varisized images

[jamesRC]

Hi all,

This feels a bit off topic but I'll try it here.
Last night I used dithering for the first time after a long time.
Apparently the resulting subs vary in size, Enough so that PixInsight and Deep Sky Stacker won't work with them. I can't use  stacking or ImageInspection (latter only works with a small part of the total).

I used to use Nebulosity for this. Surely there's a way to do it in PixInsight? Some example image sizes are 16,178, 16,141, 15,992 and so on. I'm assuming this is caused by the dithering?!

James

[pfile]

something else is wrong - are you saying that the subexposures as saved from nebulosity have different frame sizes?

dithering does nothing but move the pointing of the telescope a little bit between exposures, nothing fundamental about the sensor can be changed by this...

rob

[jamesRC]

Hi rob,

Thank you for your response.

I mentioned Nebulosity just as a kind of reference, it doesn't have anything to do with this quandary. Sorry.
Briefly, collecting images using dithering with my Canon t3 and APT, produced a lot of images with slightly different sizes. This was a surprise and has prevented any further processing. I did try ImageRegistration using stars, and then ImageIntegration using the resulting registered images. Didn't work. Resulting integrated image had off-registration parts in it. Maybe something in the software, APT?

James

[pfile]

is APT saving CR2 files or FIT files?

anyway differing X/Y sizes will prevent calibration but it should not prevent registration (and thus integration). but dithering itself can't cause this.

only thing i can think of is that you've got a lot of hot pixels in the subs and staralignment has latched onto those as "stars" resulting in no actual alignment of the subframes. you can increase the noise reduction scale and the hot pixel removal in staralignment and see what happens.

rob

[jamesRC]

Hi rob,

I am closing in fast on user error causing this problem. It's been years since I worked on the CR2 files my Canon produces, and I didn't understand a lot of how that works. I'm learning fast in the APT support group and from you. I should be able to resolve this in a little while.

THANKS, rob.

James

[jamesRC]

rob,

an additional concern:

If these (slightly) different sized images can't be calibrated, then how am I going to process and stack them? Deep Sky Stacker 64 bit won't run because of these different sizes, and I don't know just how or if to do this in PixInsight??

James

[pfile]

well they can be calibrated as long as you can come up with darks/flats which match their pixel dimensions.

generally these problems occur when capture programs convert the CR2 files to some other format. since CR2 is a proprietary format, all of the 3rd party programs that handle CR2 have essentially reverse-engineered the file format. so you end up with, say, programs based on DCRAW doing something a little different from programs based on libRaw, which are both probably different than what canon's own software does. usually you'll find that one difference between these various programs is that the pixel dimensions of the converted CR2 files are very slightly different, like by 1-4 pixels.

but if you've got real CR2 files, then as long as you use the same software to create the masters and calibrate the lights, all should be fine.

anyway something is very weird if you have CR2 files that contain different pixel dimension images. what happens if you use canon's software to open the files?

rob

[jamesRC]

rob,

It's the user that's weird here. I've been beating over what I've been learning from PI and other forums and think I understand better now.

Using my fresh, bright understanding  , I tried manual postprocessing, choosing my file sources very carefully. It worked! There's a new problem though, which I will post under its own header. I consider this question about file sizes to be answered.

Thank you again, rob.

James