What's coming up in HIPE 8

PDF Version Click here

See also the HIPE known issues and the Data products known issues pages for last-minute information on known problems.

Core system

Jython major upgrade. The version of Jython on which HIPE is based has been upgraded from 2.1 to 2.5.2. You may have to adapt your scripts for them to work with the new version.

ALERT! See this page for details on changes from Jython 2.1 to 2.5.2 that could affect your scripts.


Status bar

The maximum memory level reached is indicated with a vertical line in the memory bar. The tooltip shows the actual reached value, and allows to reset it so that further computations can record its own maximum.

Editors & Views

Variables View
  • Added contextual menu action Show contents to show the contents of a variable in the console.
  • Added contextual menu action Send To -> Text File to save the contents of TableDatasets in files.
  • Names of variables created from the Product Viewer are more consistent and meaningful.

Dataset Viewer
  • Numeric values in metadata, array datasets or table datasets can be displayed in decimal or hexadecimal format, and dates can be shown in either UTC or TAI. These options are available in the contextual menu for individual cells.

Product Viewer
  • The title (product's description) can be selected and copied.
  • The expanded sections can be resized, for example to show more metadata rows.
  • The summary of the Observation Viewer has been enriched with useful metadata of the observation being shown.
  • Products can be saved directly from the tree with the Sent To action of the popup menu, without needing to create an intermediate variable.

Jython & Text Editors
  • The supported Jython language has been upgraded from version 2.1 to 2.5.2.
  • Scripts and text files coming with the installation or with plug-ins are safe to edit without losing the original version.
  • Improved the step by step execution of scripts. See the note on this page for more details.

Help & Debug

  • The Tip of the day contains information on the Console meta commands (those starting with !)
  • If you get temporary (or permanent) freezes in the HIPE GUI, or you have created a custom task dialogue or view in HIPE, you can debug it by enabling the corresponding preference in Edit > Preferences > General > Debug. Any problem found with this debugging can be reported like previously, with the HSC Helpdesk or Give Feedback options of the Help menu.


  • More protection against problematic plug-ins. You can now disable all plug-ins without having to start HIPE. This can be convenient when you know or suspect that a plug-in causes HIPE not to start. To disable all plug-ins, run the disable_plugins utility located in the bin directory of the installation. You can use the Plug-ins panel (from the Tools menu) to re-enable your plug-ins.
  • More flexible update mechanism. If multiple updates for a plug-in are found, they are now all shown. Previously only the most recent update was shown.
  • Improved user feedback. You now get popup messages when no updates are found for installed plug-ins, and when you install a plug-in incompatible with your version of HIPE.
  • Many improvements to plug-in installation:
    • The installation process is now much faster for large plug-ins, such as those containing data pools.
    • The HIPE interface is no longer unresponsive while installing a plug-in.
    • You can follow the progress of the installation in the progress bar at the bottom of the HIPE window.
  • Fixed a problem with the scanamorphos plug-in, which would not read correctly some scanamorphos output. The plug-in assumed that the PrimaryImage HDU contained a 3-dim array. Output containing a 2-dim array in the PrimaryImage HDU was not read correctly.


If there is a new user release available, HIPE would notice it and offer to install the new version.


PlotXY is now able to display images (i.e. 2D data). Look at an advanced example here.


TablePlotter can display plot with Date/Time format.


OverPlotter can display plot with Date/Time format.

Dataset viewer

Astronomical utilities

  • Planck class moved from spire_ia_pipeline_spec_util to ia_toolbox_astro. If you are doing an explicit import (it is not needed) you have to change it to (preferred) import herschel.ia.toolbox.astro.Planck or from herschel.ia.toolbox.astro import Planck
  • New task astrometryFix: adjusts the absolute pointing of map data to give the best possible agreement in overall astrometry with some ancillary data. This is done either by cross correlation of the map(s) with ancillary map data, or by stacking the map(s) at a list of known source positions (e.g., from radio observations).

Numeric routines

Data fitting

Data interpolation

Random numbers


Trend analysis


Wavelet transform

  • Continuous Wavelet Transform synthesis operation now adds low frequencies for a reliable reconstruction
  • Continuous Wavelet Transform is able to deal with Complex1d signal

Other functions

Source extraction

  • Improvements to handling of pixel coordinates and region of interest in sourceExtractor* tasks:
    • Output source list now contains pixel coordinates of source (x, y) = (column, row)
    • Added "getWcs" and "setWcs" methods to the SourceListProduct/Dataset
    • Added "roi" parameter (region of interest) to tasks
    • Removed "corner*" and "returnPixelCoordinates" parameters from tasks
  • Non-standard columns are now accepted in a SourceListDataset, e.g., when converting from a TableDataset (an exception will be thrown if there are no standard columns at all)
  • Removed "subtractMedianBackground" parameter from sourceExtractorSussextractor (it didn't do anything)
  • Minor bug fixes, performance improvements and various internal changes



  • The methods addCircle, addEllipse and addRectangle is not longer dependent on the zoomfactor of the image.





  • The SpectrumFitter now recognizes a CUBE such that the user can give the x,y of the spectrum. The script as written by the SpectrumFitterGUI also takes the correct spectrum.

  • avg now warns if wavescales are not sufficiently aligned to get a good result

  • various changes when applying toolbox tasks to cubes, in particular avg, select and extract now allow you to work on several pixels/spaxels within the cube rather than only the entire cube.

  • resampling task now allows alternative ways to specify the grid for resampling.

Data cubes



Herschel Science Archive

  • HIPE will now store username and password in encrypted form outside of properties files (if so requested -- one can also choose not to store the username and password). In any case, it is recommended to remove HSA username and password from all properties files.

Calibration sources

  • All SED models for asteroid observations have been added to the ESAC database.
  • The Calibrators data has been reformatted, so that spectra are now stored as SimpleSpectrum objects, allowing them to be viewed with standard HIPE spectrum tools.

Products and datasets

All products

  • Removed the legacy diff methods from Product and Dataset interface. All functionality moved into ProductDiff class which is available as part of the utility toolbox (herschel.ia.toolbox.util)

Observation context

Product Access Layer

  • A "Cache" entry was added to the HIPE Preferences. Under this entry, a panel is shown with an overview of all existing caches on the system. One can also clear out these caches from this panel.
  • Product Browser: A simplified Observations tab was added, that allows user to list all observations on local disk, just by hitting the Run button. The data source selector has also been re-designed to be more flexible and intuitive.


  • General clean up of tasks:
    • All mandatory parameters must be filled in by users: removed default values from mandatory parameters, warnings issued (detection of programming mistakes)
    • Improved parameter checking and messages on bad data: Validators in primary inputs, they return false on error and log a warning (detection of programming mistakes)
    • All task GUIs and nodes should now show tooltips
  • Added some new user-friendly parameter editors (see For Task developers only)
  • Default task layout:
    • Parameter name tooltips now show the parameter name (for long names)
    • big modifiers can occupy a full row
    • groups (tabs) can be navigated with the mouse wheel
    • Parameters can be reset individually via contextual menu on parameter names (vs resetting all parameters with Clear button)
    • Parameters can list the variables they can accept via contextual menu on parameter names (to know before hand if we have the proper data to use a parameter)
  • For Task developers only:
    • Task registration will issue warnings if either the task was already registered or the taskname was already used or both.
    • Command generation: added converters so that arbitrary types can be represented with Jython strings (part of better visibility for defaults)
    • New shared Modifiers (parameter editors):
      • Added Editor for selecting a subset of a list of values (JListModifier, N values of M possible, M >= N)
      • Added a variant of JOptionModifier (Combo box) that returns some values but shows others (example: parameter uses integers but we show strings to users)

Misc Tasks

  • OpenTask parameter names have changed capitalization to comply with lower camel case

Quality control

Systematic product generation

Data input-output

  • URM entries (documentation) improved
  • Task importUFDirToPal renamed to importObservation
  • Task exportPalToUFDir renamed to exportObservation
  • Task simpleAsciiTableReader extended to support all options of asciiTableReader via Advanced tab. (eventually, asciiTableReader will subsume simpleAsciiTableReader)
  • Improved type guessing for table reading (affects (simple)AsciiTableReader): for floating-point data it will detect precision loss and maintain representation. If a floating-point number will lose precision as a float (32-bits) a double (64-bits) will be used instead: up to 17 significant digits but in these case it will use twice the memory.
  • Improved documentation and information provided by getObservation
  • saveObservation now reports the progress in the status bar and can be interrupted/stopped.
  • Added contextual menu option to Variables View to save tables with asciiTableWriter. Also added a validator to the task that will restrict its applicability to TableDatasets that can be fit into an ASCII table.

Virtual Observatory

  • Explicit support for DS9 was added. Right-clicking on a product will now show "SAOImage DS9" always, even if it is not running. In previous versions, DS9 had to be running for the menu item to appear. Requires DS9 version 5.3 or greater.
  • Upgrade to JSAMP 1.3, which targets version 1.3 of the SAMP protocol.
  • Improved GUI feedback using the HIPE progress bar when sending and receiving data.
  • The connection to the VO is now automatically established on HIPE start-up, to improve usability.


Sometimes, especially when a user has a shaky internet connection, an installation might fail in such a way that the local (in the user's machine) repository of hcss libraries might get corrupted. The workaround that existed so far can be found in the known issued page in: http://herschel.esac.esa.int/twiki/bin/view/Hcss/KnownIssuesHipeInstaller#Installer_fails_with_a_zip_error. The solution in short was to remove the local repository and all existing HIPE installation and start over. However a new solution that hopefully solves this issue has been implemented in 8.0. We know that bad proxy setup, hardware issues in routers can cause packet loss even in TCP. Such loss should always result in a file which is smaller than actually intended. Normally such a situation should result in an exception, but with packet loss in e.g. a router this will go unnoticed and a normal EOF completes the transmission. However, the resulting ZIP file will then always fail to unzip. The installer has been modified to extract the content-length from the HTTP header and validate the number of bytes actually received and raise an exception if such thing takes place. Moreover, a better exception reporting has been implemented to give more information on the basic cause of the exception. Finally the user installer reports the estimated time of completion while installing to give a rough feeling to the user.



HIFI Software Configuration and Properties

HIFI Calibration

  • Updates to Calibration tree (HICAL-30)
    • Spur table (7b)
    • FPU HK Thresholds (5a) (Prevents unnecessary Quality Flags)
    • Units added to columns (Prevents unnecessary Quality Flags)
    • Wbs Bad Pixel (saturated regions) (Causes necessary Quality Flags)
    • QC flags for bad LOU currents (Causes necessary Quality Flags)
    • Uplink auxiliary product parameter filtering (First steps in ensuring information about the way the observation was carried out is put in the meta data)

HIFI Pipeline

hifiPipeline task

  • Can now select calibration version to reprocess data with (command line only)

  • By default the pipeline now runs without saving products. Previously it always wrote output to .hcss/lstore/pipeline-out, which in theory allowed users to access processed observations from that location. This was rarely used and could cause disk space problems.

  • Improvements to tooltips in the Pipeline GUI

Level 0 Pipeline

  • DoUplink now uses the new uplink auxiliary product to set some observation parameters in the meta data, e.g., mapReadoutSep and mapLineStep

  • Suppression of unneeded Quality Flags: Quality check for mixer voltage now accounts for magnet tuning.

WBS Pipeline

HRS Pipeline

Level 1 Pipeline

  • New (optional) pipeline step allowing smooth band pass to eliminate standing waves: DoFilterLoadsTask. This is an alternative calibration scheme for strong continuum sources.

  • Substantial refactoring of CheckFrequencyGrid to correct assumptions made about the sequence of on-, off- and hc-datasets. This primarily affected HEB band data and resulted in unnecessary Quality Flags being raised.

Level 2 Pipeline

  • Sideband gains coefficients added to meta data

HIFI Products

  • Meta data calVersion is applied to all calibrated products and has format HIFI_CAL_version_number

  • Updated browse image for point mode observations.

HIFI Spectrum Toolbox

Spectrum Explorer (HIFI specifics)

Spectrum tasks (HIFI specifics)

  • Introduction of FlagToolTask, which allows you go through every subband of a SpectrumDataset and interactively flag spurs, lines, bright lines, and data to ignore. Input data can be an Observation context, a HifiTimelineProduct, or a SpectrumDataset

  • PolarPair is now a task

HIFI Data Processing Tools

Standing Wave Removal

  • Average mask option in FitHifiFringe introduced.
    • Allows a line mask to be calculated based on the average of all scans. Makes fringe removal from maps much more efficient.

  • Variable result_sds should no longer created if FitHifiFringe is run from the GUI with obs_or_htp input

  • When an Observation Context is supplied standing waves are now automatically removed from both upper and lower sidebands.

Baseline Removal

  • FitBaseline now operates in interactive mode as default

  • The automask check box now only has "True" or "False" options, previously the grey box "ignore" was operational and could allow you to misunderstand what you were asking the GUI to do.

  • You can now enter the back-end to be corrected directly into the GUI.

  • FitBaseline 'divide' option now handles negative signals robustly.


  • Some output from doDeconvolution is now written to the console as it runs - helpful to know HIPE has not crashed!

  • doDeconvolution now picks up the applied sideband gains from the meta data and "undoes" their effect on the incoming DSB data. Then the user running the deconvolution can then apply gains on or gains off. The pipeline supplied gains can in this way offer an optional starting point for the gain iteration inside the deconvolution.
    • This only has practical application for band 2 data at present.

  • New redundancy product allows you to overplot redundancy on the SSB solution. Helpful to identify where the SSB solution may not be reliable.

  • SSB spectrum now has y-axis labels "Antenna Temperature (K)", x-axis label changed to "Frequency (GHz)"

  • When diagnostic mode is turned on, doDeconvolution now produces an interim product in addition to the pop-up plots.

  • The min chi squared and the number of iterations are now included in the meta data of the SSB solution

  • Bug fix: In doDeconvolution an algorithmic error in binning the input data has been corrected. The fix has stopped the occurrence of two problems: Some band-edge channels had fluxes set to zero and some very narrow lines had their peaks shifted by as much 1/2 MHz. Both errors are fixed.

Spectral Cube creation

  • BeamSize is now taken from the calibration tree if "hpbw" is in the metadata
    • If the "hpbw" is not in the meta data, the beam size is calculated in the same way as in convertK2Jy using the rfreq in the meta data

* Can rotate maps using new parameter, FlyAngle

  • GUI behaviour modified:
    • Picks up information from meta data to provide better default input parameters
    • Changes to values in the GUI form now only go in one direction, e.g. changing PixelSize changes the Beam and Filter Parameters but not vice versa.

HiClass export tool

  • You can now create keywords when exporting to CLASS fits format.


  • HIFI User's Manual changed name to HIFI Data Reduction Guide, chapters rearranged into more logical order, addition of a changelog
    • Addition of introductory chapter, "HIFI Launch Pad"
    • Addition of a DBS Point Mode cookbook
    • Addition of chapter, "How to flag and remove flags from your data"
    • Addition of chapter, "Unit Conversion"
    • Updates to chapters on "Pipeline", "Deconvolution", "Creating Spectral Cubes", "What was done to my data" (clarifying units at each stage of the pipeline),
  • Updates the examples in the HIFI Pipeline Specification Document
  • HIFI User Reference Manual (URM) improvements
    • All HIFI tasks in URM now have a category and show up in the categorised view of commands
    • The categorised view of commands now is grouped into headers that should be meaningful to astronomers
    • Many URM entries have been improved so that decriptions are clear and examples are useful





  • interactive reduction scripts for all observing modes are now available
  • flatfielding is available in the interactive pipeline scripts
  • error propagation (and error bars) is implemented in the PACS rebinned cubes (STDDEV dataset)
  • spg pipeline: the wavelength switching pipeline is merged with the unchopped pipeline and removed
  • improved documentation and user friendliness of the pipeline tasks within hipe


  • The calibration sets that are installed on your local system can now be inspected with a 'Calibration Sets Viewer'. The viewer is located in the menu Window -> Show View -> Workbench -> Calibration Sets. Select a version from the available buttons, the you can inspect the release note for that version or the product list with the versions of the individual calibration products.



  • major improvements of the flux calibration:
    • implementation of the background normalization method for chop/nod observations
    • it is now based on observations of astronomical standards and the calibration block attached to each observation. The pipeline task specDiffCs computes the detector response of the current observation.
    • implementation of a correction of transients for unchopped observations

  • the spectral (pixel) dependency of the spatial calibration is now available for each pixel

Analysis tools

  • The Preferences now contain a Logging panel under the general tab which allows the user/developer to increase or decrease the number of logging messages send to the
  • The calibration file version is now also shown in the product history.


MapMonitor got an improved interface and allows to flip north up-or downwards.

PhotProject got a companion task MiProject (MapIndex-Project). If you have a full MapIndex, you may use MiProject instead of PhotProject. It needs frames and the MapIndex as parameters and speeds up projection time by a factor of 20 (30, 40?). Whatever the number is: its much faster than PhotProject. Use as:

 map = miProject(frames, mapindex)  

A new Task has been developed. It deglitches without MapIndex and runs with significantly less memory abuse than the IIndLevelDeglitchTask. The name and package is

. The hipe alias is mapDeglitch. It is documented as javadoc, URM and on the online deglitching documentation page:

A new Task has been implemented which removes crosstalk effects by masking (CrossTalkMask ) the first rows of detector arrays :

 frames = photMaskCrosstalk(frames) 

The HIPE Menu Pipeline -> PACS -> has been re organized and ipipe scripts added.

Automatic pipeline scripts save the Products at the end in the Observation Context also in IA sessions. That mean the scripts can be executed in a sequence without interaction.

Pacs Products (for example Frames, Ramps, PacsCube) got a method to rename a Mask :

 renameMask(String oldname, String newname) 

up to now, a mask had to be removed and created new, if the name had to be changed.

PhotProject, MiProject, IIndLevelDeglitching and some other tasks report their progress to the hipe progressbar instead of the terminal output.

The effects of 2-bit rounding on second level deglitching have been reduced in the pipeline for blue parallel maps.

The Pacs MaskViewer has been relocated to a new package:

. If you manually imported the old location (from herschel.pacs.signal import MaskViewer), the import will fail. Please just remove the import. Everything should work fine without further action. If there should be a problem nevertheless, please import the new location:
 from herschel.pacs.gui.maskviewer import MaskViewer  


  • GUIs are now available to work on spectral data: SpectrumExplorer and spectral footprint
  • implementation of user friendly viewer for the PACS specific part of the ObservationContext:
    • Each spectrometer slice now shows after the slice number a short description of its contents
    • The tooltip of the slice shows the complete list of slice info keywords.

Product changes



  • Browse products and browse image products are now available in spg
  • Red and blue camera spectra can now be combined in a single context in the level-2 processing, therefore, the lineId has changed for the red camera: 100 is added!


  • PACS Data Reduction Guide. Chp 1 is now a collection of Launch Pads: on getting Data into HIPE, doing your PACS Spectroscopy and doing your PACS Photometry. These are quick-start guides to working with your PACS observations. Chp2 tells you how to load your data, your pipeline script, and your calibration tree into HIPE, for photometry and spectroscopy both.
  • Spectroscopy: the pipeline chapters are being re-organised to follow the new pipeline scripts, and so for this 8.0 RC4 version of the PDRG only the LineScan Chop-Nod pipeline is described.


Calibration Products

  • The Spectrometer Beam Parameters calibration product used by the Point Source Flux Conversion now contains the wavenumber-dependent point-source flux conversion factors for each spectrometer detector. Previously, this product contained factors only for each spectrometer array.

  • The Spectrometer Nominal PCAL calibration product contains the weighted mean and standard deviation PCAL level separation for dark sky in Nominal Mode. This calibration product is used by the new PCAL Gain module.

  • The Spectrometer Bright Mode Gain calibration product contains the intercept and gradient for the wavenumber-dependent gain. This calibration product is used by the new Bright Mode Gain module.

Calibration Framework


Engineering Conversion

Common Pipeline

The default value for reconstruction in the sigmaKappa deglitcher is changed to "DIRECT".

Photometer Pipeline

Spectrometer Pipeline

Averaging Spectra

The Average Spectra module no longer supports the bandEdge parameter. This parameter was used only for the module's defunct outlier rejection functionality, and so it is no longer needed. Please ensure that your scripts do not use this parameter.

Spectrum Flux Calibration

  • The Flux Calibration module has been replaced by two modules, the Point Source Flux Calibration module and the Extended Source Flux Calibration module. References to SpecFluxConversionTask should be replaced with SpecPointFluxConversionTask or SpecExtendedFluxConversionTask. Read more in the pipeline specification manual at Spectrum Point Source Flux Calibration and Spectrum Extended Source Flux Calibration.

  • The Point Source Flux Calibration module can now be applied to more than just the central detectors.

Spectral Unit and Unit-Flexibility

It is now possible to specify the spectral unit of the initial spectra. The Fourier Transform module now accepts an optional spectralUnit string parameter to specify the desired output spectral unit. For example, if spectralUnit='GHz', then the output spectra will be in units of V/GHz versus GHz. If no unit is specified, the output spectra will be in units of V/(1/OPD_UNIT) versus 1/OPD_UNIT (the OPD unit is typically cm).

To accommodate this new feature, all spectral modules now accept spectra and calibration data products in any unit of wavenumber or frequency. For example, if a module is provided with spectra in units of V/GHz versus GHz and a calibration product in units of cm-1, then the cm-1 calibration product will be converted to GHz, and the output spectra will be in GHz.

Bright Source Mode

Two modules have been created for processing bright source mode observations. Both of these modules are in the new herschel.spire.ia.pipeline.spec.gain package:

  • The PCAL Gain module computes the ratio of a measured PCAL flash level separation to the mean PCAL flash separation measured on Dark Sky in Nominal Mode.
  • The Bright Mode Gain module computes and then applies a wavenumber-dependent gain factor for observations which have been made using the Bright Source Mode.

Phase Correction

  • The Phase Correction module's interface has been simplified. As before, the module provides three ways of applying phase correction, but executing the module is simpler now:
    1. From the observation data:
      sdi = phaseCorrection(sdi=sdi, avgSdi=avgSdi)
    2. From a calibration file:
      sdi = phaseCorrection(sdi=sdi, nonLinPhase=nonLinPhase)
    3. From a polynomial fit to the observation:
      sdi = phaseCorrection(sdi=sdi, sds=dsds, phaseCorrLim=phaseCorrLim,\
            polyDegree=2, pcfSize=64, convolApodName="aNB_15")
            polynomial fit to the observation

  • The Phase Correction module creates four new quality control metadata on the phase-corrected interferogram. These values will be used by the Quality Control module to assess observations:
    • phaseWrapSLW/SSW counts the number of times that in-band phase changes by +/-1.5pi in any of the nominal SLW/SSW detectors.
    • residualPhaseSLW/SSW gives the maximum absolute phase value (in units of degree) inside the phase correction limits in any of the nominal SLW/SSW detectors.

Create SDI from SDS

  • InverseFourierTransformTask has been created to transform SDS objects into SDI objects by applying the inverse Fourier transform:
    sdi = inverseFourierTransform(sds = sds)
  • InverseFourierTransformTask is not the inverse of the FourierTransformTask because it does not remove zeropadding from the output interferograms.

Convenience task for spectral cube creation

  • SpireProjectionTask has been created to provide a convenient single API for creating spectral cubes.
  • The task can make spectral cubes using any of the following algorithms:
    • Naive projection
    • Nearest neighbour projection
    • Gridding projection

Interactive Analysis and Tools


The following new features are introduced in Hipe 8:
  • Level 2 deglitching is available through setting l2DeglitchRepeat parameter.
  • The 5 individual output parameters are combined to one ArrayList parameter where each of 5 parameter can be retrieved base on its position.
  • The new diagnostic product is introduced which provides with useful diagnostic information.
  • Both Tod and TodBuffer are acceptable input for Tod parameter.
  • The input parameter "array" now is MANDATORY.
  • Pixel size can be changed through setting pixelSize parameter.
  • brightSource parameter is changed to boolean.
  • A boolean parameter, "fitWithWeight" is introduced.
  • Interruption is added so that the destriper can be terminated at the user's request.
  • Tod files stored in temporary directory are available until Hipe is terminated.
  • The destriper speed is increased 30%.
  • The input parameter name "Level1" changed to "level1".
  • The input parameter name "Tod" changed to "tod".
  • The input parameter name "L2DeglitchRepeat" changed to "l2DeglitchRepeat".

Baseline Removal

  • New "roi" (region of interest) parameter for baselineRemoval* tasks, specifying the region of the sky to be used in estimating the baseline
  • Removed "doInterpolation" task parameter
  • Minor changes and bug fixes (for example, the tasks now do ignore masked data samples when a TodBuffer is used internally)


This topic: Public > WebHome > HipeWhatsNew > HipeWhatsNew8x
Topic revision: r50 - 2011-11-14 - KatrinaExter
This site is powered by the TWiki collaboration platform Powered by Perl