Re: Disk Set 18

[<aah@nofs.navy.mil>]

Tom wrote:
>...There is 
>also more long term drift than I expected.  So it looks like I need to add 
>the air temperature into the fits header so we can study drift vs air 
>temperature.   Possibly we can calibrate it out.
  If I understand your statement correctly, the drift is just
electronic bias drift.  If you will rewrite the sequencer PROM
so that you get overscan pixels, you can remove such drift easily
without having to resort to reading temperatures.  This is the
preferred way for separating out bias, dark and sky signals.

As long as you continue to have ice crystals, dust, differing
moonlight, etc. on a night-to-night basis, you need to make a
flat for each night.  So include sufficient data for making such
flats.

For compression, look at hcompress as we have discussed before.
It will in general give more compression than gzip since it uses
knowledge about astronomical images; lossless, it can usually get
2:1 reduction.
Arne