Page 1 of 1

Deconvolution mask question

Posted: Mon May 16, 2022 2:36 pm
by dx_ron
I generally have good luck with the 'extra-sensitive' automask. But I am curious what I should do about touching-up a couple of spots.
apodization_mask.jpg
apodization_mask.jpg (395.43 KiB) Viewed 1513 times
For the very bright star, I normally expand the mask to include all of the brightest main circle. Is that necessary (or even incorrect)?

And peculiar to globs - should I leave the masking in the core, or should I only mask stars where they are more individually distinct?

Or maybe neither tweak makes much impact on the process.

Re: Deconvolution mask question

Posted: Tue May 17, 2022 4:24 am
by Mike in Rancho
AFAIK the 1.8 mask is just for SVD sample selection, and also aids with specific deringing. Deconv will be applied across the entire image based on the variance you have set it up to be applied.

So, touching up the big star (or any others) to more correctly match the stellar profile would be useful if you want to use that star in a blue box, or if it needs a little help if the deringing doesn't seem to be working the way it should.

I think 1.8 is actually much better for globs than the 1.7 version because of that. I would get some smaller glob stars all nice and sharp, but brighter neighbors (masked for protection against artifacts) would still be all fuzzy like they were unfocused!

Re: Deconvolution mask question

Posted: Tue May 17, 2022 5:11 pm
by dx_ron
Thanks. That makes sense. I do sometimes use bright stars for sampling - but not that one. It shows up in the Select screen as a green donut not a circle. It is fun to use decon on globs. Before vs after is like out of vs in focus.