Subject: 
        MAPCOR stuff
   Date: 
        Fri, 31 Oct 97 10:40:28 MST
   From: 
        jhe@noao.edu (Jay Elias)
     To: 
        tchester@ipac.caltech.edu
    CC: 
        evans@ipac.caltech.edu, roc@ipac.caltech.edu, jhe@noao.edu




MAPCOR Parameters:

MAPCOR can be thought of as having two main functions. One is to remove
artifacts due to bright stars, which consist of:

1. Persistence effects due to array "memory", which cause residual
images at the position of a bright star in subsequent reads.

2. Ghosts -- referred to as "glints" -- in the images due to bright stars.

3. Elimination of sources that are unreliable due to proximity to
bright stars and their diffraction spikes.

The second function is to put all objects found on a uniform instrumental
magnitude system. This consists of:

1. Correction of bright sources (those that saturate Read2 - Read1) to
the Read2-Read1 system. This consists of using the Read1 magnitude plus
an empirical correction factor or, for still brighter sources, the first
persistence artifact plus an empirical correction.

---------------
response from t. evans:

the "still brighter sources" you
mention below are those that are saturated in Read1; perhaps that should
be mentioned specifically, to emphasize that the persistence artifact
method is only used for those extremely bright srcs.

---------------


2. Correction of measured magnitudes to "infinite" aperture. This is actually
correction to the point where the curve of growth increase is below
a certain amount or a maximum aperture size. This also includea mormalization
of PSF magnitudes to aperture magnitudes.

Note that this is done for subsets of data within a scan, depending on
seeing variations. Much of the information required for this processing
is produced by the previous processing step, PROPHOT.

=============================

Adjustable Parameters:

For the artifact removal, essentially all of the parameters are empirically
determined. Because they are related to both array parameters and 
optics properties, they should in principle be determined whenever there
is a major adjustment to the optics or an array is changed. They should,
in addition, be determined near the end of the survey to confirm their
stability.

Although the current approach for the start of operations is to use
observations of bright stars during normal operations to define
the parameters, I would argue that it is more efficient overall
to devote a small number of nights to measurements of bright stars
using the scan mode applied to standards. One night will produce a
large enough number of measurements, but may not provide enough range
in seeing properties. I believe that this approach is needed to
ensure that enough bright stars -- K<5 or so -- are measured to reliably
determine the parameters for artifact removal for these bright objects.

----------------
response from t. evans:

The need for a range in seeing is absolutely true; unfortunately, we
have yet to do much work on that, because with the range in seeing we
also need more PSFs, which are just now being made.  It will be very
important in the future though.

-------------

Note that the nights need not meet the requirements for photometric
conditions for regular survey data, but photometry should be good to
~0.1 magnitude or so.

The issue is whether it is a more efficient use of resources to find
the required data within survey results from many nights, or to
devote some number of nights to non-survey data in order to facilitate
subsequent analysis -- i.e. trade off effort at IPAC against telescope
time.

2. The parameters associated with magnitude corrections are for
the most part the magnitude ranges used to determine corrections
or for which the different data sets are valid. These are, again,
empirical, and can probably be checked using the data needed for
(1) to the extent that they need checking at all. 

These parameters are probably less sensitive to instrument properties
than those in part (1), except for the correction from Read1 to 
Read2-Read1, which is easily checked but which must be
accurate.

=================

Issues:

Everyone involved agrees that the large number of parameters required
by MAPCOR must be emprically determined. The only issue is that mentioned
above, namely how best to get the data needed. There is not really a
science driver for one particular approach; the concern is instead
optimal use of resources and it therefore is best reovled by the
people who manage those resources.



-Jay Elias


---------------
response from t. evans:

Thanks for the writeup.  I especially appreciate the emphasis that there
is a trade-off between telescope and IPAC time in gathering data on the
bright stars;
that's an issue that I feel the team needs to address.  

---------------