Re: Exposure comparisons with CCD cameras

Daniel Borcard

Here is a point that may be important:

Actually, the stack is not summed, rather it is Median combined. Don't know if that makes a difference. It allows stray hot pixels to be subtracted out in the final image.


Median combined can be an issue for very slight background variation.
How confident are you of the software auto-scaling... I would do a
simple mean for the comparaison, even if you end up with a few hot

-- benoit

Statistically speaking, as an estimator of central tendency, median has a larger standard error than arithmetic mean: the median value varies more from one estimation to the other than the mean value. Here, every pixel shows an independent estimation of the sky background level. Consequently, a median combination does show a noisier background, especially with a small number of subs, because the value of the sky background as estimated by the median varies more from one pixel to the other than the mean does.
I second Benoit's suggestion: try a mean combination. 
In true life I always use mean, with rejection measures for hot pixels or extreme values (e.g. satellites): hot pixel mapping, sigma (i.e., standard deviation) rejection where pixels are rejected from the mean calculation if their value are too extremely outlying.


Daniel Borcard
Observatoire du Geai Bleu
Les faits sont têtus. Les nier ne les fait pas disparaître.

Join to automatically receive all group messages.