New to Startools - can you do better with this 53 hour O3 master light?

Questions and answers about processing in StarTools and how to accomplish certain tasks.
User avatar
ramdom
Posts: 13
Joined: Mon Sep 07, 2020 8:56 am
Location: Youngstown, NY USA
Contact:

Re: New to Startools - can you do better with this 53 hour O3 master light?

Post by ramdom »

Thanks Ivo. I have observed what you've observed - that the two thermal noise patterns match perfectly and the subtraction does appear to remove the thermal noise pattern but then when I do the full integration and stretch, the pattern/banding reappears and is hardest to deal with in the faintest targets. Your explanation for why that is would seems to fit with my observations but it is not the gain. It's something however that produces a similar effect (i.e., for whatever reasons, the bright parts may be throttled relative to this thermal noise and it is most problematic in this case of a faint target).

The gain I'm using is 200, and the offset is 32, whereas the manufacturer's recommended gain setting for this camera is 120 and offset is 70 (https://www.qhyccd.com/index.php?m=cont ... =31&id=199). I can't imagine the offset of 70 vs. 32 is the reason for the difference but that is one thing I've not tried. I've tried gains of 300 and 350 and then someone said my stars are oversaturated so I cut it back. The gain of 200 for NB imaging with this camera is what other people appear to use. It is definitely set above unity gain unless the hardware or software is making a mistake.

It's true in this case it's a faint target and so it is a confounder but if we go to a brighter image, like this one which is only a couple of hours or so of integration:

http://ram.org/images/space/downloads/s ... v0_tmp.jpg

You can make out the bands but it is less prominent (this is a soft stretch, just so I can see the framing).

Still though from a naive perspective it is weird to me why the preprocessing workflow in PI doesn't remove it out entirely, as well as the amp glow in the corners. It does do it partially but not entirely. Perhaps I should be using some exotic settings but what I'm doing is something very simple: it's just a straight forward subtraction. I don't check the "dark frame calibration", etc. But it seems to be amplified along with the light signal in a weird fashion. I mean even in your tutorials and guides you have sections on dealing with amp glow - if we had great calibration, why would we even deal with amp glow? But yet with this camera I find leaves me with amp glow esp. with the 3nm filter masters after doing the best calibration I can achieve. I think the two problems may be related and I've dealt with the amp glow in the end also by masking.

One thing I've been thinking of is that my master darks exposure total at 40 frames may be too short for these long integrations I'm doing (632 light frames vs. master dark frames) but again it really seems like a stretch.

I think my first step though is to get away from using this dataset to play with ST and go back to a cleaner/easier data set or wait until I get one. But as far as this particular problem, it is not present in my OSC so that's another option - that when I do my OSC image I can try out ST for that.

--Ram
Tubes: C925, SV70T, UC18, Tak FC100DF, FS128. Mounts: AVX; Paramount MyT. Glass: SV 0.8x, Tak 0.7x, 0.66x FR/FF. Cameras: QHY163M; QHY247C. Filters: Astrodon 5nm Ha, 3nm O3, S2. http://ram.org/ramblings/cosmos/equipment.html for a full list.
Post Reply