MSRED: flat fields and creating multiextension format

Frank Valdes wrote on Jun 26, 1998

Q:  I thought I knew the answer to this question, but have managed to get so 
confused that I had better ask you.

How are the flat fields scaled in when using ccdproc in MSCRED?
Are they scaled by one number for each chip, or one number for the whole image? 

I ask because I have some data from the FLWO 4-shooter, on which
one of the chips has a much lower QE than the other three. For example,
in a B dome flat, the median on this chip is about 3800, while the other
3 are about 5700. It seems like it would make a difference 
which scale is used - or does this not matter if the standard
star fields are reduced the same way? 

Also, what effect does the scaling done in mscstack have on 
photometry? The standard star fields are not going to have this
done to them, which kind of worries me..

Is there any way to put images into MEF format after
copying those for individual chips to separate files?
(You can guess why I might be asking this..)


A:  For reducing mosaic type data, such as your 4-shooter, the correct thing
to do and what mscred.ccdproc does is use a single scaling for all
CCDs.  When a flat field is processed mscred.ccdproc combines the means
over all CCDs to get the single normalization factor used when applied
to object data.  You may find that the flat field equalizes, that is
flattens, the 4 CCDs pretty well but not perfectly.  In this case you
will need to produce a sky flat, either a sky flat observation or combining
disregistered object exposures.

The MEF format is created by copying the images to the same file with
the FITS kernel append option.  If you have MSCRED loaded then that
option is already set in the "fkinit" variable.  You can explicitly
set append with a syntax such as:

	cl> imcopy a mef[im1,append]
	cl> imcopy b mef[im2,append]
	    etc.
	    
The MEF format used by MSCRED expects all data to be in extensions with
the primary FITS header being just global keywords.  I don't know if
the 4-shooter uses a global header.  The simplest case is the above
example.  By giving the extension name, "im1" in the example but it
can be anything, then if mef does not exist it knows to create a dummy
primary header and put a in the first extension with the specified extension.
The other appended extensions will be tacked on with the specified extension
name.

If you have a particular global header you want to use and have inherited
by the extensions then you have to do a first IMCOPY to set up the
primary header.  Suppose you have some MEF data, called orig, which
has the global header.  Then

	cl> imcopy orig[0] newmef		# Create primary global header
	cl> imcopy a newmef[im1,append,inherit]
	cl> imcopy b newmef[im2,append,inherit]
	
In the inherit option any identical keywords in the images a,b,...
with those in the global header will be eliminated from the extension
and will be taken from the global header when needed.

---

Last post on Jun 26, 1998