backcorr makes a background spectrum appropriate to the accumulation region and position used for the source spectrum. The output of the task is an “EPIC FITS source background spectrum” as defined in the SSC data products ICD [1].
backcorr is a metatask that calls arfgen twice in its so-called extended source mode to compute the integrated effective area curves over the source and background regions. The extraction regions, which should be in detector co-ordinates, are read directly from the input spectra (both source and background), as the information is stored in the files using the data-subspace. This explains the presence of the source spectrum (i.e. the spectrum extracted from the source region, in channel and counts) as an input of backcorr. Then backcorr takes the input background spectrum (also in channel and counts), the arf files and makes the appropriate correction (in a task called make_backcorr).
The user can choose to keep the intermediary arf files produced in the calls to arfgen. Should the user wish to correct for the bad pixels and columns in the chosen regions, he/she can choose to provide the event file from which the spectrum was constructed. This file contains a bad pixel extension with the necessary information for making this correction.
Before going into the details of the task description, it is worth stressing that backcorr is primarily intended for use in the PPS. It could also be used in the IAS, but in many cases more sophisticated analysis will be required to get an accurate and reliable background spectrum. backcorr must be robust, work for point sources and for observing modes which allow the extraction regions to be defined. Its main input is a background spectrum which has corrections made to it. The details on how this input is determined is beyond the scope of the task, and should be investigated separately.
