Re: Bits and pieces

[Tom Droege <tdroege2@earthlink.net>]

Ahhhh!  I just realized why this is true.  (I think)  I recall that Michael 
generates photometric corrections that depend on all the data taken.  If 
one does one image, then the photometry may be different???

For reprocessing I have been doing two CD's at a time.  This is usually 80 
V and 80 I images.  Takes about 4 hours.  The normal processing is 
everything taken on one night.

My current scheme is to put darks and flats in dated sub directories under 
darks and flats.  The program goes to darks and flats and gets whatever is 
there and puts it in the output directory.  Before the run, I just have to 
cp the darks I want into the darks directory from the dated sub 
directory.  Same with the flats.

I have used only light box flats for the "new" data.  The previous data 
used flats generated from the sky.  I have been looking at data from the 
combined set.  Looking for a break at the time of the process switch.  I 
don't see any.  It will be a huge program taking massive amounts of time 
just to try to see any differences from flats, I bet.

The folk lore is that you fuss infinitely with flats.  Possibly one can 
find examples where it makes a difference.  So far, I have not been able to 
see any affect of flats.  We made a huge change going from sky to light box 
flats with field flattening.  It should have made a huge difference in the 
data.  It did not.  So far nothing I do makes much of a 
difference.  Conclusion:  "It is the sky, stupid"  The only data that has a 
tighter error distribution is data taken on good nights following a 
star.  I argue that this is because the data was taken under similar sky 
conditions.

Having struggled with this for several years, I have concluded that I 
should do what I can do.  That is take data that has a noise floor 
somewhere around 0.03 mag.  I would like it to be better, but it may never 
be better here in Batavia.  I await data from ROB in a dark location for 
comparison.

Tom Droege

At 10:51 AM 6/3/03 -0400, Stupendous Man wrote:
>       b) we have a separate tool to perform basic image reduction
>                on a single frame, and then analyze the cleaned frame
>
>   The pipeline I wrote is not designed for step b).  One could
>use IRAF or XVista or any commercial package to do this ... but one
>would then _not_ get exactly the same results as the pipeline,
>because the exact steps and parameters used won't be the same.