Date: Mon, 20 Mar 2006 15:22:17 -0800


On Mar 20, 2006, at 10:54 AM, Michael Koppelman wrote:
>
> The 1/SNR method basically tells us how many photons you got. The 
> standard deviation of a constant star tells us (in aggregate) about 
> your seeing, whether clouds came through, how well your aperture 
> centering worked, etc. You can't measure a variable better than you 
> can measure a standard star (with some assumptions about magnitude and 
> color).
>
> My original post was to suggest that the comp star standard deviation 
> method is barely more difficult than 1/SNR but gives us much more 
> information.
>
> Cheers,
> Michael Koppelman
>

Last year, during the last 'great error debate' I had quite a time 
trying to understand how to actually derive the numbers for the two 
methods.   Now that I've done it a few times it does seem easy, but I 
still remember the struggle to come up with numbers.  Frankly, I didn't 
find the CCD manual very helpful.  I've also been surprised at how 
different the numbers are.  The STD of comps is always more pessimistic 
so I've elected to just report that, tho I guess the purest would 
combine these figures.

It would have been helpful to me to have had a cookbook style tutorial 
on how to derive these numbers.  Getting ramped up to report them would 
have been easier and it would have given a clearer picture of what was 
needed.

Bill

_______________________________________________