Since ca. Oct 2009 (or version 1.7) MIA+EWS takes into account that the PRISM spectrum is moving on the detector. Previously the information about the curvedness of the spectrum had been stored in the WCS (World Coordinates) part of the FITS headers. There is now a function wcscurve (in MIA/mask_utilities.pro) that extracts this information to display lines in the miamask GUI.
For MIA
The change is twofold:
- On the one hand MIA/utilities.pro now includes some extra lines (marked by CAH, Christian A. Hummel) that call c/bin/oirChopPhotoImages in such a way as to include a so called curve reference file. This specifies where on the detector the spectrum lies and what its shape is. This is only really important for GRISM data where the spectrum is significantly curved. The curve reference file had been around for some time but hadn't been used in MIA+EWS before. When called with the option for this curve reference file, oirChopPhotoImages takes these into account for the background reduction with the sky windows.
- On the other hand, MIA/mask_utilities includes new code (marked by CAH) that reproduces this functionality for miamask. Note that miamask calls oirPhotoImages for displaying the backround image, i.e. the background image is already background-subtracted. Nevertheless, the newly shown yellow curved lines (when you select one of the photometry files) are shown to display where the sky windows have been set. These are curved, which isn't necessary for PRISM (see above), but this seems to have been the safest way to implement this functionality without altering too much of the code.
For EWS
The above doesn't apply since EWS doesn't call MIA/utilities.pro but EWS/idl/ews/midi/utilities/midiUtilities.pro where this changed has not (yet?) been applied, i.e. if you want to take care of the shifting spectrum in EWS to do a decent background subtraction for your 'photometry' (uncorrelated spectra), you have to do it manually. There are some tools in MIA+EWS that can help you with this:
- c/bin/oirShiftMask takes an A and a B photometry file and a reference mask (e.g. the
prismshiftmask.fits
that you can get on request) and shifts the mask to maximise the flux under both spectra, see the description in /c/src/exec/oirShiftMask.c.
Note that in oirChopPhotoImages.c there is a DEBUG
option that saves some intermediate results into (ASCII) files.
You should do the mask shifting with a bright calibrator star and then used this shifted mask for your science files.
Ideas for improvements
- Store the position of the sky windows in the mask file FITS headers.