SPLOT pixel coordinates and interaction with V2.11 CCDPROC

Frank Valdes wrote on Sep 15, 1998

Q:  One of the things I've done for years is use (abuse?) splot when I'm
taking spectroscopic data to get an estimate of the intregral-over-the-spatial
direction.  If I'm using a chip oriented so that dispaxis is 1, and is
formatted 2048x500, say, I'll do an splot image[1024,*].

This will give me a spatial plot, and I can then locate the "peaks" of
corresponding to (multiple) stars along the slit, and measure how many
counts I REALLY have by using the "e" key and looking at the "flux" to
give me a good background-subtracted-integrated-over-the-spatial-direction
count.

The one funky thing I've noticed is that now the x axis in such a plot
seems to be the original CCD locations---rather than going from 1 to 500
in the above example, it might go from 900 to 1400, if I had gotten
my 2048x500 chip by reformating it from a 2048 x 2048 and taking
a swarth near the middle.

I can live with this, but it's arkward---it's nice to be able to use
the imagge display to locate the individual spectra I might want to
measure!

Is there someway around this new feature?

One thing I've just noticed is that this business about the "x" scale
being the original CCD coordinates---this happens only AFTER the image
is run through ccdproc.  Prior to that, the coordinates are "local".
i.e., go only from 1 to 500, rather than 900 to 1400!  Has it always
done this?



A:  In V2.11, released in April 1997, there was indeed a change in CCDPROC
to maintain the CCD coordinate system (which corresponds to what is
called the "physical" coordinate system in some IRAF tasks).  SPLOT has
always (for at least for 5 years) plotted the physical coordinates.  So
it is the change in CCDPROC which you are seeing as you noted.  One
reason for the change is to keep track of bad pixels through operations that
change the size of the data; i.e. the bad pixel masks match data
through the physical/CCD coordinate system.

SPLOT can currently toggle between pixels and wavelength (if there is a
dispersion solution).  Maybe SPLOT should allow a third coordinate system,
the "logical" coordinate system which corresponds to the pixel array of
the image (first pixel is 1,1).  However, the workaround for you is quite
simple.  If you want to remove the CCD pixel mapping from CCDPROC use
the following command (which can work on a set of images)

cl> wcsreset <images> wcs=physical

This command removes the mapping between logical and physical
coordinates (the LTM/LTV keywords) so that the two are the same.

---

Last post on Sep 15, 1998