models to be used in ODEON it is good idea to have a look at the pre calculated rooms in your ODEON installation You will see how
306. tudy the examples which are installed in the ODEON Rooms Manual samples directory along with the ODEON program open the room s in ODEON then click the Open the ODEON Editor icon on the toolbar in order to study the geometry file Components in the modelling format The basic function of the modelling format is to allow modelling of surfaces in room geometries The surfaces can be modelled point by point surface by surface however it is also possible to make use of symmetry and to create repeated features in a room such as columns using programmatically loops finally it is possible to use hybrid functions which creates points as well as surfaces in terms of shapes such as boxes cylinder and domes Constants variables and counters Constants and variables can be defined and used in the file format It is a good habit to use constants whenever a value is used more than a few times in a file this reduces typing errors and it also makes it easier to make general changes to geometry such as changing the height of a room 156 Mathematical expressions Mathematical expressions can be used to express any real or integer number in the file e g coordinates constants variables counters point numbers surface numbers etc If you use a value that is not an integer to describe a point or surface number then that value will be rounded to the nearest integer value You may describe coordinates using mathematical expressions like Len
307. tween units is d 0 20 m Whereas the single loudspeaker unit spreads the sound in a wide fan the level decreases rapidly with distance With increasing number of units in the array the sound gradually concentrates in a beam with very small splay angle 97 my kn 2 z ma e a os Ly GS N i ami Pee eR TT eee eee eee ee Nett Figure 10 1 The radiation at 1 kHz from arrays with increasing number of units from a single unit to 11 units 1 3 5 7 9 and 11 With 7 units in the array the sound radiation in the octave bands from 250 Hz to 8 kHz is shown in fig 11 2 per seereengggas Leet sree ggags y TURATI t we D He woe m O G o G a w Figure 10 1 The radiation in octave bands from 250 Hz to 8 kHz for a line array with 7 units TTTRREROUS eS ee 20D He Figure 10 2 Similar to fig 11 2 but the array is bent see Table 1 98 In order to increase the splay angle to fit an audience area it is usual to apply different elevation angles to the units and thus creating a bent array An example is shown in fig 10 3 and the coordinates and elevation angles of the units are shown in Table 10 1 below At high frequencies it is obviously a problem in this example that the sound radiation splits according to the number of units and there are gaps wit
308. two figures show an impulse response recording containing 4 hand claps in a row Only one of the impulse responses has to be processed ODEON detects the one that has the highest peak in the broadband version i e the 2 one The beginning of the 34 impulse response defines the end of the noise floor for the 274 impulse response so that the truncation point is correctly placed with respect to the noise floor associated with the 274 impulse response If you wish to process another impulse response you should crop the desired part of the recording and save it as a new file zoom the relevant part then save it using the c shortcut Once the file has been saved it is automatically opened in the Measured Response window You can also redefine the onset and truncation times as described in section 11 5 and save the impulse response with a new name Decay range less than 45 dB 24 48 dB at 250 Hz Measured response C Odeon13Combined Measurements Errors Handclap1_2_3_4 wav Co a mE Raw Impulse response at 1000Hz Raw decay curve at 1000Hz Decay curves at 1000Hz T 30 0 53 seconds Decay curves all bands Energy parameters Parameter curves Freque a C Odeon13Combined Measurements Errors Handclap1_2_3_4 wav tt i sss Ray Impulse response at 1000Hz i Omni microphone M Figure 8 microphone V Onset time Truncation time 1 1 0 1 Time seconds A series of hand clap recordi
309. ulation time as the number of individuals has a linear influence on it By limiting the range of absorption coefficients the search process also becomes more efficient as the GA will only search where there are possible valid solutions e g if it is known that there is mineral wool in the ceiling the GA should not specify absorption corresponding to wooden floor and vice versa This will not only make the search faster but it will also prevent unrealistic solutions which match the target well but they are obviously wrong wooden floor on the ceiling and mineral wool on the floor It is recommended that the user initially assigns materials as realistic as possible In the ODEON Genetic material Optimization tool it is possible to assign a search range between 0 and 100 to each material Best Fitting Last Error Decrease Crossover method Gene exchange Restore Coefficients Frequency optmzaton Indiwdual Bands Copy range to other matenals W Orig Fitness IB Best Fitness Evoluton Method gm htst Ir vicuals per material 2 Seconds Crossover _Probablity 8 0 Average error in JND s Inversion Probability 10 0 lt Mutaton_Probabiity 15 Eitist percent spo i Caladate 6 5 Z0 SO 1000 2000 4000 8000 63 125 250 500 1000 2000 4000 8000 Frequency Hertz Frequency Hertz J 63Hz V 125 Hz V 250Hz V 500 Hz V 1000 Hz Y 2000 Hz V 4000 Hz 7 8000 Hz Area Search Instial Abs uaa Absorpti
310. uld be twice the highest frequency of concern In room acoustic applications the sampling frequency is 44100 Hz or 48000 Hz which offers fine sampling up to 22050 Hz and 24000 Hz respectively These frequencies are well above the highest octave band of concern 8000 Hz 107 Tip If during the measurement a sudden peak forces the recording to clip ODEON terminates the measurement showing a warning message In this way you are not able to store such faulty measurement A measurement can be initiated by pressing the Measure button The status of the measurement is displayed at the bottom of the window When the sweep signal has been played the extra decay period is recorded afterwards keep as quiet as possible during the period of sweep as well as during the Impulse response length period Load Impulse Response You can load an impulse response file in uncompressed wav format by pressing the Load Impulse Response button or by clicking Tools gt Load Impulse Response SHIFT CTRL L In principle you can load and process any sound wav file in the ODEON measuring system You should be able to see even a music signal in the ODEON measuring system editor but calculated parameters may not make any sense and the room acoustic parameters may not be calculated at all for such a file The larger the file is the longer it will take for ODEON to open A folder called Measurements is included in the ODEON main folder where impulse responses are stored b
311. ults of mathematical expressions are automatically rounded to the nearest whole number Operation Addition Subtraction Multiplication Division ra p lt II Power Base Exponent 253 8 or or Power Exponent Base Power 3 2 8 Root Root Index Radicand Root 3 8 2 Round Round X Round 2 67676 3 Truncation Trunc X Trunc 1 7 1 or Int X Sine of an angle in radians Sin X in O O Cosine of an angle in radians Cos radians Cos PI 4 0 707 106781 186547573 Tangent of an angle in radians Tan radians Tan PI 4 1 UN Cotangent of an angle in radians Cotan 180 0 Hyperbolic Sine of an angle in radians Sinh 0 0 Hyperbolic Cosine of an angle in radians osh 0 1 Sine of an angle in degrees inD 90 1 Cosine of an angle in degrees osD 0 1 Tangent of an angle in degrees TanD 45 1 Cotangent of an angle in degrees CotanD 90 0 Inverse Sine in radians ArcSin Sqrt 2 2 180 PI 45 Inverse Cosine in radians ArcCos Sqrt 2 2 180 PI 45 Inverse Tangent in radians ArcTan 1 180 PI 45 Inverse Tangent II in radians ArcTan2D 1 1 180 PI 45 Inverse Sine in degrees ArcSin Sqrt 2 2 180 PI 45 Inverse Cosine in degrees ArcCos Sqrt 2 2 180 PI 45 Inverse Tangent in degrees ArcTan 1 45 Inverse Tangent II in degrees ArcTan2D X Y ArcTan2D 1 1 45 Exponential Exp 1 2 71828182845904509 Natural Logarithm Ln 2 718281828459045091 1 Logarithm base 10 Log10 100 2 Logarithm bas
312. umber of points as the first curve lt SectionsInRevSurf gt 167 The number of surfaces to be created by the RevSurf statement If creating a cylinder a number between 12 and 24 is suggested Although it is easy to create many surfaces in a revolution surface too many small surfaces should be avoided If the FirstSurfaceNumber is 100 and SectionsInRevSurf is 3 surface 100 101 and 102 will be created lt Optional name gt Optional user defined name for easy identification of the surface e g cylindric wall Example RevSurf 1000 100 200 6 Cylinder creates a revolution surface divided in 6 surfaces surface 1000 1005 This call requires two curves of each 6 1 points to be defined namely point 100 to 106 and point 200 to 206 If the two curves of points define corners in the lower and upper edge of a cylinder a cylinder of 6 sections is created see example room RevSurfCylinder Par Loops using the FOR END construct The For statement must follow the syntaks For lt CounterName gt lt CountFrom gt lt CountTo gt lt CounterName gt Name of counter to be used by the For statement The counter is automatically defined by the For statement and becomes undefined when the loop finishes The counter can be referenced within the for end loop as an ordinary constant or variable if desired so lt CountFrom gt First value the counter takes The CountFrom value is considered an integer value If the number entered here is
313. umming over many particles a global energy decay function for the room is obtained The decay curve is corrected for energy which is lost due to the truncation of the decay curve This is analogous to an ordinary decay curve except there is no specific receiver The summation process may be carried on for as many rays as desired Evaluating results When the decay curve seems smooth derive the results It is often the case that Tso values are longer than Tz values If Tso values are shorter than T it is likely that the number of rays used were too small in that case press the Recalculate button If the reverberation times are the Impulse response length defined at the Room Setup page is too short to derive the reverberation parameter 6 4 Calculation of Response from Sources to Receivers This section describes the methods used to predict the response from a source to a receiver This is the process used to in order to predict Single Point Multi Point and Grid Response results from within the Job list The results of the simulations are similar to what can be obtained from impulse response measurements in a real room Source types calculation methods Responses from point sources are calculated using a hybrid calculation method where the early reflections are calculated using a mixture of the Image source method and ray tracing and the late reflections are calculated using a special ray tracing process generating secondary sour
314. unique number from 1 to 2 147 483647 for identification of the first point and surface in the Dome Using the same number but with negative sign defines the dome and its mirrored counterpart in the XZ plane Y 0 A Dome will take up several point and surface numbers which must all be unique lt Radius gt Radius of the Dome must always be greater than zero lt Revangle gt Revangle must be within the range 360 and different from zero If RevAngle is 180 a half Dome is generated if its 360 a full Dome is generated Positive revolution angles are defined counter clockwise Connection points The right side vertical points in Dome are stored in PlistA The left side vertical points in Dome are stored in PlistB In the special case where the revolution angle is 180 all points are stored in PlistA and the number of vertical subdivisions is stored in ONVert The example shown was generated with the following code 178 HHH const N 16 const R 15 Dome 1 N R 270 This is a dome HHH Hint The dome can be made elliptical using the MScale statement The Dome2 statement The Dome2 statement is a Dome shell of the calotte type where the vertical revolution angle is not necessary 90 Rather then specifying the dome by a revolution angle it is specified by the width and height Dome2 may typically be used for modelling dome shaped ceilings The syntax for Dome is Dome2 lt Number gt lt NumberOfSurfaces gt lt W
315. update file cif dialog e g select User2005_Dongle102009_ODEON Industrial_V5__Restricted_Oh cif Update installation When the license has been downloaded to your dongle you can find and install the new version or edition from the webpage www odeon dk updates You may choose to uninstall the previous version before installation We recommend that you keep a safety copy of the installation file e g if you at some point want to reinstall the downloaded version The window below shows all the update options that may be available whether all these options will be displayed available depends on the license currently stored in your dongle Update option Update to most recent version When a license of ODEON has been purchased for one of ODEON version of ODEON it is valid for the full version number If the license was purchased for ODEON 9 1 then license stored in the dongle will be valid for version 9 x e g 9 1 9 2 9 21 or whichever versions are released before the next full release number e g ODEON 11 22 The software for updating versions in between the full version numbers can be obtained free of charge from www odeon dk Update user name must be For copy protection your user name and country is supplied separately embedded in the dongle If you should wish to you can change your user name by checking the box Update user name the user name can only be updated separately from the other updates When you send the reques
316. uralisation options in ODEON Auditorium and Combined the hardware requirements how to publish calculated sound examples on the Internet or on audio CD s etc Chapter 5 introduces the calculation principles used in ODEON giving an idea on the capabilities and limitations the simulation part of the program Chapter 6 describes important calculated room acoustic parameters available in ODEON how they are calculated and how to interpret the results Chapter 7 describes the various calculation parameters available in the program Most of the parameters are automatically set to reasonable values by ODEON however for special cases you may need to adjust some of the calculation parameters Chapter 8 is the discussion on quality of results and how to achieve good results This chapter may be relevant once familiar with the program Chapter 9 describes how to extend the library of directivity patterns available for point sources and the use of directivity patterns in the Common Loudspeaker Format CLF Chapter 10 introduces the line array sources discussing some basic principles behind array source design Chapter 11 presents the integrated impulse response measuring system in ODEON The chapter explains which measuring methods are used how to obtain a healthy impulse response and how to obtain room acoustic parameters as the results The chapter ends with an annex giving examples of good and bad impulse responses Chapter 12 is describes how to
317. use An extruded surface is a flat 2D outline drawn at a specified drawing depth the third coordinate and with an extrusion height When assigning an extrusion height to the 2D outline it becomes a holster outlined by the edges of the extrusion surface if so desired this holster can have a bottom and a top In the extrusion modeller it is possible to make one drawing which contains multiple extrusion surfaces each described by a 2D outline a simple drawing and the line properties drawing depth extrusion height bottom check mark top check mark and a name If the extrusion is created in the XY plane then one extrusion surface may form a volume with walls and optional floor and ceiling whereas other extrusion surfaces define tables chairs or screens Odeon 1985 2004 The Extrusion Modeller and file formats The output from the extrusion modeller comes in two formats the extrusion model can be saved in its own native format in an oes file This file can be edited and extended at a later point in the extrusion modeller e g if wishing to change width of the auditorium above to change some of the points in the drawing or to add other features 33 The other format is the ODEON par format which is loaded into ODEON for calculations The parametric format cannot be edited in the extrusion modeller on the other hand it may be edited to any degree of freedom if needed and it is possible to make use of the benefits of
318. values on the surface list left part which will only affect the surfaces that use this material Remember If you have to define new materials in the library you have the chance to make changes only for the library associated with the room Press the Toggle between local and global room library button The title on the top of the material library will change from Global material library to Room material library Now when you add a new material or delete an existing one it will not affect the default global library PE Material list ca C fea aera pel pa F Room material library auditoriume1 at DTU5receivers GK Materials Z5 a a i Normal q E Saad 100 reflecting Nomad el ra 10 absorbent Norma Ea T 20 absorbent ERT eg ce 30 absorbent BE f 40 absorbent E m 50 absorbent Backwall in aud 21 first guess 50 W 0 14000 o 10000 0 06000 0 00000 0 00000 0 00000 0 00000 Transmission data unassigned pall type Normal b Change absorption coefficients here in order to leave the material in the list unchanged The best approach is to focus your calibration procedure on a few most critical materials two or three ones and try to vary them until you get the best possible match Example In a classroom all materials are pretty well known except from the ceiling An initial estimate for the ceiling is chosen and the first simulation shows higher simulated Tso than the 127 measured one The user has to go back in t
319. ve file residing in the directory set in the Options Program setup Paths Wave folder To play the chosen input file make sure this cell is selected then press the Alt s shortcut or the Play wave button 4 Choose a Job from the Job no column to convolve with the selected signal Exit the Job no cell but stay on the same row Immediately information about the Job is revealed Press the Run Single Job button J This will calculate the point response between the source s and the receiver in the selected Job if it is not already calculated and afterwards it will perform the convolution between the response and the selected anechoic signal in the list If the convolution line becomes red at the end of the process the output level is too high above 0 dB and it is clipped This has to be fixed otherwise you will experience EIEN unpleasant distortion when __ cenove snauai RIR win sna Fopdesc Red lev Max out listening to the auralisation D 1 Voice Sabine Short 0 11 01 Average 3 1 towards P1 No description 2 pe Average none 0 00 99 00 Reduce the recording level in the 3 none Average El none 0 00 99 00 4 none Average none 0 00 99 00 Rcd Lev column by the same 5 none Average none 0 00 99 00 lt 6 none Average none 0 00 99 00 amount of overloading dBs In aa en ae 0 00 99 00 8 none Average none 0 00 99 00 this example the level 1S 9 none
320. w much the dips and notches have been exaggerated If M is 0 then the effect is neutral a value of 3 0 improves localization without too much undesired colouration A M value greater than 3 0 does not seem to give any noticeable advantages whereas a value less than 3 0 gives less colouration The enhancement algorithms are further developments of those used in earlier versions of ODEON as a result the M factor can be set as high as 3 0 allowing a significant improvement of the 3D experience through headphones without noticeable drawbacks A When an A is appended after the M factor this tells that Minimize colouration effects diffuse field approach was not enabled so the HRTF were optimized for Anechoic conditions i e no reflections present If no A is found in this place in the name then ODEON has attempted to Minimize colouration effects The enhancements applied may give some colouration effects because some frequencies are amplified for individual directions When this option is checked ODEON will try to accomplish that colouration is kept at a minimum in a multi reflection environment that is the average frequency response for all directions in the set of HRTF s is kept neutral Sample rate The sample rate of the HRTF This sample rate should be the same as the sample rate of the signal files anechoic recordings to be used The supplied HRTF s are sampled at 44100 Hz Apass Ripple of octave band filters in dB Smaller is better 0 5
321. w operation Mouse operation Scroll drawing area Right mouse button Zoom In Out Snap to grid Snap to grid enables points new points or points being moved to be positioned exactly at the intersection of the grid lines In special cases the point can also be inserted at only one grid line when the other coordinate is that of an existing point see below snap to existing points Snap to existing points Snap to existing points enables points to be precisely located on either or both reference coordinates of existing points 35 The snap point The snap point is a special case of snap to existing points In some cases you may want to move a surface to a precise location e g 0 33 0 46 not being a point on the grid nor an existing point of another surface In that case e Create a new surface e Click the approximate position of the reference point e Change the coordinates to the exact position e g 0 33 0 46 in the Point editor e Press Insert or Esc to finish editing the surface the point will appear with the mark Snap point e Select the fix point in the surface to be moved left mouse button and move it to the location of the snap point Do note that Snap to existing coordinates option must be checked Once the surface has been moved the surface containing the snap point may be deleted this is not a strict requirement as surfaces containing only one point will not be transferred to the par format to be used in
322. wall while the rest 90 is kept inside the source room for simulation of room acoustics Compensation is made concerning the energy so the calculation result in the source room is not affected by some rays being transmitted to another room only the absorption data for the wall are used In general it is a good practice to use higher amount of rays eg double number when making transmission calculations to ensure enough rays will reach the receiving room behind the wall Assigning transmission data PE Material ist Surface List Once a Type of a wall has been set to Transmission umber materiai _scatter Transp Type _ Surface name 1 100 0 050 0 000 Normal walls room 1 in the Material List it becomes possible to specify 2 ms oos 0 000 nomal walls room 1 TOn 3 703 0 050 0 000 Normal walls room 1 transmission data using the Edit transmission data 4 703 0 050 0 000 Normal walls room 1 5 703 0 050 0 000 Normal _walls room 1 for surface gt button or Alt y shortcut This iiss m oo oo nomai wals room 1 1 p gt i 000 Transmissiv wall opens a dialog where Reduction indexes can be pa s vo 000 gamma troom 172 specified in one third octave bands from 50 Hz 3 m 0 0 00 Freon to 10 kHz The data can be entered directly or copied from a spreadsheet or from a text file using the common shortcuts Ctrl C and Ctrl V For an example of transmission open the room Transmission rooms i
323. when the point source is not visible the contribution can be added to the impulse response without considering the phase interaction between direct sound and the diffracted component just like the reflections are added to the impulse response Only diffraction from one edge of the diffracting object is taken into account the one with the shortest path length The edge of diffraction is considered infinitely long More information and validation is found in Rindel Nielsen amp Christensen 2009 By using the algorithms by Pierce rather than less complicated ones ODEON is capable of handling more complicated objects than the single surface screen e g objects such as book shelves 78 the corner in e g a L shaped room and buildings out doors or diffraction over a balcony front edge or over the edge in an orchestra pit Estimating diffraction points around objects In order to calculate diffraction it is essential that the diffraction path is established In rooms with just moderate complexity it becomes virtually impossible for the user to describe which surface edges will act as diffracting edges and therefore ODEON has a built in detection algorithm In order to limit the complexity and time consumption of the detection there are some limitations to the type diffraction paths ODEON can detect In order to understand the limits here is a short and slightly incomplete description When a source is not visible from the receiver ODEON wi
324. with virtually no delay arriving some time before the start of the impulse response is actually detected by the microphone This peak is undesirable and should be removed Changing the noise floor window length can help in placing the truncation time at a correct place but the onset time will be still placed wrongly at the peak of the magnetic feedback The best solution is to zoom the healthy part of the impulse response then save this cropped version using the c shortcut Manual setting of onset time is also possible according to section 11 5 It is of course better to avoid long parallel cables in the first place if you are still in the field then by all means repeat the measurement le Unable to derive decay parameters for 125 250 500 1000 2000 4000 Hz bands Measured response C Errors Wrong_onset_ElectroMagneticFeedback_Corridor_r fo xg Raw Impulse response at 1000Hz Raw decay curve at 1000Hz Decay curves at 1000Hz T 30 seconds Decay curves all bands Energy parameters Parameter curves Freque gt C Errors Wrong_onset_ElectroMagneticFeedback_Corridor_r10 wav Ray Impulse response at 1000Hz p x E 1 0 1 2 3 4 Time seconds Electromagnetic feedback in an impulse response recording 118 Unable to derive decay parameters for 125 250 500 1000 2000 4000 Hz bands Measured response C Errors Wrong_onset_E Raw Impulse response at 1000Hz Raw decay cu
325. xt file as input Once the text input file has been created in one of the formats specified above e g in the ODEON editor ODEONEdit it can be translated into an ODEON Directivity file which can be applied to any point source from within ODEON To translate the created text file into an ODEON directivity file e Select Tools Create directivity So8 from ASCII file DAT e Open the input file you have created e Specify the name of the directivity file pattern you wish to create e Select whether you wish a Calibrated source or not e Apply calibration data as prompted for Applying Calibration Creating a new directivity file you will be prompted whether to create a calibrated source or not Calibrated source Sound Power Level 0 dB re 10E 12 W at 1 kHz YES NO Calibrated Sources Press YES if a generic source with an adjustable level is needed for ODEON calculations an example on this could be the OMNI or SEMI directional directivity pattern When selecting a calibrated source no data apart from the ASCII input file are required The directivity represented by the text file is preserved but the values are simply shifted by a constant amount the same for all bands such that the sound power level of the source is 0 dB re 10 2 Watts at 1 kHz Please do note that the power in the other bands may differ from 0 dB You may still alter the overall power response of the source by applying an EQ however the power at 1 kHz will alw
326. y default High quality Impulse responses can be measured using the sweep facility in ODEON if loudspeaker amplifier microphone etc are available It is also possible to record handclap popping of balloon or paperback etc e g using a smart phone if an App has been installed for that and then load the impulse response file into ODEON If a file contains multiple impulse responses such as hand claps ODEON will try to make use of the one providing the best signal to noise ratio S N The next figure shows a typical impulse response loaded on the Measured Impulse response window Distortion Noise 12 13 dB at 4000 Hz use longer sweep Measured response C Odeon12Combined Measurements Impulse response filel wav o B Raw Impulse response at 1000Hz Raw decay curve at 1000Hz Decay curves at 1000Hz T 30 0 29 seconds Decay curves all bands Energy parameters Parameter curves Frequency Response C Odeon12Combined Measurements lmpulse response file1 wav Ray Impulse response at 1000Hz Omni microphone Figure 8 microphone Onset time Truncation time p x E 2 0 1 0 0 1 0 2 0 3 0 4 0 5 0 6 0 7 0 8 Time seconds Figure 11 Q0 1 The measured response window in ODEON Raw Impulse Response The Raw Impulse Response displays the broadband and filtered pressure impulse response see Fig 12 1 You can switch between different octave band filters by using the Up and Down arrow 108 keys on the keyboard A ver
327. y Group The University of York http www york ac uk inst mustech 3d_audio Meyer E Kunstmann D amp Kuttruff H 1964 Uber einige Messungen zur Schallabsorption von Publikum in German Acustica 14 119 124 Naylor G M amp Rindel J H 1994 Odeon Room Acoustics Program Version 2 5 User Manual Publication No 49 Lyngby The Acoustics Laboratory Technical University of Denmark Ondet A M amp Sueur J 1995 Development and validation of a criterion for assessing the acoustic performance of industrial rooms J Acoust Soc Am 97 1727 31 Oppenheim A V amp Schafer R W 1989 Discrete Time Signal Processing J Prentice Hall International Parati L amp Otondo F 2003 Comparison of Directional Sources in Simulating a Soprano Voice Proceedings of the Stockholm Music Acoustics Conference SMAC 03 Stockholm Sweden Parkin P H Humphreys H R amp Cowell J R 1979 Acoustics Noise and Buildings London Faber and Faber Petersen J 1983 Rumakustik in Danish Horsholm Denmark Statens Byggeforskningsinstitut Pierce A D 1974 Diffraction of sound around corners and over wide barriers J Acoust Soc Am 55 941 955 Press W H Flannery B P Teukolsky S A amp Vetterling W T 1990 Numerical Recipes in Pascal The Art of Scientific Computing Cambridge University Press Rindel J H 1986 Attenuation of Sound Reflections due to Diffraction Proceedings o
328. yers in order to avoid that ODEON glues these surfaces together described below If a drawing is subdivided into layers this also makes it easier to assign materials to the surfaces in the Material List in ODEON because materials can be assigned to all surfaces on a layer in one operation Exporting a geometry from ODEON to IntelliCAD or AutoCAD When ODEON exports surfaces containing more than 4 points each these surfaces are exported using the 3DPOLY entity whereas all other entities are exported using the 3DFACE entity The 3DPOLY will appear as 3DPOLY lines in the CAD program and does not respond to the HIDE and RENDER commands like entities such as 3DFACE do However using the REGION command it is possible to convert 3DPOLY s into REGION s which do respond to the HIDE and RENDER commands Before exporting from the CAD program Remember that BLOck s are not supported by ODEON BLock s containing relevant 3D info must be exploded using the EXPLODE command in the CAD program in AutoCAD this may also be done by exporting the file to the 3ds 3D Studio Max format and importing it again as described in the section On 3DSOLIDS 3DSOLID REGION and BODY entities are not supported by ODEON try using one of the approaches listed above in order to make the geometry compatible A final remark is that it is always a recommended practice to make backup copies of your CAD files before making any conversions Performing the import in ODEON To

User's Manual

Contents

Download Pdf Manuals

Related Search

Related Contents