In the interest of shamelessly copying Marc, here's an M45 from Astrofest
I've been wanting to do this for a long time but I can't see it from my house and last year's Astrofest was too early in the year. None the less long nights paid off.
So here we are, shot on QHY10, Meade ED80, 3h 25min worth at 5min subs.
Side note: 5min subs were too long for this target. I severely clipped the center of the stars which for the first time has played havoc with my processing. But I learnt something interesting: convolution as a tool to make things blurry has a point, and there's a pun in there somewhere
Wow, that's a very good M45 Chris, nicely processed excellent colour and contrast and the slightly tilted framing works well too, nice job indeed. I don't mind the slightly blurred bright stars actually, a bit Akira Fuji like
Looks pretty good apart from the blurry stars, Chris
I blame that on the lack of deconvolution. Because of the clipping issues I wasn't able to deconvolve. I skipped the step and it wasn't until later in the processing step that I realised I had issued in other parts of the processing and had to fix the clipping.
I blame that on the lack of deconvolution. Because of the clipping issues I wasn't able to deconvolve. I skipped the step and it wasn't until later in the processing step that I realised I had issued in other parts of the processing and had to fix the clipping.
Maybe one day I'll go back and do it again.
I was taking the piss just a little, Chris I presume it was just the star cores that needed fuzzing up. Decon won't have much value when you're undersampled.
I was taking the piss just a little, Chris I presume it was just the star cores that needed fuzzing up. Decon won't have much value when you're undersampled.
Cheers,
Rick.
Oh. No this image is legitimately looks slightly less sharp than some of my others. In the linear stage I normally apply deconvolution that tightens the stars up a bit. I thought you were referring to the fact that it looks a bit flat at 100%, this time it's not because of overcooking the erosion. Yeah I know I have a different excuse every time
Delicious colour balance and saturation. Very nice contrast in the wispy detail. A lovely image.
(Aside to anyone who cares: There is no mathematical reason not to deconvolve when some of the stars are burned out. (I think I understand the maths well enough). You just create your point spread function from the brightest of the ones that aren't burned out. You'll need an anti-ringing algorithm, but that's pretty standard. In my view, the purpose of deconvolution for our kind of deep sky photos is not to make the stars look tiny but to get better detail in shock fronts, dust lanes, etc in the nebulosity.)
(Aside to anyone who cares: There is no mathematical reason not to deconvolve when some of the stars are burned out. (I think I understand the maths well enough). You just create your point spread function from the brightest of the ones that aren't burned out. You'll need an anti-ringing algorithm, but that's pretty standard. In my view, the purpose of deconvolution for our kind of deep sky photos is not to make the stars look tiny but to get better detail in shock fronts, dust lanes, etc in the nebulosity.)
Mike,
Best practice for decon in PI is to estimate the PSF from a selection of unsaturated stars so I don't see some burnt out stars as a problem either.
However, I've never had much luck with decon on wider fields/larger image scales. These images tend to be fairly sharp anyway and I assume that undersampling leads to a poor PSF estimate. Perhaps I've been spoiled by good gear and nice data and there is a place for decon with blurry wide field images?
Delicious colour balance and saturation. Very nice contrast in the wispy detail. A lovely image.
(Aside to anyone who cares: There is no mathematical reason not to deconvolve when some of the stars are burned out. (I think I understand the maths well enough). You just create your point spread function from the brightest of the ones that aren't burned out. You'll need an anti-ringing algorithm, but that's pretty standard. In my view, the purpose of deconvolution for our kind of deep sky photos is not to make the stars look tiny but to get better detail in shock fronts, dust lanes, etc in the nebulosity.)
In theory it should work, in practice however the anti-ringing is the single biggest problem and in my case I could find no combination of dark / bright compensation, using a star support, or even masking brighter sections of the whole image which didn't introduce some kind of strange result in this case. But then I'm by a long shot not an expert, and my computer is slow so I got frustrated after an hour and moved on.
...
Only to hit the same problem when I got to masked stretching
That's a beautiful M45 Chris. I here you about Decon mate. I rarely use it any more, I cannot get it to work without introducing artifacts. I found ATWT does a great job of sharpening when used with a light hand, so I use that instead.
