DISPLAY: Where is the overscan?

Frank Valdes wrote on May 27, 1998

Q:  I am using now the new IRAF version and when I display an image the
regions of overscan don't appear. Doing "epar display" there is a
parameter called "nsample" that I have changed to 1124 because the size
in pixels of the digitised frame is 1124 (CCDXSIZE and CCDYSIZE on the
header), although the X and Y sizes of useful imaging area on the
header is 1024 (CCDXIMSI and CCDYIMSI), but it doesn't show these
regions. I need to visualize them for the next step, where I have to
give the coordinates of the biassec and trimsec.


A:  There is nothing in the display process that specifically excludes
a part of an image such as the overscan.  The automatic scaling
algorithm is designed to provide a greyscale range near the median
level.  Usually the overscan is significantly lower and so my best
guess as to the problem is that the overscan level is just appearing as
black because of the automatic stretch chosen.  You should turn off the
automatic algorithm and set the z1/z2 values.  Changing nsample will
have very little effect on the automatic values chosen.  The contrast
parameter can be adjusted to increase the range around the median for
the automatic algorithm.

The second, less likely, possibility is that the frame buffer size is
smaller than the image.  If you do not use the fill option in DISPLAY
then only the central portion of the image that fills the frame buffer
will appear.  The solutions are either to set a bigger frame buffer,
with stdimage, or use the fill option.

Finally I would comment that if you want to accurately measure the
position of the overscan columns then I find using plots, specifically
IMPLOT, to be a better way.  Use IMPLOT and then set the x extent, with
":x <value> <value>" around the overscan region.  Then the 'C' can be
used to measure specific positions.

---

Last post on May 27, 1998