As for SPIRE FTS spectroscopy, semi-extended sources provide the user with a very special set of problems, as they are sources of a size intermediate between point sources (i.e. unresolved) and extended sources (clearly resolved), for which neither point source reduction techniques, nor extended source techniques work well.

We assume that you have watched the Herschel Science Centre short videotutorial called Starting with PACS Spectroscopy:

https://www.youtube.com/watch?v=rcY487Z9Kyg 

Or, at least, are familiar with the contents of this short videotutorial and its associated documentation. There is a Herschel Science Centre short videotutorial that will guide you through the process of processing and analyzing semi-extended sources. You can find the short videotutorial here:

https://www.youtube.com/watch?v=PNSuxM105UY

A pre-requisite for processing semi-extended sources is to know:

1. The surface brightness distribution.

and

2. The off-set of the source from the centre of the central spaxel.

If you do not have this information, you can still attempt to use these techniques (the short videotutorial will advise you on this), but may have larger errors in your calibration. 

You can find a package of useful support materials for this short videotutorial in the Herschel Explanatory Legacy library (HELL) in the Level 2, Data reduction, PACS spectroscopy section.

This short videotutorial you will be shown how to use the useful script that was developed to deal with semi-extended sources that extracts the spectrum of the central spaxel of a re-binned cube and applies a point source correction to it. It then compares the surface brightness of a point source with the known surface brightness distribution and creates an extended-to-point correction too that is used to generate the semi-extended source correction. It will show you how to apply the models and which data products to use for different observing modes. Finally, you will be shown some examples of PACS data processed with the script.

We recommend, as background information, that you consult the following Herschel Science Centre short videotutorials too:

Point Source Spectroscopy with PACS, https://www.youtube.com/watch?v=nYWTccEQz_M

PACS Cubes in Detail, https://www.youtube.com/watch?v=CxZF9-9D8UM

Or the following guide:

Dealing with extended sources observed with PACS spectroscopy

Extended sources provide the user with a particular set of problems that require special techniques.

We assume that you have watched the Herschel Science Centre short videotutorial called Starting with PACS Spectroscopy:

https://www.youtube.com/watch?v=rcY487Z9Kyg 

You should also watch the Herschel Science Centre short videotutorial called PACS Cubes in Detail:

https://www.youtube.com/watch?v=CxZF9-9D8UM

Or, at least, are familiar with the contents of these short videotutorials and their associated documentation. There is a Herschel Science Centre short videotutorial that will guide you through the process of processing and analyzing extended sources. You can find the short videotutorial here:

https://www.youtube.com/watch?v=j270VI0Ah1g

A pre-requisite for processing extended sources is to be familiar with the different spectroscopy products: cubes and tables that are produced when processing PACS spectroscopy.

We assume here that you have watched the Herschel Science Centre short videotutorial called Starting with PACS Spectroscopy:

https://www.youtube.com/watch?v=rcY487Z9Kyg 

And the Herschel Science Centre short videotutorial Point Source Spectroscopy with PACS:

https://www.youtube.com/watch?v=nYWTccEQz_M

Or, at least, are familiar with the contents of both short videotutorials and their associated documentation. If you have/are, you may find that using the PACS Pointing Off-set Correction Script will improve the quality of you results. There is a Herschel Science Centre short videotutorial that will show you how to use it and teach you the background information that you need to make effective use of it. You can find the short videotutorial here:

https://www.youtube.com/watch?v=f861D8CoYqc

The Pointing Off-set Correction Script will improve the spectra of point sources that suffer from pointing off-sets, or pointing jitter, provided that the off-set is no greater than 10 arcseconds. The script measures the average pointing off-set and its variation with time (the jitter). The script works best for:

1. Sources with a flux of about 10Jy or more.

2. Sources that are off-set by no more than 4 arcseconds.

For these objects you will get a much improved continuum shape and will be able to detect broad features much more reliably.

In contrast, the script will not work for:

• A source located at the edge of the central 3x3 spaxel box.
• Un-chopped spectroscopy.

And it will be unreliable for faint sources.

You can find the Pointing Off-set Correction Script in the HIPE Scripts menu and must be run from Level 0: i.e. with raw data. It is fully described in the PACS Data Reduction Guide. It will produce two spectra from the extractCentralSpectrum task - the so-called c9 and the c129 spectra - of which, the one with the higher flux is the one to use. You will see several examples of spectra of sources with off-sets from 2-6 arcseconds that have been processed with and without the Pointing Off-set Correction Script to illustrate the differences that can be found.

Assuming that you have watched the Herschel Science Centre short videotutorial called Starting with PACS Spectroscopy:

https://www.youtube.com/watch?v=rcY487Z9Kyg 

And the Herschel Science Centre short videotutorial Point Source Spectroscopy with PACS:

https://www.youtube.com/watch?v=nYWTccEQz_M

Or, at least, are familiar with the contents of both short videotutorials and their associated documentation, you will find that the Herschel Science Centre short videotutorial on Point source spectroscopy with PACS: point sources in the central 3x3 spaxels, which you can find here:

https://www.youtube.com/watch?v=v1W5pT9DYUk

will show you how to proceed if your point source is in the central 3x3 spaxels, but not in the central spaxel itself. This case is valid if you have a point source, with a pointed observation, in which the target is centred within one of the eight spaxels surrounding the central spaxel, or for a mapping observation, in which, for one of the raster positions, the point source is centred within one of the eight spaxels surrounding the central spaxel. You will learn to identify cases for which the process may not work well (contamination, poorly centred point sources, ...) You will learn how to identify the spaxel coordinates of a source and which HIPE scripts and tasks to apply in the different possible cases. It will also show you a "last resort" case that you can use if your source is not even centred in the central 3x3 spaxels. You will also be guided towards the other processing and post-processing that may be helpful, such as the Pointing Offset Correction pipeline, or the Extended-to-point Correction.

Assuming that you have watched the Herschel Science Centre short videotutorial called Starting with PACS Spectroscopy:

https://www.youtube.com/watch?v=rcY487Z9Kyg 

And the Herschel Science Centre short videotutorial Point Source Spectroscopy with PACS:

https://www.youtube.com/watch?v=nYWTccEQz_M

Or, at least, are familiar with the contents of both short videotutorials and their associated documentation, you will find that the Herschel Science Centre short videotutorial on Point source spectroscopy with PACS: point sources in the central spaxel, which you can find here:

https://www.youtube.com/watch?v=_cVncMct3NE

will show you how to proceed if your point source is in the central spaxel itself. The short videotutorial will show you how to use the Product Decision Tree to decide which products and tools to use. This case is valid if you have a point source, with a pointed observation, in which the target is centred within the central spaxel, or for a mapping observation, in which, for one of the raster positions, the point source is centred within one of the eight spaxels surrounding the central spaxel. You will learn to identify cases for which the process may not work well (contamination, poorly centred point sources, ...) You will learn how to identify the spaxel coordinates of a source and which HIPE scripts and tasks to apply in the different possible cases. 

The short videotutorial will show you how to use the extractCentralSpectrum task, which is designed for these cases, how it operates and the corrections that it applies. It will show you how to decide which extraction to apply to the data (c1 - extracts from just the central spaxel, c9 - extracts from the central 3x3 (i.e. 9) spaxels , or c129 - which extracts both and scales the central spaxel to the central 3x3 spaxel box), given the various possible combinations of source flux and centering, although as a general guide, whichever extraction gives the highest flux is the best one to use.

You will see examples of extractions of data showing you what the different extractions show and how to select the most adequate one for you, as well as simple Quality Control checks to apply on the results. It will also show you what other documentation and scripts may be useful to you, as well as giving you indications on how you can do these analysis outside HIPE too.

Answer supplied by David Teyssier:

 

The calibration uncertainties vary from one instrument to another. A summary of those is provided in each of the instrument overview pages, namely:

HIFI: https://www.cosmos.esa.int/web/herschel/hifi-overview

PACS: https://www.cosmos.esa.int/web/herschel/pacs-overview

SPIRE: https://www.cosmos.esa.int/web/herschel/spire-overview

The calibration uncertainties are provided in the Performance and Calibration section, typically with a distinction between absolute and relative (repeatability) accuracy. For HIFI a dedicated calibration uncertainty budget table exists per observation, and can be retrieved as an HPDP - see https://www.cosmos.esa.int/web/herschel/highly-processed-data-products.

The Herschel Science Centre has produced a number of short videotutorials about the analysis of data with the Herschel instruments. These contain information on the uncertainties in photometry and on the flux calibration of spectroscopy and how to treat the different effects that may influence it. You can find the videotutorials, which we recommend strongly viewing, here:

https://www.cosmos.esa.int/web/herschel/legacy-training-materials

Answer supplied by Katrina Exter:

The PACS Spectrometer Quick Start Guide has a summary of all the PSF information for the PACS spectrometer, and for more detail and tables, go to the PACS Handbook (calibration chapter). The Beams section of HELL has some more information and the beam files to download.

The PACS spectrometer beams are 2-D, i.e. they are maps, one per key wavelength. The maps have a small pixel size (0.5”) so the details of the beams can be seen clearly. Bear this in mind if comparing to images created from your PACS cubes, as their spaxels/spatial pixels are usually much larger (see the PACS Products Explained document for more detail). Bear in mind that PACS pointed observations spatially under-sample the beam; mapping observations were taken to compensate for this and the spatial sampling better. (For more information on this, see the PACS Spectrometer Quick Start Guide and the PACS Handbook)  

Note that the PA of the observation is important in the shape of the beam for any single observation — the beam rotates with the PA (this is true for photometry and spectroscopy).

If you wish to know what the actual shape of a point source is on an observation, you could look at observations of point sources: some point sources (and their ObsIDs) are published in the Herschel Cross Calibration paper. 

Note that if you want to know what your point source looks like in your specific observation, the location of the source in the FoV is important. The native spaxels of the PACS IFU are 9.4”, which is more or less the size of the beam in the blue and only slightly smaller than that in the red (see the QSG for numbers). However, if a point source is offset from the centre of the FoV by even as little as 3”, this can result in slight distortions in its appearance compared to it being located dead centre.

Answer supplied by Katrina Exter:

The filter transmission curves can be found via the link in HELL: http://archives.esac.esa.int/hsa/legacy/ADP/PACS/PACS-P_filter_curves/