Re: TASS data reduction

[aah@nofs.navy.mil]

Chris wrote:
>1) extract stars list with DAOFIND.  Produces very rough
>   photometry but good center locations even for blends.
  The centroids for the daofind portion of daophot are not all that accurate,
probably a quarter of a pixel or so.

>3)...
>   frame.  (I do think the PSF varies in ra.  The VCO drifts
>   and seeing can change.)
  The VCO may drift, but seeing has nothing to do with these images
(remember, fwhm ~ 2 pixels ~ 28arcsec).

>4) Determine intsrumental mag for each star in the list
>   by fitting the local PSF at each star's x,y location.
  If you are already using daofind, use allstar to do this part of the
extraction.  It handles blends nicely.  It also improves the centroiding
by about an order of magnitude.  One problem with using Chris' 4)
technique is that you get psf magnitudes.  The psf changes frame-frame,
and so it is difficult to get 'all-sky' photometry (i.e., one zero point
for the night plus some extinction corrections).  It works really well for
many observations of a single field, where you are doing differential
photometry.
  Someone should investigate DoPHOT as well.  This is an automated system
used by the MACHO or OGLE group (what a memory!), and might work well.
It doesn't hurt to explore a number of avenues at this point, but probably
TASS should converge on one extraction program before starting the real
survey so that the results are consistent.  It will be interesting to
decide whose version of reality is correct.
Arne