Sound Absorption in Free Hanging Textiles
Contents
1. Kilo Serge Calculations Measurements Frequency Hz Figure 11 Relative errors of reverberation times for calculated textiles without edge effect Calculated dissipation coefficient is used transparency is set to 1 T The red line indicates the upper 5 range The results for 4 kHz and 8 kHz are better but another deviation emerged between 250 Hz and 1000 Hz All textiles except Scene Molton and Fibertex F2B broke the 5 limit in this frequency range The results for Fibertex F2B are a little better but that may be because the edge effect calculations aren t effective for this textile The results for Scene Molton are rather good This gave averagely worse results than using the edge effect calculations 17 SOUND ABSORPTION IN FREE HANGING TEXTILES MAGNUS OGNEDAL 4 AbsTex A short user manual 41 The GUI The right of the screen gives the user direct access to the most commonly changed settings The calculated data 1s plotted in the right side of the screen 95 AbsTex Joe File Output Quickload aes Plot Help Default name Default description Thickness mm P rfkow resistivity kPas m Mass ym Sound incidence Diffuse field f Angle Model Moveabie Free hanging Texte w E Bs tu B in C ca Calculate dissipation ex 800 T1000 1250 1600 2000 2500 3150 1000 5000 6300 ey E Figure 12 The AbsTex user interface The two first text2. ccceeeeeeeeeeeeeeeeees 14 Figure 10 Relative errors of reverberation times for calculated textiles with edge effect 16 Figure 11 Relative errors of reverberation times for calculated textiles without edge effect 17 Figure 12 The Abs Pex user nterface e ooa etri orar ropa SENE e nO e en Un NEU tease 18 Picute 13 ODHUORBS TOP COp TH Gald sooo otc ond doo bof Dodo oot oua Voc boue Io AR acme 19 Figure 14 Parameters Tor edge effect estimation ic ori esee tee tae nete ente t asap Ine tu er eae e tuo 20 Figure 15 The gre nu StEHOBUEG oodd iidese to fais ideato ur ance ote to esee Pole Deiode iet seva die ein iae 21 Figure 16 Custom compare with data from e g MS Excel sss 21 SOUND ABSORPTION IN FREE HANGING TEXTILES 1 Measured parameters for the textiles The absorption coefficients were measured in a reverberation chamber with the textiles free hanging 100 mm from a wall and in a standing wave tube The transmission coefficients were also measured 1 1 The measured textiles Table 1 show the textiles used in this master thesis together with the weight given by the manufacturers and a brief description of typical use of them in acoustical applications Please note that the weights given by the manufacturers may differ from what measured weights see Table 3 Table 1 The measured textiles Weight g m Description Textile 1 Scene Molton 2 Ullseviot 3 Fibertex F2B 4 Velour 5 MI Wo
3. 1 0 o Ln Ln Z which can be used to find the reflection coefficient R We may then express the transmission as SOUND ABSORPTION IN FREE HANGING TEXTILES MAGNUS OGNEDAL ikd ikd rP I R _ e 1 R 2 2 7 P t and we may find the dissipation coefficient with p 1 R T These expressions can then be used with the Rayleigh and Delany Bazley models which give complex wave numbers and impedances based on the given airflow resistivity r the porosity c and the thickness d 2 3 Results and discussion The three developed models have been tested with parameters for the seven measured textiles The effect of the velocity of the textile has been tested A proposal for calculating the edge effect for low frequencies 1s also presented 2 3 1 Results from the models for free hanging textiles The new models have been tested with measured airflow resistivities densities thicknesses and calculated porosities The results are compared to measurements of the textile samples from a reverberation chamber All calculations were done in AbsTex All three models give better results as compared to earlier test done with the Mechel model 1 2 It may be assumed that the reason for this is that the new models also take the vibration of the textile into account For the banners most alike Ullseviot Velour MI Wool Serge and Super Wool Serge we see that the measured absorption coefficients are somewhere between the
4. 7 Frequency Hz Figure 10 Relative errors of reverberation times for calculated textiles with edge effect Calculated dissipation coefficient is used transparency is set to 1 T Both mass correction and dissipation correction for the edge effect is used The red line indicates the upper 5 range The calculated values for Fibertex F2B are lower than the measure values This is not surprising considering the results the MFT model gives for this textile If the edge effect 1s taken into account the calculated dissipation coefficients for Velour gives the best fit to the measured absorption Despite of this Velour gives the second worst results when the calculated values are used in Odeon The 5 limit is broken for 8000 Hz by Scene Molton Velour M1 Wool Serge and Super Wool Serge But overall these results may be judged as acceptable since the all banners are between 7 of the measured data If Odeon is calculating the edge effect for high frequencies disregarding the edge effect from the input absorption will result in higher calculated reverberation time for these frequencies see Figure 11 for results 16 SOUND ABSORPTION IN FREE HANGING TEXTILES MAGNUS OGNEDAL 12 96 11 96 10 9 8 7 6 5 4 3 2 1 96 0 1 96 2 96 3 96 4 96 5 96 6 96 Scene Molton Ullseviot Fibertex F2B Velour M1 Wool Serge Super Wool Serge
5. E mail magnus ognedal com PS This is only a fraction of the total document If you want to learn more please contact me 23
6. 16 4 ABSTEX A SHORT USER MANUAL eeeeeee eee eee eee eee eto sonas asses eee ee eee ette ette epos 18 4 Te Gi Oh NN 18 Be Jeunes cceeasevm set Ite eee OU eeu eve te Mete us Mae te e uU use too etes uu es Gees 19 4 3 Liequentb asked CICS FLOM aeo dioe bo uu cubi bitu dn DeL deo 22 4 4 System requirements installation and upgrading eeeseeeeeseeeeeese 22 S JOREERERENCES u diinii perti torio icol Doo ido DEus aa 23 6 CONTACT INFORMJA LION i usse eoe ea paee ao ee eoa iei cte t eos eue noe ia E eae ao eo PA Pea EEDE 23 Figures Figure 1 Measured one sided absorption coefficients for all seven textiles freely hanged 4 Figure 2 Measured transmission coefficients for all seven textiles sss 4 Figure 3 The basis for a free hanging porous textile cccccccccccccccceceeeceeeeeeseesssseeeeeeeeeeeeeees 6 Figure 4 Results for Ullseviot for all three models in diffuse field 0 0 ce cececeeeeeeeeeeees 10 Figure 5 Comparison for the MFT model and Ingards Volume absorber 11 Figure 6 Dissipation coefficient for a textile with Rs 2poco and different masses 12 Figure 7 Simplification of wave around an oblong textile in 3D left and 2D right I2 Figure 8 The dissipation coefficient for Velour with mass correction sss 13 Figure 9 The dissipation coefficient for Velour dissipation correction
7. Absorption Reflection and the Transmission coefficient and display the results in either octave bands and 1 3 octave bands The edge effect may also be calculated as described in Section 2 3 3 To calculate the dissipation with the influence of diffraction select Edge effect and On If the user selects Edge effect Set parameters a new dialog box will appear see Figure 14 letting the user set the width and height of the textile in metres The user may then turn the edge effect calculation on or off If a banner is loaded from file or quickloaded AbsTex will automatically set the width and height to the dimensions for this banner Keyboard shortcuts are also available Ctrl E will turn the edge estimation on and Ctrl R will turn it off Note that the edge effect is only estimated for the dissipation coefficient Width m 2 Height m L 5 Iv Correction of mass i Correction of dissipation Close Figure 14 Parameters for edge effect estimation The Quickload menu The data for the measured banners might be quickloaded from this menu The dimensions from Table 2 and the airflow resistivity and mass from Table 3 are used Shortcuts are available For example will the key combination Alt 1 quickload the data for Scene Molton 20 SOUND ABSORPTION IN FREE HANGING TEXTILES The Compare menu If the user chooses one of the seven banners for comparison the program displays the calculated dis
8. Kilo Serge File Output Quickoad Compare Plot Help 1 Scene Molton 2 Ullseviot 3 Fibertex F2B 4 Velour 5 M1 Wool Serge 6 Super Wool Serge 7 Kilo Serge Custom comparison w Mo comparison File Output Quickoad Compare Plot Help Set maximum y value w Enable legend w Enable background Ele Output Quickoad Compare Plot Help Quick Help About Figure 15 The menu structure The comparison might also be turned off It will automatically be turned off if the users change output or the octave band setting Custom Compare C Current calculation Custom data OK Cancel ance 10 447013500 454899410 460485910 464597900 466719110 468 185410 46919530 4698 10400 470157 100 4 704 1 1630 4 705661 5 ss la w zs j vo 252 sus aoo soo exo soo sooo i252 teon 2000 2503 sto anoo sooo aama 8000 O08 0 06 0 08 0 11 0 15 0 21 028 0 31 0 35 0 33 O41 1 0 45 048 048 o lt 7 O47 o lt 7 o lt 7 O47 O47 o lt 7 Figure 16 Custom compare with data from e g MS Excel The Plot menu These options are also found when right clicking in the plot area The maximum value for the plot is default set to 0 7 Values lower than 0 1 or higher than 1 are not allowed The Help button QuickHelp displays a short guide to the program About displays version number and contact information 21 SOUND ABSORPTION IN FREE HANGING TEXTILES MAGNUS OGNEDAL 4 3 Frequently asked questions Questio
9. NET Framework on your computer If you receive a message asking if you want to install the NET Framework 1 1 click the Yes button After 22 SOUND ABSORPTION IN FREE HANGING TEXTILES MAGNUS OGNEDAL the installation of the Framework is completed or if you already have the Framework installed the setup program will install AbsTex The AbsTex homepage is found here software ognedal com abstex The version of AbsTex available on homepage requires that the user sends an e mail to magnus ognedal com with his hers name An unlock key will then be returned to you This is done to get an overview of the users of the program Registering is currently free The unlock key will be necessary in order to unlock the full potential of AbsTex The following limits apply until the user enters his hers name and unlock key e The program will use ca 30 seconds to start e Exporting disabled e Copying results to memory is disabled e Values are not shown in plot only the graphs are shown 5 References 1 Ognedal M 2004 Sound absorption in textiles 2 Buen A 2004 Absorption in free hanging banners 3 Ingard K U 1994 Notes on Sound Absorption Technology 4 Microsoft 2004 NET Framework Version 1 1 Redistributable Package http www microsoft com downloads details aspx FamilyID 262d25e3 f589 4842 8157 034d1e7cf3a3 amp DisplayLang en last visited 24 05 2005 16 15 6 Contact information Web http software ognedal com
10. specific airflow resistance which can be relatively easily found it s the product of the specific airflow resistance r and the thickness d o is the porosity and v 1s the particle velocity This can be solved for vm A v A Pa GAP 2 1 2 OR och The total particle velocity v may then be written as V V OV ei SOUND ABSORPTION IN FREE HANGING TEXTILES MAGNUS OGNEDAL From the connection between sound pressure and velocity we get an expression for the specific acoustic impedance for the model A A l l A v v tov P io a Pot 2e O m S l gt MEE VE 2 1 4 yi This may be interpreted as a parallel connection between the limp mass impedance and the specific airflow resistance We can rewrite the equations to Z Z Z Z 27 2 1 5 35 Reo und Tan ien Z2 d zz The dissipation or absorption if you prefer may then be found with p 1 R T The model is named Moveable Free hanging Textile or MFT for short After finding these expressions it was brought to my attention that this is almost the same to what Uno Ingard 3 has derived although with a slightly variation Ingard also uses parallel connection between the limp mass impedance and the airflow resistivity but adds the reactance of the textile in form of iwdv This gives little lower dissipation coefficients for higher frequencies but for a textile with thicknesses as used here this has no practical effect This model pre
11. E 1 41 3 Fibertex F2B 35 192 2 1 20 4 Velour Banner 668 545 4 2 30 5 MI Wool Serge 515 548 1 1 43 6 Super Wool Serge 708 482 0 1 42 7 Kilo Serge 807 886 2 2 56 SOUND ABSORPTION IN FREE HANGING TEXTILES MAGNUS OGNEDAL 2 Models for free hanging textiles The models presented here are based on a purely resistive layer the Rayleigh model and the Delany Bazley model and the assumption that the difference in sound pressure between the two sides of the textile will cause the textile to vibrate slightly The following models are based on the combination of the velocity of the banner and the particle velocity through a porous material Since these velocities are directly linked to the difference in sound pressure between the two sides the dissipation 1s too 1 L 4 vi p D X YZ m OV Sa D x y z bp x y Z X Figure 3 The basis for a free hanging porous textile Illustration of the sound pressure mass textile velocity and particle velocity Figure 3 show a sketch of the problem which may be described by the incoming sound pressure wave p X y z the reflected and transmitted sound pressure wave the velocity of the banner v and air particle velocity within the banner vm 2 1 The Moveable Free hanging Textile model If we assume that the banner 1s infinitely thin and the particle velocity 1s constant through the banner we can write P Py Vm OR 2 1 1 where Rs is the
12. Modified Rayleigh model from now MR and the Modified Delany Bazley model from now MDB for frequencies up to 1 kHz see Figure 4 For frequencies lower than 1 kHz the MFT model is always higher for these banners SOUND ABSORPTION IN FREE HANGING TEXTILES MAGNUS OGNEDAL I Measured pues Moveable Free hanging Textile Modified Rayleigh Modified DelanyBazley 0 9 0 8 0 7 t n 0 6 e e c 0 5 B NETS i p F i 0 4 Sue 0 1 F ob E i I i i i i li I i I i l l l i I l i i L I L 63 125 250 500 1k 2k 4k 8k Frequency Hz log Figure 4 Results for Ullseviot for all three models in diffuse field r 647 kPa s m m 477 5 g m d 1 41 mm and c 0 656 2 3 2 Differences between the MFT model and Ingards Volume absorber As earlier mentioned the main difference between the MFT model and Ingards Volume absorber is that Ingard has added the reactance of the textile in form of icdv This addition actually makes the dissipation coefficient lower especially for higher frequencies For a typical textile with r 700 kPa s m m 600 g m and d 1 50 mm the maximum difference between the two models occurs at 8 kHz and is approx 0 35 which 1s so little that it may be neglected see Figure 5 10 SOUND ABSORPTION IN FREE HANGING TEXTILES MAGNUS OGNEDAL I I T Moveable Free h
13. Pressure P 4BYXpocket Out 4 ea ency Response PON pms gt NEn tts i F ei ag OU ie Sound cueorptionlt in free hi sH nam ieme jA ds COU I iptio on Wu ie ee Bees wi ere ERJ at i tud os LOSZAReatLett win b p k E doen bade em i Studios d ios xk 32AReat Riot w 600 2Dn E 355 x EE uS o A n 12 Is SOUND ABSORPTION IN FREE HANGING TEXTILES MAGNUS OGNEDAL Summary Hanging banners are used in many concert halls theatres and multipurpose halls to adjust the reverberation times A banner will absorb more sound energy when it is freely hanged compared to banners hanged at a distance to a wall The database of absorption coefficients for freely hanged banners is limited Therefore three models for free hanging banners have been derived The first model 1s based on a purely resistive layer The second and third models are based on porous models the model by Lord Rayleigh and the Delany Bazley model respectively These three models are combined with the vibration of the textile itself The vibration of the textile 1s found by the principle that the difference in sound pressure between the two sides of a free hanging textile will cause the textile to vibrate slightly This effect will cause the models to be mass dependent The first model has an advantage It 1s fairly easy to calculate and it 1s only based on the airflow resistivity and thickness practically th
14. anging Textile Ingards Volume absorber 0 8 0 6 0 4 U 9X90 92 OO0O e 02 0 3 0 1 i i i i I i li i I I i l i i 63 125 250 500 1k 2k 4k 8k i l l i i I l i i i Frequency Hz log Figure 5 Comparison for the MFT model and Ingards Volume absorber Calculated with r 700 kPa s m m 600 g m and d 1 50 mm 2 3 3 A rough estimation of the edge effect There is a need for a practical method to incorporate the edge effect How much will the low frequency absorption or dissipation change with the size of the banner We know qualitatively that the loss coefficients will tend to be larger the smaller the sample size At the same time will a smaller sample interact less with the sound wave of the same wave length at least when freely suspended Practical tests using WinFLAG also indicate that the edge effect depend on the thickness of the sample However with the MFT model the textile is regarded as infinitely thin The dependency of mass may act as a kind of high pass filter for the dissipation coefficient as seen in Figure 6 If the mass is increased the model returns more dissipation and lower mass give less dissipation This is because the textile velocity v and the limp mass impedance has a inverse linear relationship If the diffraction around the sample is taken into account some effect of the dissipation is lost due to short circuiti
15. boxes are for naming and describing the material currently being modelled for later reference If the user selects Angle as Sound incidence a new textbox will appear letting the user select the angle in degrees If the user right click in the plotting area the user has the option to turn the legend on and off turn the background on and off and set the maximum value for the plot See Section 4 2 under The Plot menu If the user selects either the Modified Rayleigh or Modified Delany Bazley model a new textbox will appear under the Model dropdown box that let the user set the porosity for the textile If the user wants the program to choose the porosity clicking the Suggest button can do this 18 SOUND ABSORPTION IN FREE HANGING TEXTILES MAGNUS OGNEDAL Example of data copied into Excel Name Default name Description Default description Description labels 1 3octaveband 50 63 60 100 125 160 200 T FCociweliand lahek Dissipation 0 03 0 05 0 08 0 11 0 15 0 21 0 2 Frequency 44 5 45 7 46 9 48 0 49 2 50 3 51 5 T 3 octaveband results Absorption 0 03 0 03 0 03 0 05 0 03 0 04 0 04 Continous labels Continous results OK Figure 13 Options for copying data The Copy to clipboard button opens a new window see Figure 13 Here the user can select the data he wants to copy export to for example Microsoft Excel The data 1s copied to the Windows Clipboard in tabulated format which is directly supported by Microsoft Exc
16. dicts a maximum in dissipation coefficient of 0 5 for Rs 2poco for normal incidences Values lower and higher than this value will according to the model result in a lower dissipation 2 2 Modifications of the porous models by Rayleigh and Delany Bazley We are searching for a general principle for free hanging textiles that can be used for more than one porous model This can be achieved by defining the sound pressure and particle velocity as P x Ae Be v x 4e Be m k 2 2 1 where A and B are pressure amplitudes for the incoming and outgoing sound waves At the borders x 0 and x d where d 1s the thickness of the banner see Figure 3 we get SOUND ABSORPTION IN FREE HANGING TEXTILES MAGNUS OGNEDAL P P 0 A B p p d Ae Be 2 2 2 and Ap p p 1 e A 1 e JB and the particle velocity Vm Vm 0 A B 2 2 3 v a V d Z 4e Be k The total velocity on each side of the banner can then be written as y Wt OV T y V O Vp l At the right side of the banner x d we know that Z p v which gives us ikd ikd Z 0 ikd ikd Z 0 ikd ikd Ae Be ses A l e 8 e 7 4e Be m k A gt E 2 2 5 ikd lee JA MZ ikd where n lr 1 pue L Z 55 to e Ln Z At the left side of the banner x 0 we can find the input impedance 7 A B s 1 e A peg 5 EEIE fn 2 2 6 gt Z ES B e 4 e B e
17. e measured The formulas presented in this section are developed based on observation of the measured values in combination with differences in phase between the two sides They should therefore be used with a critical eye and only as a very rough guide on the absorptive properties of a free hanging textile in a reverberation chamber A new and hopefully better method of calculating the diffraction based on Huygens principle may be developed 15 SOUND ABSORPTION IN FREE HANGING TEXTILES MAGNUS OGNEDAL 3 Absorption in Odeon The results from the models described here may be used as input in Odeon Therefore the dissipation and transmission coefficients for the seven textiles were calculated in AbsTex see Section 4 The transparency was as before set to the squared average transmission coefficient for the 500 Hz and 1000 Hz octave band For the absorption coefficients in Odeon the dissipation coefficients from AbsTex calculated with edge effect estimation as described in Section 2 3 3 were used The MFT model is used and the scatter coefficient in Odeon was set to 0 7 for all textiles See Figure 10 for results 7 4 6 5 4 2 3 o 3 Scene Molton 2 Ullseviot o 2 Fibertex F2B D 1 Velour o M1 Wool Serge 0 Super Wool Serge o o c 1 Kilo Serge 9 d amp 2 9 O 4 5 6
18. e specific airflow resistance in combination with the mass It also has no recommended upper limit in airflow resistance as the Delany Bazley model has The models are named the Moveable Free hanging Textile model MFT the Modified Rayleigh model MR and the Modified Delany Bazley model MDB The MFT model seems to fit the measured values best A rough estimation of the edge effect has also been developed SOUND ABSORPTION IN FREE HANGING TEXTILES MAGNUS OGNEDAL Contents 1 MEASURED PARAMETERS FOR THE TEXTILES eeeeeee eee eee ee eene enun 3 1 1 These asure TERES naon Renee rn ene tado Longue b aD ipee I auc Mh 3 1 2 Measured absorption and transmission Coefficients cccccseesssssesssssesssessssssesseees 4 1 3 Airflow resistivities dimensions and weights sse 5 2 MODELS FOR FREE HANGING TEXTILES ccccsssssssssssssscccccccsssssssssssssssees 6 24 The Moveable Free hanging Textile model eeeeeeeeeeeeeeee 6 2 2 Modifications of the porous models by Rayleigh and Delany Bazley 7 2 3 Results and discussio Menantu neee E E ure pae dina rbd eue udi qu etia 9 2 3 1 Results from the models for free hanging textiles csse 9 P Differences between the MFT model and Ingards Volume absorber 10 2 349 A rough estimation of the edege effect sse 11 3 iABSORPTION IN ODEON 25 12 51 8 0 tI III ti rie
19. el or Notepad Commonly used operations and options are also available from the Toolbar 4 2 The menus The File menu Selecting New will open a new instance of AbsTex Settings for the textile mass airflow resistivity porosity etc the name and description may be saved to a textfile for later usage or distribution to others using the Open and Save options This text file is editable in for example Notepad The lines in the textfile are 1 Name Description Thickness in mm Airflow resistivity in kPa s m Mass in g m Incident angle in degrees Porosity Width of the banner in m Height of the banner in m pe e cde re 9 The user may Export the results and model information to a text file or to a Microsoft Excel file if MS Excel is installed on the system The Generate report option will export the textile data results and graphs to a Word document if MS Word 1s installed on the system However on some rare cases MS Word is not allowed to be started by other programs so Word may need to be started by the user The plot may be exported to a file in PNG JPEG GIF or TIFF format JPEG is not recommended due to its high compression of information and thus loss in quality and should only be used if the target application only can handle jpg files The image will be exported with dimensions of 552 414 pixels 19 SOUND ABSORPTION IN FREE HANGING TEXTILES MAGNUS OGNEDAL The Output menu The program can calculate the Dissipation
20. fferent T li Measured With diffraction estimation Without diffraction estimation 0 8 0 7 di CRT nz x um o d 0 4 e p U 90 1 92 OoO0D 02 0 3 QE MED E L pu r Pu 3 4 o r P d d 4 Sof a 0 2 x E we ee p 4A 1 I 1 Ld i i Ll ui rg 1 Ll rog 1 pita f ot eee 1 Ll 4 63 125 250 500 1k 2k 4k 8k Frequency Hz log Figure 8 The dissipation coefficient for Velour with mass correction In addition to the correction of mass we may assume that the diffracted sound waves may hit the backside of the textile This results in a higher or in some cases lower dissipation coefficient for the textile The addition or subtraction may be expressed as 13 SOUND ABSORPTION IN FREE HANGING TEXTILES MAGNUS OGNEDAL H C log x 2 3 3 where Ho is 1 m K 8 as with 2 3 1 and K2 is also found with an iteration process Kz 2 5 seems to fit well The corrected dissipation coefficient is then expressed by p p 1 C 2 3 4 The basis for 2 3 3 and 2 3 4 is that high frequency sound waves will bend more around the edges and therefore hit more of the backside of the textile resulting in higher dissipation coefficients Low frequency sound waves will only pass the textile resulting in lower dissipation coefficients The effect of this correction is seen in Figure 9 T Mea
21. implification of wave around an oblong textile in 3D left and 2D right Blue indicates the sound sound wave However if we assume that this correction of mass is dependent of the difference in phase between the incoming and the diffracted wave the correcting of the mass has to be frequency dependent For example it may be assumed that if the difference in phase is 10 the mass should be reduced by 90 It may also be assumed that the waves will diffract around the 12 SOUND ABSORPTION IN FREE HANGING TEXTILES MAGNUS OGNEDAL shortest dimension because this is the easiest way around the banner The banner is assumed to be in the form of a rectangle with an area of H L where H is the shortest edge and L 1s the longest edge From Figure 7 it can be seen that the diffracted sound wave has gone H longer than an incoming wave so the difference between the incoming and diffracted sound wave can be written as A A KH 2 3 1 where 4 is the wavelength 4 co f of the sound wave in m and K is found in an iteration operation This expression seems to fit best the measured values when K 8 The mass used for calculation may therefore be written as m m 1 C 2 3 2 where m is the original mass In Figure 8 this effect is calculated for the Velour textile It may seem like this correction has little impact on the overall performance but for other textiles with different masses and dimensions the results may be di
22. l D5 and E5 The scatter coefficient should be set to 0 7 Question May I request a new feature in AbsTex Answer Yes please do Send an e mail to magnus ognedal com and explain your idea If your idea is good and or there is a demand for this feature the feature will be incorporated in a new version of AbsTex Question I think I ve found a bug What should I do Answer Please e mail the error message if any and details of your operating system computer and settings in AbsTex to magnus ognedal com 4 4 System requirements installation and upgrading AbsTex will run on IBM compatible computers using Windows 98 ME NT 2000 or XP as long as NET Framework 1 1 is installed Computers using Windows XP with Service Pack 2 is likely to have this installed The Framework can be found on the internet 4 The computer will run on virtually any computer but the higher system specifications the better performance AbsTex will run on a computer with a Pentium 75 MHz processor and 32 MB RAM but it s not recommended since calculations and drawing of the graphs is very slow AbsTex was also tested on a laptop with a Pentium II 300 MHz processor and 64 MB RAM and it performed well although the plot used some time to be drawn The installation is started by double clicking on AbsTex Setup exe found in the AbsTex directory on the CD or using the Install AbsTex option on the CDs autorun program The setup program will search for the
23. n Is it possible to hold the current values in the plot Answer Yes the Custom Comparison function can do this Just go to the Comparison menu and select Custom Comparison Click the OK button and the current calculation will be set as the Custom Comparison If you want to update it quickly this 1s possible with a series of quick key combinations Alt C C and Enter will turn Custom Comparison Alternatively you can start several instances of AbsTex Question Why do I not see the entire graph when choosing Absorption coefficient or Transmission coefficient under Output Answer The maximum y value of the plot is set to 0 7 by default Some of the calculated coefficients for these settings are usually higher than 0 7 and you may need to set it to 1 0 This 1s done with the Set maximum y value in the Plot menu Question How can I calculate input data for use in programs such as Odeon Answer The following procedure is recommended Switch from 3 octave band to 1 1 octave band Turn on the edge effect estimation Calculate the dissipation and transmission coefficients using the MFT model and copy or export them to for example MS Excel Then use the dissipation coefficient as absorption in Odeon and set the transmission to the squared average of the transmission coefficient for the 500 Hz and 1000 Hz octave band in Excel this is calculated as AVERAGE D5 E5 2 if the transmission coefficient values for 500 Hz and 1000 Hz are placed in cel
24. ng of the sound wave For lower frequencies it may be assumed that the sound pressure behind the sample is reduced while the airflow resistivity and the velocities for the banner stay the same This may be interpreted as one option To reduce influence of the mass For free hanging textiles with larger dimensions than the textiles measured here the effect of diffraction will be reduced and we may expect to find an increased low frequency dissipation If the textile has smaller dimensions than the measured textile the effect of diffraction will be more significant 11 SOUND ABSORPTION IN FREE HANGING TEXTILES MAGNUS OGNEDAL Mass 0 g m2 Mass 10 g m2 Mass 50 g m2 Mass 100 g m2 Mass 500 g m2 7 Mass 1000 g m2 0 8 Mass 2500 g m2 Mass 5000 g m2 0 9 0 7 0 6 2 OUVE 63 125 L f 1 L L L L L L 1 L L L L 250 1 L IL 1 1 L 1 1 500 1k 2k 4k 8k Frequency Hz log Figure 6 Dissipation coefficient for a textile with Rs 2p c and different masses Calculated for normal incidence using the MFT model If the weight of the banner is greater than ca 400 g m trial and error have lead to the suggestion that the dissipation coefficient for frequencies lower than ca 900 Hz fits better to the measured data if the mass for the textile is reduced by for diffuse field calculations KE N bi bm CN Figure 7 S
25. ol Serge 6 Super Wool Serge 7 Kilo Serge 300 500 140 530 650 500 1000 A typical inexpensive scene textile A tighter textile Used as variable absorber strong textile used as cover for mineral wool in e g slatted panels Strong fine textile Used as scene curtain Used as variable absorber Used as variable absorber Used as variable absorber MAGNUS OGNEDAL SOUND ABSORPTION IN FREE HANGING TEXTILES MAGNUS OGNEDAL 1 2 Measured absorption and transmission coefficients The seven textiles were measured in the reverberation chamber at NTNU When measuring the absorption coefficients the textiles were hanged freely in the middle of the reverberation chamber When measuring the transmission coefficient the textiles were mounted in a window opening between the reverberation chamber and the adjacent room Lydrom 4 see 1 1 00 0 90 0 80 4 Scene Molton 0 70 c Ullseviot 0 60 Fibertex F2B Q 9 Velour c 0 50 A M1 Wool Serge o z Super Wool Serge 5 0 40 Kilo Serge m g lt 0 30 0 20 0 10 4 0 00 L e e O o LO e e e lo O e e e O o e e e e O oc e e LO co co O CN co eo LO O e e Q O 10 e e e LO O O e e v v v CN CN oO st LO co co O C co O LO O o e Q c N N e t LO co eo Frequency Hz Figure 1 Measured one sided absorption coefficients for all seven textiles freely hanged Measurements
26. performed in the reverberation chamber at NTNU 1 00 0 90 0 80 1 Scene Molton 0 70 2 Ullseviot 0 60 3 Fibertex F2B 4 Velour 5 M1 Wool Serge 6 Super Wool Serge e T eo 7 Kilo Serge Transmission coefficient O Cc eo 0 00 e eo e ce LO e ce e LO ce ce eo ce ce e ce ce ce e e ce ce ce LO co co e CN co je LO w e e eo eo e Ww e ce e Ww ce ce eo ce v v v CN N eo LO co co e CN co LO v e e eo e v v CN CN eo LO co co Frequency Hz Figure 2 Measured transmission coefficients for all seven textiles Measurements performed in the reverberation chamber and Lydrom 4 at NTNU SOUND ABSORPTION IN FREE HANGING TEXTILES MAGNUS OGNEDAL 1 3 Airflow resistivities dimensions and weights Measured widths lengths and areas are presented in Table 2 The airflow resistivity weight and thickness for all banners were measured see Table 3 Table 2 Lengths heights and areas of the textiles Textile Height m Length m Area m7 1 Scene Molton 0 80 1 95 1 56 2 Ullseviot 1 5 2 20 3 34 3 Fibertex F2B 1 16 2 10 2 44 4 Velour 1 47 2 16 3 18 5 MI Wool Serge 1 10 1 53 1 68 6 Super Wool Serge EIS 1 52 2 7 Kilo Serge 1 14 1 50 1 71 Table 3 Measured airflow resistivities weights and thicknesses Textile Airflow resistivity Weight Thickness E 3 g m4 mm m l Scene Molton 2994 363 2 0 99 2 Ullseviot 647 ZE
27. sipation coefficient in blue and measured absorption coefficient in black The calculated dissipation effect is calculated by the MFT model with weights as given by the manufacturers Table 1 and the measured airflow resistivity Table 3 The comparison for the measured banners 1 7 is only available when the output is set to Dissipation in 1 3 octave bands For other comparisons use the Custom comparison option Custom comparison displays a new window that lets the user paste data from for example Microsoft Excel for comparison The data has to be tab delimited in 1 3 or 1 1 octave bands dependent of the choice in the Output menu Set the cursor on the textbox in the upper left corner and press Ctrl V or right click and select Paste Press the Transfer to memory button If the import was successful the form should look something like Figure 16 Then press the C ose button If Current calculation 1s selected the data from the last calculation is used MAGNUS OGNEDAL File Output Quickoad Compare Plot Help New Open Ctrl a Save As Ctrl 5 Export k Exit Ctrl Q File Output Quickload Compare Plot Help Parameter k vy Dissipation coefficient 1 R T Edge effect Absorption coefficient 1 R Octave band Reflection coefficient R Transmission coefficient T File Output Quicdoad Compare Plot Help 1 Scene Molton 2 Ullseviot 3 Fibertex F2B 4 Velour 5 M1 Wool Serge 6 Super Wool Serge 7
28. sured a With diffraction estimation Without diffraction estimation 0 9 0 8 ail 0 7 t n 0 6 NL i oe C o5 n 5 o 0 4 n d P Qn Pu ot puo 0 I I 63 125 250 500 1k 2k 4k 8k Frequency Hz log Figure 9 The dissipation coefficient for Velour dissipation correction 14 SOUND ABSORPTION IN FREE HANGING TEXTILES MAGNUS OGNEDAL These two methods of correction may be regarded as semi empirical In 2 3 1 the height 77 is multiplied with 8 to move the reduction to the desired frequency range The same adjustment is done with 2 3 3 These coefficients are based on calculations for all the seven textiles and the K values that gave the best results where chosen The reason why these adjustments are needed may be because the expressions have to account for different sound incidences and the assumptions stated earlier may be wrong It must be emphasized that this 1s a generalization based on the measured coefficient for only seven banners and for only one size per banner For example for Scene Molton the measured values are reduced for frequencies above ca 4 kHz This 1s a situation not accounted for in these expressions Very oblong textiles for example length of 20 m and width of 0 5 or 1 m the edge effect may be over predicted for higher frequencies To predict the effect of diffraction more precisely more textiles with different areas should b
Download Pdf Manuals
Related Search