Re: Photometry

[Arne Henden <aah@nofs.navy.mil>]

I will state once again: as long as everything is reversable,
it doesn't matter whether you keep raw or processed frames.  The
only question is whether the processing steps are reversable.
For example, Michael R. could be trimming the non-imaging pixels
from the raw image; in that case, you cannot ever recover the
original image from any possible processing step.
   And again, for the common, normal survey mode of operation, you
will rarely go back to the original images anyway; you will be modifying
the extracted starlists, with improved astrometry, improved
photometry, different ensemble differential photometry, etc.  As Tom
mentioned, my original philosophy was to not save any images at
all, which was then modified to saving processed but compressed
images; however, that was for my planned operational mode of purely
survey, purely photometric calibration work.  My concern regarding
Tom's images was for the other uses of his data that were discussed
on this list.  If you do not intend to look for transient objects/events,
then proceed forward.  I'm not talking about changing the processing
procedure and redoing every image; I'm talking about archival checks
of a few images where a question occurs.  Even if you were to reprocess
all of the images taken so far, that is not impossible, as Tom has done
this already over the past few months; but whether Tom wishes to do
that again, or let someone else do it perhaps, is not my concern.
   It is Tom's decision, and I'm not going to second-guess.  Just
quit making blanket statements like you've done the best that is possible
or that you will never have to go back to the original images.
Arne

Tom Droege wrote:
> First, we are keeping the astrometry and photometry calibrations 
> separate.  So that takes care of that.
> 
> Next, the fits files we are keeping have only dark and flat corrections 
> and a flattening step.
> 
> What could we possibly do differently in the future?  I say nothing.  
> The darks and flats that have been used on the .fits files we are saving 
> are the ones we have at the time we take the data.  There is no 
> possibility of any improvement in the darks and flats in the future.  So 
> it is really the same as if we were keeping the object darks and flats.  
> Only I am presently keeping them combined in one image.  This means that 
> we do not have to make a decision in the future as to which dark and 
> flat set to use.
> 
> I realize that if we find some better way to do processing in the future 
> we can do it.  We just start at the step after the dark and flat 
> correction.  We can't do that any better in the future.
> 
> Tom Droege
> 
> 
> 
> 
> 
>