>>Registar is reckoned the best and easiest but it is expensive
>I know, I have tried the demo version and you can add and/or average
>the images.
RegiStar is a very good software, but it is intended only for make things easier when registering the images, not for stacking them. There are some other programs that makes things better in this area, like "Sigma".
>I understand that most people stack and average two or more images
>to improve the SNR.
>What I wanted to know is if anybody is stacking and adding several film >images
>(e.g. 4 or 5 ten minute set) to get deeper. This seems to be routinely >done with CCDs
I have been stacking up to 8 images (film) to go deeper. Just see this page:
(english version not avalaible yet).
The differences between all the methods described in Jerry's page (averaging, SMI, etc) are just a curves transform and how they make use of the resulting dinamical range. In fact they are almost the same if you are working in 16bits (or 32bits), so I just make things easier and just average them. An "upgrade" is to use some rejecting algorithm, like MIN/MAX or Sigma to avoid the hot pixels and some scratches. You can make this with "Sigma" or IRAF (in Linux).
Other big advantage of stacking several film images is that the noise is very effectively reduced. If you look in the other processing pages (Can Mayor and The Pudu specifically) you can see the big difference of stacking just 2 or 4 images.
