Home

The HIFI User`s Manual - Herschel

1. running a status El progress 0 E Console x Zo Processing spectrum 1340 Processing spectrum 1341 Processing spectrum 1342 Processing spectrum 1343 k Processing spectrum 1344 HIPE gt zl EH o 407 of 3495MB_ Here an HTP MyHTP is being exported to MyClassFile The blanking value has been left at the default value while values for the source position and velocity are given as well as a frequency ref erence frame Figure 11 1 The hiClass task GUI 11 3 How to read HIFI data in CLASS First make sure you use a recent version of CLASS The version from august 2009 works nicely and we can assume that any posterior version will work as well It is important for you to use a recent version because Old versions use Fortran 77 and will not be able to dynamically allocate the memory needed to read big spectra like WBS ones 8000 channels Old versions do not know about the subscan number and will not be able to make any difference between the different subbands of a spectrum Old versions have troubles with reading double precision values from FITS files Some versions first half of 2009 have a broken code which totally prevents reading any FITS file with a long header CLASS is not able to work directly within FITS files So you have to
2. ce When a Product is selected for display the bottom part will show a loading datasets message as long as the Product is being processed Each SpectrumDataset found in the Product is added to the selection panel The location of the divisor between both panels can be changed through drag drop interaction Clicking on one of the little black arrows displayed on the left edge of this divisor extents a single panel to its full size 5 4 2 Selecting Spectra The attribute columns in the selection panel can be used to find spectra that one wishes to plot A single click on a header of such column sorts the rows according to that column s entries Clicking it again inverts the sort order A double click removes the sort and therefore brings the ordering back to its initial state With drag and drop the columns themselves can be reordered A right click on one the headers shows a dialogue box with a selection list of all column headers With this list the columns can also be reordered or even hidden from view Hold the shift button to hide display a whole range of columns at once Furthermore specific spectra can be selected by applying a filter on the attribute columns Open the filter panel by selecting Dialogs gt Filter from the right mouse click menu or by clicking on the filter icon in the button toolbar at the top of the HIPE screen Specify the attribute name from one of the column headers and enter the filter values that
3. Editor x t obs refs dataset x 0 0 5 3800 4000 4200 4400 4600 4800 5000 5200 n ng MHz SEDER no oe ey oe oe oe a EE ITE true 6031 1 19984 3 1522 Log El Console x poo lname 268435841_obs storage ProductStorage pool PoolManager getPool poolname storage register pool query herschel ia pal query AttribQuery herschel ia obs ObservationContext p 1 result storage select query obs storage load result 0 urn product Caption Example of a HIFI spectrum TableDataset which contains the rowflag column with a value of 256 12 Chapter 4 Quality Flags Last updated 10 Feb 2010 Quality Flags are raised during standard processing of HIFI data Flags should be created from every processing step of the pipeline from the initial creation of the HifiTimelineProduct Level 0 through to the final product of Level2 processing If all goes well the flags will have their default values but if a certain processing step is unable to perform the action it was designed for the flag will take a different value If the pipeline produces a flag other than the default value this flag is promoted to the Quality Report Thus the quality report is by definition a list of things identified as have gone wrong A quality report is found from the ObservationContext obs refs quality product Please note the difference between a quality flag and flagging data In flagging data you identi
4. FPU MIXER CURRENTI lIekef iWrent FPU Check Mixer current is Out Of Limit false 2xnom value for SIS maybe a degraded baseline 30uA 60uA for HEB maybe a degraded baseline quality Quality Flags Quality Flags with specified thresholds Range Consequences for sci ence data Action required Check Mixer current variance is Out Of Lim it false FPU MIXER CURRENT VARIANCE mixerCuragitVarlagcedetFPU baseline quality Check Mixer Voltage is Out Of Limit false FPU MIXER VOLTA GB6mixelWeolt gi Yy FPU nom value 100guV maybe a serious prob lem Inform engineering team Check Mixer Magnet Current is Out Of Lim it false FPU MIXER MAGNE nGtHiRRHNI 96ixerMa nom valuex1 04 eimest Timereautt ARRUN stable FPU MIXER MAGNE I nRMSISMANCE magnetRtsitqgrodle PU Inform engineering Check chopper mea sured values differ from the commanded false nom_offset 0 05 V readout problems Check Mixer Magnet nom_valuex1 2 team Resistance is Out Of Limit false FPU CHOPPER fpuCh pper ofERS3OO 05 V potential pointing or Check other pointing out of limit flags Check Diplexer Resis tance is Out Of Limit false FPU DIPLEXER RESIS edNGHeifl xerResis nominal valuex1 2 faeces pPbblem Inform engineering team FPU LNA Ina FPU Check IF Amplifier values are Out Of Lim it fals
5. 10 3 Making a Spectral Cube via the com mand line Some examples of usage are below Data selection Make cubes for all the subbands then display the first one cubes doGridding htp htp cubes count len cubes cube cubes 0 Display cube Or you might automatically create a separate variable for each cube as in the following routine get a separate variable for each cube computed for each subband for subband in range len cubes cube cubes subband subband cube meta subband value cube_name cube_ d subband vars cube name cube The medata of each cube will include a subband parameter stating the subband of the spectra which was used to compute the cube This can be checked with print cube meta subband 34 How to make a spectral cube You may select just a part of the spectrum for each subband to be processed that is to generate the cube for a range of the channels of the given spectra This can be done by providing a channels input which is an Int2d array This has to contain as many rows as subbands are to be processed Each row must have two elements the start and end channel to be read The next example shows how to create a cube for the first and fourth subbands of a given spectrom eter reading just the channels 200 to 1200 in the first one and the channels 400 to 700 in the second channelRanges Intd2 channelRanges append Intid 200 1200 0 0 means appe
6. 11 append 1 repeat the above for the 2nd plot to be overlaid valid flux2 where IS FINITE l 2LayerXY freq2 valid flux2 valid 11 append 1 define the plot layers that have just been created p layers 11 get out of batch mode This actually creates the plot p batch 0 And this is how some common features of the plot are modified p setYrange 0 1 5 p setTitleText This is an example plot 5 3 Viewing with SpectrumPlot It is also possible to display spectra without taking apart the data format as is described in the previous section All Herschel spectra types can be displayed with the Spect rumP lot package If spect rum is a Herschel Spectral type Spectrumld Spectrum2d then splot SpectrumPlot spectrum useFrame 1 will simply display the spectrum along with some standard header information The useF rame 1 allows for the possiblility of creating a plot without actually viewing it at first but as the last step The SpectrumPlot module is build on PlotXY and so many of the features you would use in Plot XY you can also use for Spect rumP lot Below are a few examples from herschel ia toolbox spectrum gui import SpectrumPlot from herschel ia gui plot renderer StyleEngine ChartType import HISTOGRAM LINECHART Creating the plot Sp SpectrumPlot spectrum useFrame 1 adding a second spectrum to the plot sp add spectrum2 Start fresh again p SpectrumPlot spectrum useFrame 1 get graphs g0 p
7. Better to wait until you know you have something you want to save Another thing to note from this is that pipelineout will become very large and you should delete it from time to time simply delete the directory with rm 2 2 1 hifiPipeline task in the GUI Opening the hifiPipeline GUI The hifiPipeline task is run from the GUI in the following fashion Click once on an Observation Context in the Variables pane and the hifiPipeline Task will appear in the Applicable Tasks folder double click on it to open the Task dialogue in the Editor view Alternatively open the hifiPipeline Task by double clicking on it under the Hifi Category in the Tasks view A Hifi Pipeline View is also available from the HIPE Window menu under Show View but it is not fully implemented yet Running the HIFI pipeline File Edm Run Window Help FETE eN arpw x 56g C Editor x 2a 2 x 2a di Tasks x 28 O hifiPipeline x X An s Applicable iow fas go to expert mode E 2 E ObservabonConte xt MyObs st simplefitsWriter Instruments wi HRS H vi MRS V yi W8S M yi WaS V un K By Category fromLevel 0 0 upToLevel 2 0 palStore lt Novariable gt hrsAlgo e wh sAlgo e level Algo e level Algo e Output Variable name for obs obs Info yeady status r J Clear Accept J D Console x a HIPE laj
8. fastQDC Fast Quantization Distortion Correction processed Not optimal false NOPOWCOR noPowerCorrection No Power Correction could be processed false Level 1 0 Quality Flags Check data structure Quality Flags OBSERVINGMODE observingMode Observing mode not recognized consult the pipeline configuration xml file false UNKNOWNBBTYPE unknownBbType Bbtype not known false Check freq grid Quality Flags FREQUENCYDRIFT maxFreqDrift Unacceptable maximum drift in the frequency grid detect ed false FREQUENCYCHECKS noFreqChecks Frequency checks and or frequency grouping failed false Check phases Quality Flags CHOPPERPATTERN chopperPattern Pattern observed for the Chopper not as expected in all datasets false CHOPPERV ALUES chopperValues Number of distinct Chopper values not as expected in all datasets false LOFPATTERN lofPattern Pattern observed for the LoFrequency not as expected in all datasets false LOFVALUES lofValues Number of distinct LOF values not as expected in all datasets false BUFFERPATTERN bufferPattern Pattern observed for the buffer not as expected in all datasets false BUFFERV ALUES bufferValues Number of distinct buffer values not as expected in all datasets false PHASECHECKS noPhaseChecks Not all phase checks could not be carried through or com p
9. 29 Sideband Deconvolution The deconvolution tool is run AFTER the level 2 pipeline The level 2 pipeline performs the following tasks splits the data into upper and lower sideband representations applies a gain correction specific to the LO frequency and sideband of the spectra corrects frequencies for spacecraft radial velocities resamples the spectra onto a fixed grid For WBS this is done at 0 5 MHz spacing with the first frequency snapped to the nearest 0 5 MHz Any HIFI observation context will contain Level 2 products if run through the standard product generation 9 2 Running the Deconvolution Tool Assuming the observation context is named MyObsContext the user can run the deconvolution task on the command line with the default parameters by simply invoking result doDeconvolution obs MyObsContext The full range of parameters and their defaults are as follows decon_result doDeconvolution polarization 0 bin_size 5 0E 4 max_iterations 200 tolerance 0 0010 gain 0 channel_weighting False ignore_mask 524288 plot_dsb 0 use_entropy False lambdal_channels 0 0 lambda2_gains 0 0 cont_offset 0 0 expert polarization Observations contexts store H and V polarisations You can specify which to decon volve with this option O H 1 V bin size Tells deconvolution the sampling interval of the single sideband solution A value of 0 5 MHz is recommended to match the WBS sampling max iterations Tell
10. HIPE gt MyObs getObservation 1342180798 i Il Jython Interpreter 100 71 of 3495 MB The hifiPipeline task appears in the Applicable Folder in the Tasks view after clicking on the Ob servation Context MyObs in the variable view Figure 2 1 HIFI pipeline task default view Running the hifiPipeline GUI The default or basic dialogue allows you to re process an already existing observation context e g from the Herschel Science Archive through the pipeline The default set up of the pipeline is to re process data from level 0 to 2 for all four spectrometers or as many as were used in the observation The way the data is to be reprocessed is defined in the Input section 1 Ifthe hifiPipeline task was opened from the Applicable Tasks folder then the ObservationContext selected in the Variables View will automatically be loaded into the Task dialogue and you will see its name by the observation context bullet which will be green Alternatively drag the name of the observation context to be reprocessed from the Variables view to the observation context bullet 2 Select the spectrometers you wish to process data for by checking the desired instrument s and polarisation s Both H and V polarisations of both the Wide Band Spectrometer WBS and High Resolution Spectrometer HRS are checked by default 3 Select which levels to re process from and to via the drop down menus By default the pipeline will process level 0
11. containing different numbers of segments the invalid segments are disabled and displayed with a grey x An example is shown in the figure above 5 4 4 Button Bar Helak OUAIS E y Ne T m Y EE At the top of the HIPE screen the SpectrumExplorer buttons following the New and Open File buttons have the following meaning button 1 save the plot as a PNG PDF EPS or JPEG file button 2 send the plot to the printer button 3 zoom mode This is the default mode when SpectrumExplorer is started Change the hori zontal and vertical plot ranges by drawing a rectangular box using the left mouse button Control left mouse button will un zoom the plot or use the Autorange option under the right mouse button button 4 select spectra A clicked spectrum will be displayed with a bold line Any operation such as the Tasks under the right mouse button will then only apply to this particular spectrum Also 21 Viewing Spectra 5 4 5 the selected spectrum can be dragged to a new panel note that dragging to the left and top of the original panel is not possible The spectrum can also be dragged to the Variables window where it will be stored as a new variable button 5 pan mode Pan through the spectrum by clicking the left mouse button and moving the mouse If one only wants to pan along the x or y axes click on the axis with the left mouse button and then move the mouse or use the mouse wheel button 6
12. convert the FITS file into a CLASS file file out MyHIFISpectra hifi mul fits read MyHIFISpectra fits Now you have a CLASS file named MyHIFISpectra hifi you can use whatever you want as an exten sion you can access like you always do in CLASS file in MyHIFISpectra hifi eine 43 Exporting HIFI data to CLASS get first SEE Wingite dr b plot 44 Chapter 12 Memory Issues Last updated 28 Feb 2010 On occasion one can run into the java heap space error when using HCSS software especially when running the pipeline Here are some things to help 1 User release Choose the Advance installation and increase the maximum amount of memory available to HIPE the User installation allocates 1 Gb by default 2 Modify the memory allocation java vm memory min and java vm memory max in hcss Hipe props 3 The garbage collection command System gc is also useful to force clearing memory HIPE will automatically do this when memory becomes too full 4 Swap Store Properties Itis possible to use the hard disk as swap space to preserve the memory available in HIPE and HIPE does this by default The following properties are defined to preserve computer memory This becomes especially useful when pipeline processing long observations on a laptop or on a pc with a 32 bit Operating System TBC and with average or limited memories capacities However any Task that uses or changes any HifiProduct e g HifiTimel
13. let s play with the result cubes each cube is a SpectralSimpleCube which in its turn is an SpectrumContainer hence we profit all the spectrum toolboxes arithmetics statistics etc and we can e g directly obtain a point spectrum as for any other SpectrumContainer Hb db db db OSE dk row 0 column 10 spectrum cube getPointSpectrum row column print spectrum getLongitude print spectrum getLatitude print spectrum segmentIndices you can check that the cube hence its spectra has a single segment or subband segment spectrum getSegment 0 PTOL segment spectrum getSegment spectrum segment Indices 0 plotSpectrum PlotXY segment getWave segment getFlux xtitle Frequency MHz ytitle Intensity 39 How to make a spectral cube There are several ways to visualize a cube such as the CubeSpectrumAnalysisToolbox cat CubeSpectrumAnalysisToolbox cube you can also visualize it with the SpectrumExplorer since the cube is an SpectrumContainer or simply display it as a cube of images display Display cube Optional inputs for the Gridding task Most of the optional inputs of the DoGridding task are also applicable to the Gridding task namely weightMode filter Type mapSize refPixel refPixelCoordinates pixelSize smoothFactor filterType filterParams detail extrapolate and the input Wcs In addition there are other optional inputs which are specific to this task namely co
14. of filter are thought to optimize the convo lution Specify the size of the pixels The user can choose a pixel size different from the pixel size computed by default based on other inputs and on the angular dimensions of the observed area The pixel size must be given in seconds of arc For example to assign a pixel size of 20 arcsec along both axes the user can specify the pixelSize input 36 How to make a spectral cube cubes doGridding htp htp weightMode selection filterType gaussian pixelSize Doubleld 15 And to assign a different pixel size along the x and y axis the given Doubleld must have two elements For example to get pixels 15 arcsec wide and 25 tall the pixelSize input should be Double1d 15 25 cubes doGridding htp htp weightMode selection filterType gaussian pixelSize Doubleld 15 25 By default the pixel size is computed so that it is optimal based on the other parameters given to the task If neither beam size nor smooth factor have been provided the task will compute a default HPBW and then it will choose pixel size equal to the half of this HPBW i e it will assume that the sampling was done with following the nyquist criterion The default pixel size will be the biggest of the values HPBW 4 and HPBW 2 smoothFactor By default the smoothFactor is 1 0 no smoothing factor applied so that the default pixel size becomes the half of the beam size If the map dimensions in
15. pixels were specified the pixel size will be simply the division of the area actually observed by the number of pixels specified in the map size parameter If an smooth factor is provided the pixel size will be the largest of HPBW 4 and HPBW 2 smoothFactor Specify the dimensions of the map The user can specify the size of the map in pixels by means of the mapSize parameter For example cubes doGridding htp htp mapSize Intid 10 20 cube cubes 0 will create a cube 10 pixels wide and 20 pixels high When this parameter is not specified the task computes the optimal dimensions taking into account the antenna beam size as well as the area of the sky covered by the input spectra Specify the reference pixel The user can specify which is the reference pixel of the grid It is also possible to define the coor dinates of that reference pixel If the latter is not provided the reference pixel will provide the co ordinates measured in pixels of the projection centre which is in its turn computed as the center of the coordinates of the input spectra usually the centre of the map Hence if the user provides a reference pixel the user is defining where in the regular grid lies the centre of the observed area Please note that the convention for the pixels computed for the regular grid of the output cubes is that the 0 0 pixel corresponds to the center of bottom most left most pixel of the regular grid Please note that t
16. result can be viewed with the product viewer The single sideband result ssb is a dataset that can be viewed with the SpectrumExplorer On the command line it can be extracted from the product as follows ssb2decon result ssb This contains the deconvolved spectrum and is the primary output of the tool The dataset gain can be viewed with dataset inspector On the command line it can be extracted from the product with gains decon result gain The deconvolution tool can estimate 31 Sideband Deconvolution the sideband gains due to the redundant nature of the data taking These estimates are stored per LO tuning in this product The meta data added to ssb includes number of iterations and the tolerance as can be seen in the HIPE screenshot below 000 X HIPE Herschel Interactive Processing Environment File Edit Run Window Help fe decon result x 9 HistoryScript HistoryTasks 9 HistoryParameter on amp o c T Spectrumid name value unit description wavename freq Actual name of the WaveColumn waveunit MHz Units of the WaveColumn wavedescription Single Sideband Frequency Description of WaveColumn bbin size los max iterations 200 tolerance 0 0010 Data decon result decon resul t ssb E f 9 gain gt amp History 10 T T T T T T freq MHz spectrums 2 5 510 5610 5710 5810 5910 601
17. since no off baseline data available false DATALOSSINAVERAGRE average Some data has been lost while computing the average over many datasets false Chapter 5 Viewing Spectra Last updated 19 Dec 2009 5 1 Introduction HIFI spectra can be visualised in several ways at various levels of sophistication and user friendliness Here the PlotX Y and SpectrumExplorer packages are described 5 2 Basic Spectrum Viewing the PlotXY Package PlotX Y is the basic package to plot arrays of data points in the HCSS and it can be used to plot HIFI spectra as well It has a lot of options making the plots highly configurable Here is an example of plotting a HIFI spectrum e Get the frequency and flux data to be plotted from the spectrumdataset sd freq sd getWave get 0 flux sd getFlux get 0 The simplest possible plot out PlotxyY freq flux When plotting multiple spectrum datasets say sd1 and sd2 in one figure get the wavelengths and fluxes to be plotted freql sdl1 getWave get 0 fluxl sdl1 getFlux get 0 freq2 sd2 getWave get 0 flux2 sd2 getFlux get 0 create the plot variable p P1lotxyY create the plots in batch mode p batch 1 define the layer variable 11 remove any non numbers NaN s Infinites etc valid fluxl where IS FINITE create layer for first plot l 2LayerXY freql valid fluxl valid 17 Viewing Spectra append to layer variable
18. subtract this from each scan DEFAULT process each scan separately doglue False Determine SW on individual sub band spectrum This is desired for HRS but often not for WBS as long period SW can only be determined on the combined spectrum DEFAULT doglue True usermask mask frequency ranges in addition to the automatically determined mask Example usermask 537 0 538 0 539 539 5 masks the ranges 537 538 GHz and 539 539 5 GHz DE FAULT only use automatically determined mask 28 Chapter 9 Sideband Deconvolution Last updated 1 March 2010 9 1 Introduction to doDeconvolution The deconvolution tool is the post Level 2 processor to separate the folded double sideband DSB data inherently produced by the heterodyne process into a single sideband SSB result See the figure below Fluxes F_DSB in the DSB spectrum are given by F_DSB nu_IF g_u F_sky nu_LO nu_IF g_1 F_sky nu_LO nu_IF where nu_LO nu_IF are sky frequencies and g 1 and g_u are sideband gain imbalance factors typically close to 1 The deconvolution is used to reduce WBS Spectral Surveys which are collections of observations taken at many LO settings so as to constrain the solution The algorithm finds a SSB solution that best models the observed DSB observations through iterative chi square minimization Comito and Schilke 2002 Synthetic Spectrum 800 0 GHz 804 816 0 GHz 812 Double sideband spectrum
19. to be appropriate to the frequency at which the observation is carried out but you may wish to simulate a different beam size beam Doubleid 40 0 Drag this to the beam bullet xFilterParams yFilterParams the appropriate values for these depend on the filter being used box filter or the default Gaussian see the next section for more notes Here an example appropriate for a box filter xFilterParameters Doubleld 0 5 yFilterParameters Doubleld 1 5 Figure 10 1 The doGridding task GUI form As with all GUI forms in HIPE clicking accept will start running the task The outputs you will be most interested in are the array containing all the cubes created default name cubes and the map context that allows you to easily browse and view the cubes default name cubesContext You can view these cubes with the SpectrumExplorer and the CubeSpectrumAnalysisToolbox There are also other output produced xPoints and yPoints give the offsets measured in radians with respect to the projection centre The convolutionTable notes which spectra have contributed to each pixel but is only generated when the detail tab in the expert GUI is checked Clicking on the expert button will toggle to a version of the GUI designed for those who want to really redesign their cubes There are many more options available and they can be passed to the GUI in the same way They are discussed in the context of the command line in the next section
20. works on HifiSpectrumDatasets or HifiTimelineProducts The task uses the keywords sds for HifiSpectrumDataset and htp for HifiTimelineProducts The conversion of frequencies is done using the to keyword The following table shows the various possibilities to Description Other keywords necessary frequency Converts to the Intermediate None Frequency scale usbfrequency Converts to the Upper side band None Frequency scale Isbfrequency Converts to the lower side band None Frequency scale velocity Converts to the velocity scale in reference reference frequency km s inupper True or False 25 Chapter 7 The Spectral Toolbox When a spectrum is active has just been created or has been selected in the Variables View HIPE automatically becomes the Spectrum Toolbox Your spectrum can be viewed with SpectrumExplor er and this can serve as your launching point for spectral analysis By right clicking on a spectrum in SpectrumExplorer and selecting the Tasks sub menu you can access the tasks in the Spectrum Arithmetics Toolbox and to the Spectrum Fitter Toolbox The tasks in these Toolboxes are also found under Applicable Tasks These are general tools available to all Herschel instruments and are described in Chapter 5 of the Data Analysis Guide 26 Chapter 8 HIFI Standing Wave Removal Tool Last updated 3 March 2010 8 1 Introduction to FitHifiFringe FitHifiFringe is a t
21. 0 6110 6210 6310 6 410 CEN completed level 1 WBS V ltart running level 1 to level 2 processing for Quality_level0 WBS H WBS V start running level2PipelineTask for apid WBS H lleveli WBS H stored into pipeline out hifi ops cal hifi ops aux Thu Nov 05 12 02 47 PST 2009 completed level 2 WBS H start running level2PipelineTask for apid WBS V llevell WES V stored into pipeline out hifi ops cal hifi ops aux Thu Nov O5 12 03 08 PST 2009 lcompleted level 2 WBS V Nevel2 stored into pipeline out hifi ops cal hifi ops aux Thu Nov O5 12 03 10 PST 2009 ipeline Task finished PE decon result loDeconvolution obs obs polarization 0 bin size 0 5 nax iterations 200 tolerance 0 0010 plot dsb 0 IPE gt r Variables x a Tasks x zZ2rees ene zu decon resut ofa Ba E Applicable doDeconvolution hifiPipeline level Pipeline level2Pipeline localstoreWriter simplefitsWriter By Category 900000 Do Nee 32 Chapter 10 How to make a spectral cube Last updated 1 March 2010 10 1 Introduction to doGridding 10 2 Spectral cubes from OTF mapping observations are produced as part of the SPG pipeline and are a level 2 product However re processing of spectral cubes from a Level 2 product is likely desirable this is done using the doGridding Task after cal
22. DSB model spectra are derived But when the input data are of poor quality sparse sampling or con tain few lines repetitive noise structures may appear in the solution and or the fitted gain values may begin to diverge and become non physical Since the entropy of these artifacts is low we compute the inverse of the solution entropy and add it to the Chi square value at each iteration In this way the deconvolution must still match the observations but has the additional task of keeping the entropy of its solution high yielding a non highly patterned result Turning on the entropy terms helps the deconvolution behave The two lamda factors are the two relative weights of the entropy terms for the channel solution SSB spectrum and the gain values To use the maximum entropy lambda terms set the weight low e g channel weight to 10 5 and gain weight to 0 Slowly increase either of these weights 10 4 10 3 and look for an improvement The weights should not exceed 10 1 The Deconvolution Tool can also be run from a GUI by clicking on the the obs context in the Vari ables window then double clicking doDeconvolution in the Task list Like other GUIs in the system once the Accept button is hit the command line version is written in the console window so users know exactly how the task was called This output can be cut and paste into user scripts for repeatability 9 3 Viewing Deconvolution Results The output product
23. The HIFI User s Manual Hifi Editorial Board Max Avruch Adwin Boogert Tony Marston Carolyn McCoey Michael Olberg Miriam Rengel Russ Shipman The HIFI User s Manual Hifi Editorial Board Max Avruch Adwin Boogert Tony Marston Carolyn McCoey Michael Olberg Miriam Rengel Russ Shipman Table of Contents Le Data usq PEERS 1 1 1 Data frames eet hs eased ins aoe es sects Geese e ep veh 1 1 2 Data Products ner eer arr ipe ne RETO SA EEES RREPS ei Ree 1 ES CONTERIS cuin mp REN EINER DERI AMET REI a yeens 1 1 3 1 Herschel Observation Context 2 teet ertt ses dee Re e 2 2 Running the HIFI pipeline eee t eese tege ce sete reete te cerae Er se ete ye s nag ure 3 2 1 Introduction to the Pipeline 5 eene erm ertt Pere oro ERE Ee eases 3 2 2 How to run the HIFI Pipeline sssee em emm eme 3 22 1 hifiPipeline task in the GUI eet te reete ttr ia 4 2 2 2 The hifiPipeline in the command line ssseA 6 2 3 Running the Pipeline step by step eme 7 2 4 How to customise pipeline algorithms seeeee m HI 7 3 Flags 1n HIET data 5 te t ee ett Pe IR ERE E OE E TRE PO RESO E POP ip ates 9 3 1 Introduction to flags ereire t certet e cete eese ti de cen Er egre eei 9 32 Channel flags 5 iz err rte EIE Ee br e Re ERR RR iia 9 3 3 Column TOWFTlags eben P REPE ect ipei Dee EDIT 10 4 Quality Flags EE 13 De VIEWING Spectta sons ee iie iste
24. art The look of the selection panel depends on the SpectrumDataset type A typical example is displayed in the following picture When the added SpectrumDataset is a SpectralCube a cube visualizer is displayed instead with which spectra can be selected File Edit Run Window Help rie lalhsemm m se iurvr m Y B E i ePeoueBg 7 Editor x i hrsv x TITTY TT Tt 1 0 i a AE a pii c o3 Ts CO v do OC pg gg fy ge 1 0 0 8 0 6 0 4 0 2 0 0 0 2 0 4 0 6 0 8 subplot 0 18 0 34 12345 67 8910111213141516 obs time Chopper MJC_Hor packet ti dftransfer MJC Ver sequenc ACS_18P_V ACS 5M V ACS_5P_V ACS 8P V ACS_Ana AC x x x x x 1626823 4 409 0 027 1626823 315 0 021 0 17 999 5 023 5 036 8 05 1 099 x x xIx x x pc x 1626823 441 0 027 1626823 319 0 021 1 17 999 5 024 5 036 8 051 1 099 x x xIx1 Xx xf x 1626823 4 409 0 027 1626823 323 0 021 2 17 999 5 023 5 036 8 051 1 099 xixixixIxixix x x x x x x x 1626823 4 409 0 027 1626823 327 0 021 3 18 5 023 5 036 8 051 1 099 x x x x x x x x x x x 1626823 4 409 0 027 1626823 331 0 021 4 18 5 023 5 036 8 051 1 099 x x x xix x x x x x x 1626823 4 409 0 027 1626824 335 0 021 5 18 5 023 5 036
25. asks will actually change the data by resetting the frequency to upper lower sideband represen tation or velocity The GUI only changes what is seen in the SpectrumExplorer the data themselves are not changed There are four fundamental ways of representing the frequency scale for HIFI the intermediate fre quency default the upper sideband frequency the lower sideband frequency or by velocity One final note currently the HIFI pipeline is providing the final spectra represented in both USB and LSB The level 2 product names are tagged LSB or USB it is still possible from these spectra to transform back to IF or the other sideband 6 1 1 Changing Spectral Views The SpectrumExplorer provides internal means of viewing spectra These views are only for display purposes and do not change the data 6 1 1 1 LSB USB Assuming you have activated a spectrum in a SpectrumExplorer window To move between a spectrum seen in the Intermediate Frequency USB or LSB right mouse click on the frequency access not the title of the access but the axis itself A pull down menu for the access will appear 6 1 1 2 Velocity 6 1 2 Change Spectral Views from the command line 6 1 2 1 LSB USB The task to convert the actual frequency scale in a HifiTimelineProduct or HifiSpectrumDataset is called ConvertFrequencyTask Assuming spectrum is the variable name for a HifiSpectrumDataset with the frequency scale of the data in expressed as IF frequenc
26. cNominal NaN raNominal and decNominal allow the user to set the central position of the observation he she wishes to export All the positions offsets will be calculated from the point raNominal decNominal This is of primordial importance for maps and if you are exporting map data you are strongly recommended to set these If raNominal or decNominal is not provided then HiClass will try to find it in the dataset or product that you provide At present only the ObservationContext contains this information stored in its metadata parameters raNominal and decNominal these are the values you gave in HSpot The HifiTimelineProduct is supposed to contain a copy of the ObservationContext nominal ra and dec but it does not yet However another set of coordinates is available at that level and can be used to have the same origin for all the datasets contained in the HTP These coordinates are the average coordinates of the entire observation and are stored in the metadata parameters ra and dec Until the HTP does contain the correct coordinates these average values are used veloSource NaN If let to its default value then HiClass will try to find the velocity of the source in the dataset The velocity is expressed in the frame in which the frequencies of the dataset are expressed usually the observatory or the LSR As always the velocity is positive if the source moves away from the observer e specsys If left to its default val
27. can be ranges circular ranges or exact values 20 Viewing Spectra The filters are combined by applying the AND operator Clicking on the green circle next to a filter temporarily disables that filter Clicking on the red cross removes it from the panel X filters attribute fier f df transfer 300 400 eX Chopper 4 44 0 1 eX bbrumber i X ER IL 5 4 3 Displaying Spectra In the general selection panel at the bottom each row depicts an individual spectrum The numbers in the first column show the index of the spectrum within the SpectrumDataset If SpectrumExplorer was opened on a Product the index is preceded by the index of the SpectrumDataset within the Product For example 2 3 denotes the fourth spectrum within the third SpectrumDataset within the Product given that both indices start with 0 Clicking the button in the first column displays all segments in that spectrum A double click removes them from the plot The same accounts for the top row of buttons clicking displays a single segment for all spectra while double clicking removes them from the plot The ALL button in the top left corner of the selection panel displays all segments of all spectra Finally individual segments can be displayed by the clicking the approprate box The colour of the button is changed to the colour of the spectrum displayed in the plot In case a Product is displayed with SpectrumDatasets
28. cea cet EEI beo ETE ee RERO E ask eG oS 17 3 1 Introduction eia ois een E E SE Re RR ER CHE ERR dp RegS 17 5 2 Basic Spectrum Viewing the PlotXY Package ssseeA 17 5 3 Viewing with SpectrumPlot tnit trek ett ege e ae E apre Eat 18 5 4 The SpectrumExplorer Package sssssee em eme 19 5 4 1 Starting the SpectrumExplorer ssssse ceca cena sean ssie 19 3 4 2 Selecting Spectra 4 eut pe ert Rupe net ee debs AE ER 20 25 4 3 Displaying Spectra oio ett bos bes de e gre er te E ETES 21 3 4 4 Button Bat 5i e eoe Resa eere eee eee 21 3 4 5 Plot Interactions on ier enr ERE TEESE TE SAARSO R 22 25216 Raster Panel uou a lee loss in odes Uu DRE Ie tire Mp 23 2 17 Preferences dette Uere aedi eg o E e RCER E eMe 23 6 Changing to LSB USB and Velocity sesssessessesee HH mH eene 24 6 1 Changing HIFI Frequency Scales esesseesseeeeeeeeeeee eene neenene 24 6 1 1 Changing Spectral Views 20 0 0 cece cece ence cece HH re 24 6 1 2 Change Spectral Views from the command line esee 24 T The Spectral Toolbox isr o irt tet tercer e REN eod eR E SER Teed Bea 26 8 HIFI Standing Wave Removal Tool sese He mH meme eene 27 8 1 Introduction to FitHifiFringe 2 0 0 0 eee cece I emen hene 27 8 2 Running FitHifiEnnge eite reet ig ettet rer reb RR itp rre beanies 27 9 Sideband Deconvolution 0205 2 or eget t
29. cur although the pipeline will try to have the best performance possible hcss hifi pipeline product savedisk true This property determines whether an existing observation in the swap store should be overwritten or not It is strongly suggested to keep the value true otherwise the space used in the hard disk will increase in proportion to the number of times a product is saved in the swap store Note SwapUtil Class At the moment the pipeline does not clean the swap store after the processing To avoid the swap store completely filling the hard disk when many 45 Memory Issues observations are processed it is suggested one manually remove the swap store by either deleting the swapStore directory or in HIPE from herschel hifi pipeline product import SwapUtil SwapUtil delete 46
30. d from the HK packets if possible noLoCodeOffset 12 2048 No valid LO code off set information Set when the flagbit is ze Flags in HIFI data Flag Name Bit Value Description ro in the DFs extracted from the HK packets if possible noLoCodeMain 4096 No valid LO code main information Set when the flagbit is zero in the DFs extracted from the HK packets if possible BbidCorrection MixerCurrentDeviation 14 8192 16384 Correction of Bbid see SPR 1963 Not relevant any more It was during SOVT testing but the onboard software has been corrected since Difference in mixer currents exceeds tol erance when applying DoRefSubtract MixerCurrentDeviation 32768 Difference in mixer currents exceeds tol erance when applying DoOffSubtract MixerCurrentDeviation 65536 Difference in mixer currents exceeds tol erance when applying DoFluxHotCold or Mk FluxHotCold NoHotColdCalibration SuspectLO 19 131072 262144 Division by the band pass has not been car ried through LO Frequency is listed in the Bad Frequency Table Data not neces sarily is corrupted SpurDetected 524288 Spur detected in the cold load Data partial or total is corrupted IgnoreData 21 1048576 User has the option to set this flag Some tools e g doDeconvolution will honor it Flags in HIFI data
31. d spectra 10 3 1 Using Gridding Task Another task is available called Gridding to make cubes of images It works with any dataset or product that happens to implement the SpectrumContainer or SpectrumContainerBox interfaces It can also work with a collection of SpectrumContainer s Said without using the Java jargon means that it can accept various simple inputs such as an Spectrum2d or an Spectrumld since these are SpectrumContainers You may also create your own collection of datasets and pass it to the Gridding task in order to provide the spectra to be read to make a cube by performing an spatial regridding a convolution of these spectra onto a regular grid computed based on the coordinates of the given spectra and on optional inputs about the shape of the grid which can be given by the end users The Gridding task and the Spectrum Toolbox The user can make use of the spectrum selection tools of the spectrum toolbox to perform any selection of spectra followed by the usage of the Gridding task to create a cube for each segment of the spectra in these selections The following example shows how to combine SelectSpectrum with the Gridding task first create an instance of the SelectSpectrum task 38 How to make a spectral cube selector herschel hifi pipeline util tools SelectSpectrum use SelectSpectrum to get a single HifiSpectrumDataset with the spectra that mup Cerca Inm eriteria e g here one selects tho
32. data up to level 2 Data taken from the Herschel Science Archive HSA can be re processed from level 0 option 0 to levels 0 5 option 0 5 1 option 1 or 2 option 2 If you try to re process from a higher Level data than exists in the Observation Context then the hifiPipelineTask will automatically select the highest existing Level For example if you try to re process from Level 0 5 to 1 but the ObservationContext only contains a Level 0 product then the pipeline will automatically run from Level 0 to Level 1 Running the HIFI pipeline 2 2 2 4 You can supply your own algorithm to the pipeline see Section 2 4 Click on the folder to browse for the file or supply the full path in the text box The ways you might want to modify the pipeline algorithms are discussed in Section 2 3 See the notes below about customizing pipeline algoriths n the Output section choose the name of the observation context that will be produced or use the HIPE default obs e Click on accept to run the pipeline The status running if all is well error messages if not and the progress of the pipeline are given in the Info section at the bottom of the Task dialogue You will also see more informative messages about the status of the pipeline written in the console and terminal Saving the output There are several methods you can use to save your reprocessed observation Right click on the output ObservationContext obs and select Send t
33. e Check Hot load tem perature is Out Of Lim it false 1 5 V 0 5 V IF power level ok gt maybe unstable base line Level dropped gt transistor faulty the heater or with the readout Inform engineering team FPU_HOT_LOAD hotl 94931 JU RU serious problem with Inform engineering team Check Cold load tem perature is Out Of Lim it false SFPU COLD LOAD cpidP0aqQ FPU serious problem Inform engineering team Check Level 0 Tem perature is Out Of Lim it false FPU_LEVEL_TEMP l llemf S KRU serious problem with the thermal environ ment or with the read out Inform engineering team Level 0 5 Quality Flags WBS Quality Flags COMBFLAG QWbsFreq VALIDATE Flag for all COMB of the observation false ZEROFLAG QWbsZero VALIDATE Flag for all Zero of the observation false SPIKENUMBER QWbsSpikes NUMBER Maximum number of spikes detected in a Comb 0 gle spectrum 01 SATURATEDNUMBER pixelSaturated Maximum number of saturated pixel detected in a sin SDARKFLAGY darkFlag Spectrum contains saturated dark false Quality Flags Quality Flags BADPIXELS badPixels Number of channels marked as BAD due repeated saturations O1 Level 0 5 Quality Flags HRS Quality Flags NOQDC noQDC No Quantization Distortion Correction could be processed false FASTQDC
34. e edere erret ver tese o Ue ert voe TU 29 9 1 Introduction to doDeconvolution cece cece cece cece cece cece nme 29 9 2 Running the Deconvolution Tool cece cece nec cece cece HH emen 30 9 3 Viewing Deconvolution Results ssssssse Hem 31 10 How to make a spectral cube 20 00 cee een ccc ee HH HI mH ehem 33 10 1 Introduction to doGridding ssis siirre iresi rres cence eee ce teen ceca cena eeae eeu EE SETAE STOREN 33 10 2 Using the GUI to make a Spectral Cube sses A 33 10 3 Making a Spectral Cube via the command line see 34 10 3 1 Using Gridding Task sese 38 11 Exporting HIFI data to CLASS eed reete EES PESES Err SEEE TESSA ERTE 41 11 1 Introduction to hiClass uirinn eee eee HI emer 41 11 2 hiClass examples cte t te soi Te teet bete e rea aay ab bot ESS 41 11 3 How to read HIFI data in CLASS ssssssee HH 43 12 Memory ME HEURE 45 lil Chapter 1 Data Primer A short introduction to the structure of Herschel HIFI data storage Last updated 9 Oct 2009 1 1 Data frames The Herschel spacecraft stores data onboard up two days worth until transmited to Earth Science data such as a WBS spectrometer readout come naturally in sets or Frames Data frames are packe tized for transmission from HSO to Earth Along with House Keeping HK data they are downlinked to the tracking station and thence to the Mi
35. fhf FitHifiFringe fhf sdsl sds in nfringes 2 typical period 150 sds out fhf result sds For HifiTimelineProducts and SpectrumDatasets the output is identical to the input but with the fitted sine waves subtracted from the flux columns For ObservationContexts the input has the sine waves subtracted as well this is a HIPE limitation which may be changed in the near future Besides obs1 htp1 and sds1 the following input parameters are allowed 27 HIFI Standing Wave Removal Tool product if the input is an ObservationContext indicate which level 2 product needs to be processed Options are WBS H USB WBS H LSB WBS V USB WBS V LSB HRS H USB HRS H LSB HRS V USB HRS V LSB nfringes number of sine waves to be fitted DEFAULT 1 start period shortest period standing wave in MHz to search for DEFAULT start period 20 end period longest period standing wave in MHz to search for DEFAULT end period 3000 typical period typical standing wave period in MHz in the data This is used for the baseline determination Features with much longer periods are considered baseline structure and will not be removed with sine waves DEFAULT typical period 150 plot False only show plot of end result for each scan DEFAULT 2 plots per scan 1 period versus Chi 2 2 the before after plot and the subtracted sine wave and the line mask averscan True determine standing waves on average of all scans and then
36. fy that for example a given channel sample is saturated if those channels are saturated repeatedly during the observation then the quality flag SATURATEDNUMBER wil be raised Below is a list of the current available types of quality flags for the HIFI pipeline for each level The format below gives flag name flag description and flag default value Level O Quality Flags Quality Flags UNALIGNED HK unalignedHK data Percentage of Dataframes which have unaligned HK 0 0 NOCHOPPER noChopperHK data Percentage of DFs having no chopper information 0 0 NOCOMCHOP noCommandedChopperHK data Percentage of DFs having no commanded chopper information 0 0 NOFREQMON noFrequencyMonitorHKdata Percentage of DFs having no frequency monitor information 0 0 NOLCOFFS noLoCodeOffsetHKdata Percentage of DFs having no LO Code offset informa tion 0 0 NOLCMAIN noLoCodeMainHKdata Percentage of DFs having no LO Code main informa tion 0 0 BBID CORRECTION bbidCorrection Percentage of Bbids corrected according to commanded Bbids 0 0 DATAFRAMES OUTOFORDER dataframesOutOfOrder Unordered or duplicate Dataframes found false MISSING DATA missingData Less data found than expected false SURPLUS DATA surplusData More data found than expected false Quality Flags with specified thresholds Range Consequences for sci ence data Action required
37. getGraphs 0 g2 p getGraphs 2 display as line graph or histogram gO0 layer style chartType HISTOGRAM 18 Viewing Spectra g2 layer style chartType LINECHART add annotations g0 layer addAnnotation Annotation 4000 1 My annotation i g0 layer addAnnotation Annotation 5000 0 98 My annotation select a range of data g0 layer xaxis addMarker AxisMarker 4200 4400 g2 layer xaxis addMarker AxisMarker 6000 6500 These last lines will produce the following plot My annotation Observation 268510098 backend WBS H activeband 1b lofre quency 627 9998 sds_type chopse 0 996 0 994 0 992 0 990 0 988 flux 0 986 0 984 0 982 0 980 y annotation 0 978 3500 4000 4500 5000 5500 6000 6500 7000 7500 8000 8500 Frequency scale MHz Figure 5 1 5 4 The SpectrumExplorer Package 5 4 1 Starting the SpectrumExplorer The SpectrumExplorer package allows one to visualize HIFI PACS and SPIRE SpectrumDatasets in a userfriendly interactive way To activate it click on a SpectrumDataset or Product in the Variables window or Observation Viewer with the right mouse button and select Open With and Spectrum Explorer If this is the default it suffices to double click on the variable 19 Viewing Spectra Initially an empty plot is displayed in the top part of the window that is opened and a selection panel is displayed in the bottom p
38. he pool to which the pipeline should write output name My pipeline out pool ProductStorage LocalStoreFactory getStore name MyNewobs hifiPipeline obs Myobs palStore pool If the pipeline is not behaving as you expect keeping old values for example try resetting it hifiPipeline hifiPipelineTask Running the HIFI pipeline The exact ordering of the arguments does not matter What is an apid Application Program IDentifier it is what the pipeline calls spectrometers Note that to implement your own algorithm you must load the algorithm script from wherever you saved it into HIPE and compile it run it with gt gt before you run the pipeline see Section 2 4 To save MyNewObs to pool Storage ProductStorage pool PoolManager getPool MyPool storage register pool Storage save MyNewObs 2 3 Running the Pipeline step by step Running the pipeline or one part of the pipeline step by step allows you to inspect the results of each step and change the default parameters of the pipeline If you wish to create your own algorithm which must be written in jython for a part of the pipeline then this will likely be your first step tis not expected that there will be much need to customise the spectrometer pipelines up to Level 0 5 and indeed there are only a few steps of the spectrometer pipelines that have some options It is more likely that you may wish to play with how off and refere
39. he x and y axis thus specifying the dimensions of an elliptical beam When this input is not provided the gridding task computes a default value for the HIFI beam size based on a known function of the observed frequency At present the formula used for the default case is HPBW 75 44726 wavelength mm specify the size of the beam cubes doGridding htp htp beam Doubleld 15 4 specify the size of the beam In this case the beam is wider along the vertical axis cubes griddingTask htp htp beam Doubleld 10 20 If the beam size is specified and the pixel size is not specified the pixel size will be function of the beam size taking into account the Nyquist criterion and the smooth factor if any given Usually for nyquist sampling the default pixel size becomes half the beam size Specify the type of filter 35 How to make a spectral cube By default the convolution is performed with a gaussian filter function however the user can specify other filter types cubes doGridding htp htp filterType box At present the available filter functions are box function best for Raster maps and a Gaussian function best for OTF Other filter functions maybe added in next releases the default filter type is gaussian cubes doGridding htp htp filterType gaussian Specify the parameters of the filter along each axis The parameters that characterize each filter can be modified For example to
40. hem in and what options you can change see the HIFI Pipeline Specification document see 2 4 How to customise pipeline algorithms The pipeline algorithm scripts can be found in WBS BuildDir scripts hifi pipeline wbs WbsPipelineAlgo py HRS BuildDir scripts hifi pipeline hrs HrsPipelineAlgo py e Level 1 BuildDir scripts hifi pipeline generic LevellPipelineAlgo py Running the HIFI pipeline e Level 2 BuildDir scripts hifi pipeline generic Level2PipelineAlgo py Open the algorithim you wish to customise in the editor edit it and save Compile your algorithm by running the script with gt gt Apply the algorithm to the pipeline as described in the sections above Chapter 3 Flags in HIFI data Last updated 11 Feb 2010 3 1 Introduction to flags Flags also called masks are identifiers of specific issues with the data such as saturated pixels or a possible spur that can affect the quality of the final product Flags are used to identify affected data and to make a caution during its processing A Flag has a defined name and a value which specifies the nature of the flag The flags are divided into two categories depending on whether they apply to an individual channel pixel or to a complete Dataframe They are called channel f lags and column rowflags respectively Note There are also Quality Flags which are found in the Quality Product in the Observation Contaxt and are used
41. hifiPipeline task links together the four stages of the pipeline described above and it can be used to reprocess ObservationContexts up to any Level for any choice of spectrometer s and polarisation s You can also make your own algorithms or modify the ones provided in the scripts hi fi Pipeline directory in the installation directory of HIPE and apply them to the pipeline Configuring the pipeline The first step in reprocessing an observation is to configure one of the properties of the pipeline In the future a means to automatically configure the pipeline for your needs will be provided but for now save the following line in a py file and run that script once in your session before running the pipeline Running the HIFI pipeline Configuration setProperty hcss ia pal store spgstore pipelineout Alternatively you can eliminate the need to run a script by setting this property in your hc ss user props file hess ia pal store spgstore pipelineout This property sets the pool to which the pipeline will by default write output You will see below Saving the output how to save the output of the pipeline to a pool of your choice an it is recommended that you follow that method Why This pool is not overwritten but appended to so you would need to set it everytime you ran the pipeline even if you made an error decided you wanted to try a different parameter or the pipeline failed this rapidly becomes tiresome
42. his differs in 1 from the usual convention for FITS images where the center of the bottom most left most pixel has coordinates 1 0 1 0 If the user specifies only this refPixel input and the user does not specify the coordinates of that pixel this will computed so that it gets the pixel coordinates of the centre of the input spectra For example if we want to force that the reference pixel is the pixel 3 5 4 0 then the refPixel input will be Doubleld 3 5 4 If no refPixelCoordinates are provided then the centre of the coordinates of the input spectra will be locate at the pixel 3 5 4 0 of the regular grid i e at the FITS pixel 4 5 5 0 from the bottom most left most pixel of the cube This means that the value of the CRPXII parameter of the result cube will be equal to 4 5 and the value of the CRPIX2 parameter will be equal to 5 0 remind that the cube header uses the usual FITS convention about pixel coordinates 37 How to make a spectral cube cubes doGridding htp htp refPixel Doubleld 3 5 4 0 By setting both refPixel and the refPixelCoordinates input explained below the user can place the regular grid at any arbitray location although the user is adviced to let the task automatically compute these so that the grid is located at a suitable place fully covering the observed spectra Specify the coordinates of the reference pixel In addition to choosing a reference pixel the user can also specif
43. ibrations of baseline sideband gain and antenna temperature It is also important that the spectra have been resampled to a linear frequency axis doFreqGrid in the Level 2 pipeline The default operation of the task is to select the science datasets from an HTP and create a cube for each given spectrometer subband Each slice of the cube is produced by computing a two dimensional grid covering the area of the sky observed in a mapping mode For each pixel in the grid the task computes a normalized Gaussian convolution of those spectra equally weighted falling in the convolution kernel around that pixel After running the task you will have an array of cubes one for each subband and in 3 0 a cubesContext variable that allows you to easily browse the cubes without need to extract them from the cube array The SimpleCube product can be viewed and analyzed in the SpectrumExplorer see and with the CubeSpectrumAnalysisToolbox see Using the GUI to make a Spectral Cube The doGridding Task can be found in the Applicable folder of the Tasks view when an HTP is selected in the variable view double click on it to open the dialogue in the Editor View You can also find the task under the Task View in By Category gt HIFT As a part of the automated SPG pipeline doGridding handles ObservationContexts but if you are making a cube yourself then you should use a Level 2 HifiTimelineProduct HTP The reason for this is that doGr
44. idding assumes that the spectra have a linear frequency axis and this may not be the case for Level 0 5 or Level 1 HTP where there can still be overlap of subbands Resampling to a linear frequency axis is carried out in the doFreqGrid step of the Level 2 pipeline Using the GUI you can pass an HTP to the task You can also specify the subbands for which to create cubes useful if you know a line falls only in one subband the beam size the weights to be used the type of convolution filter and the parameters of the filter By hovering the mouse over the parameter names in the GUI you can find more information and some tips on usage There are two drop down menus in the GUI one to select the type of weighting either all spectra equally weighted or you can read the weights column from the dataset which will carry forward any weightings you have already applied to the data and one to select either a Gaussian or a box filter for the convolution For all the other parameters you must specify a variable in the command line and drag that variable to the appropriate bullet to modify the defaults of the task Here are some examples subbands by default cubes are created for all subbands in the HTP To specify for example sub bands 2 and 3 create the variable subbands subbands Intld 2 3 and drag it to the subbands bullet 33 How to make a spectral cube beam the default half power beam width beam size is calculated
45. ies cft ConvertFrequencyTask cft sds spectrum to 2 lsbfrequency Of course it is also possible to convert to the upper sideband for this the keyword is usbfrequency cft sds spectrum to usbfrequency To convert back to the IF use cft sds spectrum to frequency 24 Changing to LSB USB and Velocity The ConvertFrequencyTask works equally well on the HifiTimelineProduct itself In this case all the internal HifiSpectrumDatasets are converted This is not something you should do in the early stages before level 0 5 of the HIFI pipeline For example on a level 1 HifiTimelineProduct cft ConvertFrequencyTask cft htp hifitimelineproduct to frequency Note Direct application of the ConvertFrequencyTask changes the data listed in the spectrum Conversion back to the original IF scale is possible just use the to frequency option 6 1 2 2 Velocity The ConvertFreqencyTask also works to convert the frequency scale to a velocity scale once given the reference frequency cft ConvertFrequencyTask cft sds spectrum to velocity reference 576 268 inupper False In the above example I had to specify the reference frequency in GHz and whether this reference frequency is for the upper inupper True or lower inupper False sideband Another call to ConvertFrequencyTask using to frequency will undo the change to velocity as well 6 1 2 3 Review of ConvertFrequencyTask The ConvertFrequencyTask
46. ineProduct will benefit from the use of swap space The following properties can be modified in the user props file or using the Hifi Product tab in propgen to set or to configure the Swap mechanism hess hifi pipeline product memory true Setting the value of this property to true en ables the swap mechanism Note that the default value is false hcss hifi pipeline product swapstore swapStore This is the name of the LocalStore where the temporary data will be saved The default location is user home hcss lstore swapStore hcss hifi pipeline product swapratio 0 25 This property determines how much the swap mechanism is used and is used to set the threshold level of free memory When a new dataset is set or retrieved from the HifiProduct the HifiProduct will check the size of the dataset and the free memory in the system If the condition memory free swapratio dataset size is met then all the floating datasets contained in the HifiProduct will be saved in the swap store This property should have value between 0 and 1 and has a default value of 0 25 If the value is O all datasets will be always stored in the swap store This is safe but it could create performance delay in the time needed to process the pipeline due to the access time to the hard disk In the case of long observations setting the property to 1 could be dangerous because memory problems like Java heap space exception may still oc
47. ing in the console print herschel hifi dp tools hiclass tools doc 11 2 hiClass examples 1 Export one dataset to a FITS file HiClassTask dataset myspectra fileName myspectra fits 2 Export one HIFI timeline product to a FITS file HiClassTask product myhtp fileName myhtp fits The fits file is written in the installation directory of HIPE 41 Exporting HIFI data to CLASS You should specify one of but never both dataset or product and an output filename The remaining properties of hiClass and their defaults are as follows HiClassTask product myhtp fileName myhtp fits exportFrequency False doublePrecision True blankingValue 1000 raNominal 38 27531 decNominal 76 949043 veloSource NaN spec sysz exportFrequency False When set to True the resulting FITS file will contain columns with the value of the intermediate or sky frequency for each channel which CLASS does not handle properly It is recommended that you leave this to false unless you use this task to export your spectra to FITS in order to read them with IDL and if you want to export the irregular WBS frequency axis along with it doublePrecision True Recommended to leave to true particularly for spectral scans blankingValue 1000 NaNs in the flux columns will be replaced by this value This is the only way to blank channels in CLASS as CLASS does not handle flags or NaNs raNominal NaN de
48. leted false Hot cold calibration Quality Flags HOTCOLDDATA hotcoldData Data measured from hot and cold loads not sufficient for hot cold calibration false TS YSFLAG tsysFlag Hot cold calibration not successful false Quality Flags Quality Flags INTENSITYCALIBRATION intensityCalibration Intensity calibration not or not for all spec tra carried through false Channel weights Quality Flags CHANNELWEIGHTSFLAGY channelWeights Problem occurred while computing channel de pendent weights No weights added false Reference subtraction Quality Flags REFSUBTRACTIONFLAGY refSubtraction Reference subtraction not processed maybe iden tification of phases not successful false Off smooth Quality Flags NOOFFBASELINE noBaseline No off baseline could be calculated false Off subtraction Quality Flags ONOFFSEQUENCE onoffSequence ON OFF datasets not in expected sequence ON OFF ON OFF or ON OFF OFF ON ON false ONOFFPAIRSIZE onoffLength Some ON OFF dataset pairs found with unequal number of rows false ONOFFPROCESSING onoffProcessing More ON than OFF datasets found in the data not all ON datasets could be processed with OFF dataset s false OFFBASELINESUBTRACTION offBaselineSubtraction No off baseline subtraction carried through
49. nce spectra are subtracted in the Level 1 pipeline although it is expected that the default settings should work well To step through the pipeline you must work directly on the appropriate level HifiTimeLine HTP the dataset containing all the spectra including calibration spectra made during an observation for a given spectrometer So the first thing you must do is extract the HTP you want to work on from your ObservationContext Drag an HTP from the ObservationContext tree in either the Context Viewer or Observation Viewer into the Variables view and rename it if you desire by right clicking on the new variable and selecting rename n the command line the formalism to extract an HTP is htp obs refs level2 product refs HRS V USB product level2 and HRS V USB should be replaced by the level and backend combination desired When you select an HTP in the Variables view in HIPE you will notice that many tasks with names like DoWbsDark mkFreqGrid These are the names of all of the steps in the HIFI pipeline mk signifies a step where a calibration product is being made Do is a step where a calibration is applied You can step through the pipeline using these tasks or more efficiently use and modify the scripts that are supplied with the software in the scripts hifi Pipeline directory in the installation directory of HIPE Forinformation on the steps of each level of the pipeline their names the order to run t
50. nd row wise channelRanges append Intid 400 700 0 cubes doGridding htp htp subband Intld 1 4 channels channelRanges Select datasets by type the default action is to take the science data sets that are on the source and this is normally sufficient However there may be observations where the dataset type to be read to make the cube has a different dataset type e g an engineering observations whose type is called other instead of science You can also select the off positions too cubes doGridding htp htp datasetType science ignoreOffs false e Select some datasets by index instead of picking all the science datasets datasetT ype is ignored if this is used here we select subbbands 2 and 4 and datasets 3 4 and 5 from the HTP The weighting can also be specified to be equal this is default or that computed in DoChannelWeights in the Level 1 pipeline selection cubes doGridding htp htp subbands Intld 2 4 datasetIndices 3 4 5 weightMode selection cubes doGridding htp htp subbands Intld 2 4 dataset indices Inti1d 3 4 5 weightMode selection cube subband 2 cubes 0 cube subband 4 cubes 1 Geometry Specify the antenna beam size You can specify which is the half power beam width of the instrument i e the beam width In the case of HIFI case the beam is symmetric hence a single value is needed However one might in principle provide two different sizes along t
51. ntainer and con tainerBox 40 Chapter 11 Exporting HIFI data to CLASS Last updated 1 March 2010 11 1 Introduction to hiClass It is possible to export all Herschel data to FITS files using FitsArchive but these are not readable by CLASS Therefore hiClass has been developed to export HIFI spectra to a FITS file that CLASS can read Please note that this task is for HIFI data only it cannot be used for PACS or SPIRE data The hiClass task can be used for SpectrumDatasets or HTP of level 0 5 1 and 2 data but not raw data level 0 The following information from is exported to CLASS The fluxes of course The frequencies in a column if you ask for it see examples ObsId BbType BbId SequenceNumber The way CLASS will store and handle it is still being discussed The name of the observed source which is computed from the BBType The Rest Frequency Image Frequency Channel References Frequency Step HiClass always choses the centre of the spectrum as the reference Dates of observation and name of the instrument HIFI plus spectrometer and polarisation Pointing information Tsys The hiClass task is a wrapper around the HiClass object defined in herschel hifi dp tools hiclass tools py Only the usage of the hiClass task is described here if you want to work directly with the HiClass object you can read further documentation about how the HiClass object works including examples by typ
52. o Local store When you run the pipeline you can specify which pool the output should be written to In the console type name My pipeline out pool ProductStorage LocalStoreFactory getStore name and drag poo1 to the palStore bullet in the GUI The Expert mode of the hifiPipeline is intended for Calibration Scientists and Engineers and is not described here The hifiPipeline in the command line Below are some examples of running the hifiPipeline task from the command line once again it is assumed that an ObservationContext called Myobs has been loaded into the session Reprocess an ObservationContext up to Level 2 for all spectrometers MyNewobs hifiPipeline obs Myobs Reprocess Myobs from Level 0 5 to Level 1 for all spectrometers MyNewobs hifiPipeline obs Myobs FromLevel 0 5 UpToLevel 1 Now reprocess MyNewobs which now contains data only up to Level 1 but only for the WBS WBS H and WBS V are the horizontal and vertical polarizations respectively MyEvenNewerobs hifiPipeline obs MyNewobs apids WBS H WBS V Reprocess Myobs from Level 0 to Level 0 5 for only horizontal polarization data MyNewobs hifiPipeline obs Myobs apids WBS H WBS V FromLevel 0 UpToLevel 0 5 Now include your own algorithm for the Level 1 pipeline for all spectrometers from Level 0 to 1 MyNewobs hifiPipeline obs Myobs FromLevel 0 UpToLevel 1 levellAlgo full pathjmylevellAlgo py Specify t
53. ool to remove standing waves from level 1 and level 2 HIFI spectra It makes use of the general sine wave fitting task FitFringe but has been adapted to read HIFI data and provide input and defaults applicable to HIFI spectra For details on the sine wave fitting method please consult the FitFringe manual FitHifiFringe is being tested on PV data It can be applied to all bands with the caveat that the standing waves in HEB bands 6 and 7 are not sine waves and hence can only be fitted in an approximate way by fitting a combination of many sine waves The GUI and input parameter names of FitHifiFringe are different in HIPE versions 2 and 3 The latter is more userfriendly although there is no difference in the functionality 8 2 Running FitHifiFringe FitHifiFringe is automatically registered to an ObservationContext i e when clicking on an ObsCon text in the Variable window of HIPE FitHifiFringe shows up as an applicable task However the user can also process a HifiTimelineProduct or a SpectrumDataset instead by opening the GUI under All Tasks and then dragging the variable to the appropriate bullet Alternatively run it on the command line as follows fhf FitHifiFringe fhf obsl zobs in nfringes 2 typical period 150 product WBS H USB obs_out fhf result_obs or for HifiTimelineProducts fhf FitHifiFringe fhf htpl htp in nfringes 2 typical period 150 htp out fhf result htp or for SpectrumDatasets
54. rid The wave and flux ranges above the plot can be altered by textual input or by scrolling on top of the text field After doing this the slide bars below the ranges can be used to slide the sub range through the plots Use the scroll wheel on top of the plot to zoom A single click on a plot opens the spectrum in the plot view of the SpectrumExplorer Preferences Default SpectrumExplorer settings can be modified using the Edit gt Preferences button at the very top of the HIPE screen The following options are available nitial tool specifies whether the Spectrum Explorer should start in zoom or select mode e ChartType display plot in line style or histogram Display grid on or off Display legend on or off Start in preview mode on or off For a specific SpectrumDataset type title subtitle and legend element can be specified Metadata fields and attribute fields can be filled in automatically by specifying the fields name between angular brack ets Optionally with a printf style format suffix For example longitude 2f in the legend element field displays the value of the longitude attribute for each spectrum in the legend 23 Chapter 6 Changing to LSB USB and Velocity 6 1 Changing HIFI Frequency Scales In practice there at two methods of altering the HIFI frequency scales using the Spectrum Explorer GUI or from the command line These two approaches differ in one fundamental way The command line t
55. ropy method The relative importance of maximizing the gain entropy of the solution along with matching the observed spectra is controlled by this weighting coefficient cont offset Used in maximum entropy method The user can insert a continuum offset value to insure that no negative fluxes enter and disrupt the entropy calculation The offset is subtracted after the solution is reached expert Toggle on an off expert use of the tool which allows viewing of interim products These are a snap shot of the solution as a function of iteration and include goodness of fit measurements such as Chi squared Note this is memory intensive and only available in 3 0 Maximum Extropy The maximum entry option is turned on as a stop gap measure to help a bad situation with the input data The bad situation can include 1 Insufficient redundancy say a redundancy of less than R 4 2 Too few lines since line strengths guide the deconvolution 3 Poor or excessively noisy data 4 Also if the solution of the nominal deconvolution contains periodic noise patterns or the solved gains deviate widely from 1 0 The inclusion of the maximum entropy method adds a term to to the quantity being min imized Without this term the quantity being minimized is the Chi square difference be tween observed double sideband DBS spectra and the modelled DBS spectra The mini mization is accomplished by altering the SSB model spectrum from which the
56. rument and observatory independent for HIFI essentially an intensity calibration It is expected that Level 1 data processing can be performed without human intervention 5 Level2 scientific analysis can be performed These data products are at a publishable quality level and should be suitable for Virtual Observatory access 6 Level 3 These are the publishable science products with level 2 data products as input Possibly combined with theoretical models other observations laboratory data catalogues etc Formats should be Virtual Observatory compatible and these data products should be suitable for Virtual Observatory access 1 3 Contexts A Context is a subclass of Product a structure containing references to Products and necessary meta data A Context can contain Contexts giving rise to Context trees Types 1 ListContexts for grouping products into sequences or lists hardly used 2 MapContexts for grouping products into key value dictionaries Data Primer 1 3 1 Herschel Observation Context A MapContext instance serves as the organisational product unit for the Herschel Data Processing system It contains the following contexts 1 2 Level 0 Level 0 5 Level 1 Level 2 amp Level 3 optional Contexts Calibration Context Auxiliary Context Quality Context Browse product Trend Analysis Context optional Telemetry Context not by default only when the HSC deems it necessary beca
57. s doDeconvolution to stop after a specified number of interations if it has not converged by then tolerance Specifies the tolerance of the solution When the rms of the residual of the fit changes fractionally by less than tolerance the algorithm stops iterating A value of 0 001 is best Below this value the algorithm may produce poor baselines gain Toggles gain optimisation on and off If on doDeconvolution will run twice first with gain factors set to 1 0 for stability and then a second time starting with the SSB solution of the first run but this time allowing the gain values to be optimized as well channel weighting Toggles whether or not the deconvolution uses the weight values in the data to weight less noisy data ignore mask Looks for a row mask identifying spurs so strong they corrupt the entire spectrm and will ignore that spectrum during deconvolution The defaults are still being determined plot dsb Toggles visualisation on and off When on the SSB output solution against the DSB input can be viewed use entropy The user can turn On or Off terms which incorporate the maximum entropy method in the deconvolution e l amdbal channels Used in maximum entropy method The relative importance of maximizing the SSB channel entropy of the solution along with matching the observed spectra is controlled by this weighting coefficient 30 Sideband Deconvolution lambda2_gains Used in maximum ent
58. se spectra where its containing dataset has bbtype equal to 6022 selected selector htp htp selection_lookup bbtype 6022 return_single_ds Boolean TRUE one might have a glance at the spectra in the selection dataset e g in the TablePlotter cube gridding selected cubes gridding cubes Making a SpectralSimpleCube with the Gridding task a UP E E E E make a dataset with all the spectra from all the science datasets isLine true gt bbtype 6022 pe es oo oi Bes Sa ee Sao a Sa SSS SSS SSeS ees ase SSeS Sees eas esas Se asas selector herschel hifi pipeline util tools SelectSpectrum selected selector htp htp selection lookup bbtype 6022 return_single_ds Boolean TRUE scienceOnIndices htp summary isLine data where htp summary isLine data Boolean TRUE bbid htp summary Bbid data scienceOnIndices another way of selecting selected selector htp htp selection_lookup bbtype bbid 0 return_single_ds Boolean TRUE the Gridding task that can work with any SpectrumContainer or collection of SpectrumContainers like the selected above cube gridding container selected get a point spectrum from this selection ds spectrum selected getPointSpectrum 1 ds_segment ds_spectrum getSegment 3 read its third subband get SpectralSegment plotSegment PlotXyY ds_segment wave ds_segment flux xtitle Frequency MHz ytitle Intensity now
59. select ranges Click and drag to select ranges in a plot the middle mouse button can be used anytime for this as well This will create a vertical grey bar Then in the spectrum selection mode button 4 only this will be saved as a new variable button 7 select points Click and or drag with the left mouse button to select one or more spectral points These points can later be flagged or removed button 8 de activate preview mode In preview mode a quick preview is displayed of all rows selected in the selection panel button 9 display hide grid in the active sub plot button 10 display hide the plot legend button 11 switch between line and histogram mode button 12 display flagged channels button 13 show hide the plot title button 14 open filter panel button 15 show metadata of the displayed SpectrumDataset button 16 open a raster panel showing all plots in the selection panel button 17 open the properties panel in the top right part of the SpectrumExplorer to view and modify any plot parameter The panel can also be opened using the Properties option under the right click popup menu If a paricular element in the context contains no changeable properties the plot properties are displayed Plot Interactions The Spectrum Explorer provides context dependent plot interactions The behaviour of mouse inter action depends on the location of the mouse cursor The actual context is displayed in the lef
60. sets of frequency and intensity calibrated spectra 4 The Level 2 pipeline removes further instrumental effects by converting to antenna temperature applying side band gain corrections and converting velocities to the local standard of rest frame Spectra are averaged folded or gridded into spectral cubes as appropriate In theory Level 2 products can immediately be used for scientific analysis but this is not recommend ed At the minimum you will need to remove baselines standing waves see Chapter 8 spurs and other outliers in the case of spectral scans you will need to deconvolve the spectra see Chapter 9 You may also wish to change the temperature scale or reference frame see Section 6 1 Particularly in the early stages of the mission data may well need to be looked at much more carefully before scientific analysis can be done Indeed you may wish to re run all or part of the pipeline to change defaults use your own pipeline algorithm or examine each step of processing To that end the ObservationContext that is obtained from the HSA contains along with the Level 0 2 data Prod ucts everything you need to reprocess your observations calibration products satellite data as well quality logging and history products which you can use to identify any problems with your data or its processing The following sections explain how to re run the pipeline using the HifiPipeline task 2 2 How to run the HIFI Pipeline The
61. ssion Operation Center MOC at ESOC in Darmstadt or to the latter directly The data packets then flow from the MOC to the Herschel Science Center HSC at ESA s European Space Astronomy Centre ESAC in Madrid The HIFI ICC copies the data from HSC as well At ESAC the data packets are ingested into a database and the science data frames are reconstituted The combination of HK and science data creates a Level 0 Observational Data Product 1 2 Data Products refs Herschel Data Product Document partI v0 95 pdf ftp ftp rssd esa int pub HER SCHEL csdt releases doc ia pal doc guide html pal guide html A Herschel Data Product consists of metadata keywords tables with the actual data and the history of the processing that generated the product There are various product types Observation Calibration Auxiliary Quality Control User Generated The types of Observation Data Products 1 Level 1 Raw data packets separate HK and science frames as described above 2 Level 0 HK and science frames grouped by time and building block ID and perhaps other param eters As close to raw data as the as the typical user would find useful to be 3 Level 0 5 data processed to an intermediate point adequate for inspection for HIFI they are pro cessed such that backend spectrometer effects are removed essentially a frequency calibration 4 Level 1 Detector readouts calibrated and converted to physical units in principle inst
62. t bottom corner of the plot panel Next to the context you ll find the location of the mouse cursor in plot coor dinates The following table provides the some contexts and the mouse interaction behaviour Context Click Ctrl click Drag Scroll Subplot Set as active Zoom Select Pan Zoom Axis Pan Zoom Spectrum Select spectrum Extend selection Move spectrum to another subplot Select point Extract spectrum to a new variable Use spectrum as task input parame ter Selection Same as above 22 Viewing Spectra 5 4 6 5 4 7 Context Click Ctrl click Drag Scroll Marker edge Resize marker A right click on a plot shows a popup menu with global and context specific options Right clicking below or besides a plot gives the option to add another subplot in that place The new subplot becomes active New selected spectra are displayed in the active subplot To activate another subplot right click on that subplot and check the radio button named active Raster Panel When SpectrumExplorer is used in raster mode selected using the Raster button at the top button bar a single spectrum is plotted plot for each row in the selection panel This selection can be altered by making use of the filter panel When all spectra contain pointing information the plots are laid out on a latitude longitude plane Otherwise the plots are displayed in a rectangular g
63. ta 3 3 Column rowflags Column rowflags the rowflag column in the HIFI spectrum TableDataset apply to the complete Dataframes DF or rows in a HifiSpectrumDataset HSD For bit n the value is computed according to value 2 The first 5 bits are about the packets from which the DataFrame DF is reconstructed and are unlikely to ever occur Below is a table showing the current names and values of HIFI rowflags Flag Name Bit Value Description PacketOrder 1 1 Error in the packet or der while constructing the DataFrame PacketLength 2 2 Error in the packet length while construct ing the DataFrame TooMuchData 3 4 More data than can be fit in a DataFrame FirstPacket 4 8 Error in the start packet while constructing the DataFrame NoBlocks 5 16 No block information present while construct ing the DataFrame spare 6 32 spare 7 64 UnalignedHK 8 128 HK could not be aligned with DataFrames When the columns df_transfer and hk_transfer in the TableDataset are differ ent bit 8 is set noChopper 9 256 No valid Chopper in formation Set when the flagbit is zero in the DFs extracted from the HK packets if possible noComChop 10 512 No valid Commanded Chopper information Set when the flagbit is ero in the DFs extract ed from the HK packets if possible noFreqMon 11 1024 No valid Frequency Monitor information Set when the flagbit is zero in the DFs ex tracte
64. to provide you with means to make a quick assessment of the quality of your data they are discussed in chapter Chapter 4 3 2 Channel flags Channel or pixel flags apply to individual pixels and are added as a column in the HTP Their names are also added to the metadata of a dataset during processing and this is used for the history of the pipeline it also means that you can tell that e g the WBS pipeline has been applied if you see things like isMasked and checkZero in the metadata For each pixel there are 32 flags which can be set currently 8 are defined and the definition of the mask bits and values in HIFI data is given here Flag Name Value Description Bad pixel 0 If this bit is set the sample con tains a bad pixel Saturated pixel 1 If this bit is set the sample was saturated Not observed 2 If this bit is set the sample is not observed Not Calibrated 3 If this bit is set the sample is not calibrated In overlap region 4 If this bit is set the sample is in the subband overlap region I e it can be seen better in the adja cent subband Glitch detected 5 If this bit is set the sample is not observed Dark pixel 6 If this bit is set the sample is used to measure the dark Spur candidate 7 If this bit is set the sample is a candidate to be a spur It is a candidate since not all things flagged by the spurfinder are necessarily spurs Flags in HIFI da
65. ue empty string then HiClass will try to find in the datasets the reference frame in which the frequencies are expressed Setting this parameter to something else than will override whatever reference frame the datasets may refer to At present there are two choices topocentric the frequencies are expressed in the the satellite frame The satellite velocity cor rection was not applied e LSRK the frequencies are in the local standard of rest They have been corrected from the effect of the satellite velocity There is still no correction of the source velocity There is a hiClass GUI available accessable from the Tasks View under General gt HIFI If you use the GUI you cannot touch the exportFrequency or doublePrecision properties 42 Exporting HIFI data to CLASS AAA HIPE hiClass File Edit Run Window Help Hebel ern x Editor x 20 28 varia x P HVProfile rison py hifidemo ringe py flagspurDatasets py V PutAllSci alnDS py K obs hiClass x x Input 9 hiClassObj hiClassObj_1 e PX dataset e product MyHTP b obs blankingValue 1000 fileName MyClassFile poolname r 1 9 query raNominal amp 38 27531 decNominal 76 949043 e rac veloSource e 132 specsys LsRk 9 storage hiClassObj 9 lt No variable gt Output Variable name for hiClassObj hiClassObj Info
66. use a box filter with a different length parameters Doubleld 0 5 Doubleld 1 5 subbed c doGridding htp htp weightMode equal filterType box filterParams parameters The next example specifies the parameters length and sigma of the Gaussian filter function when using a gaussian filter default case the influence area is the area surrounding a grid point where the algorithm must pick up all the available data points influence_area 1 95 length in pixels sigma of the gaussian function times SQRT 2 sigma sqrt2 0 3 in pixels xFilterParameters Doubleld influence area sigma sqrt2 default case influence area 1 8 sigma sqrt2 0 36 yFilterParameters Doubleld influence area sigma sqrt2 cubes doGridding htp htp filterType gaussian xFilterParams xFilterParameters yFilterParams yFilterParameters it is also possible to pass both set of parameters in a single input parameters Doubleld 1 8 0 4 Doubleld 1 6 0 3 cubes doGridding htp htp weightMode equal filterType gaussian filterParams parameters The following example modifies the default parameters of the box filters their length customize a box filter i e set the length of the pixel measured in pixels parameters Doubleld 0 5 Doubleld 1 5 Ce qune doGridding htp htp weightMode equal filterType box filterParams parameters Note bear in mind that the default values of each type
67. use of a serious problem in the processing to level O data The uses of these Contexts will be described in Chapter 2 Note that the descriptive modifiers Product and Context are often dropped conversationally Chapter 2 Running the HIFI pipeline Last updated 1 March 2010 2 1 Introduction to the Pipeline HIFI data is automatically processed through the HIFI pipeline before it can be accessed from the the Herschel Science Archive HSA The HIFI pipeline is used for processing data received from one or more of the four HIFI spectrometers into calibrated spectra or spectral cubes and comprises four stages of processing 1 Take data from the satellite and minimally manipulate it into time ordered Data Frames a Hi fiTimeline or HTP for each spectrometer This is a Level 0 data Product which is the least pro cessed data available to Astronomers 2 Remove backend instrumental effects essentially a frequency calibration There are separate pipelines for the WBS and HRS spectrometers and the result is a Level 0 5 Product From HCSS 3 0 onward you will not see this Product in the ObservationContext unless the generation of a Level 1 product fails However you can always generate it for yourself 3 Application of observing mode specific calibrations i e subtraction of reference and off positions and intensity calibration using Hot Cold loads This is done by the Level 1 pipeline and resulting Level 1 Products are
68. y its celestial coordinates i e the longitude and latitude of the point chosen as the reference pixel of the cubes to be made by the gridding task For instance to make that the reference pixel is located at the coordinates RA DEC 308 9 40 36 degrees a refPixelCoordinates input can be provided with these coordinates Let s say in addition that the user wants that these reference pixel is the 0 0 pixel located at the bottom left comer of the image Then refPixel 0 0 Then the user should call doGridding like in the following example refPixel Doubleld 0 0 longitude 307 9 latitude 40 36 refPixelCoordinates Doubleld longitude latitude cubes doGridding htp htp refPixel Doubleld 0 0 refPixelCoordinates Doubleld longitude latitude door cubes doGridding htp htp refPixel refPixel refPixelCoordinates refPixelCoordinates one can check that cubes i wcs crvall longitude and cube i wcs crval2 latitude print cubes i wcs crvall longitude 1 True print cube i wcs crval2 latitude 1 True Please note that if only the refPixelCoordinates input is provided the user will be choosing the coordinates of the centre of the map By setting both refPixel and refPixelCoordinates the user can place the regular grid at any arbitrary location although the user is adviced to let the task automatically compute these so that the grid is located at a suitable place fully covering the observe

The HIFI User`s Manual - Herschel

Contents

Download Pdf Manuals

Related Search

Related Contents