Users who have CGS4 data since the upgrade to the 256x256 array will
have noticed their spectra having a small amount of curvature which
causes them to stray into neighbouring rows. This is caused by an
optical disortion in the CGS4 optics which has now become obvious
because of the increased wavelength coveraged and diminished spatial
pixel size.
Here are some suggestions for dealing with the above during data
reduction with various software packages.
CGS4DR
It is not possible to do curved extraction of spectra within CGS4DR.
Figaro
These routines can be used for both point and extended sources.
Commands to be used:
ICUR to mark the centre of the spectrum.
SDIST which will trace out the indicated spectrum
CDIST to correct the data using the SDIST output.
You may wish to consult the extensive documentation written by Keith
Shortridge on straightening spectra with Figaro, a copy of which
is locally
available .
IRAF
If you have been using the task noao.twodspec.apextract.appall (or just
apfit) to extract your spectra, you should be setting pfit=fit2d in
order to use the Marsh algorithm (PASP 101:1032, 1989) for optimal
extraction of curved spectra. For additional information consult the
APEXTRACT documentation (especially the revision summary dated May
1992).
Twodspec
The program COMB in the Starlink package TWODSPEC allows correction with
the same functionality as the Figaro comman SDIST. For more information
consult the TWODSPEC documentation (SUN 16).