I'm shooting almost exclusively with a 12-bit ASI183MM. When I get a frame off the camera, I see a histogram peak that makes sense, between zero and 4096.
But the max value in the Statistics pane is always 65504. Furthermore, there are a few points scattered along the histogram well in excess of the max ADU that the camera can put out.
I wouldn't care about a few random stray values, but I'm having a lot of trouble getting any significant color out of my last RGB outing and wondered if overexposure was the cause. According to the stats, the skyfog peak (median value) is 27% of full scale, so that looks fine. But it's the higher values that seem wonky. You can't see the cursor in the second screenshot, but it was hovering over a star. Note the value on the stats line at the bottom: 65504. Well, THERE'S your problem, right? Looks pretty blown-out to me, but HOW DID IT GET THERE?!! Should have maxed out at 4096, no?
Hmm. Something ate the first screenshot with the actual histogram. Here it is. If I hover over X-axis values in excess of 4096, I get some results, e.g. 280 samples around 6000, etc..
And even more confusingly, here's a flat that I shot that same night. Wow, I didn't know you could get a value of 40,000-plus from a 12-bit ADC! So now I'm really confused.
It's true the camera is 12 bits, but the reporting is often 16bits depending on SW parameters. All your values are raised by 4 bits.... I've just gotten used to seeing it one way in the FITS viewer, and another in other post-processing SW.... (I have an ASI183mc Pro)
Of course. <slaps forehead> Which means that my flat is a bit overexposed -- probably ought to be more in the 25K range -- but that my light has a skyfog peak of more like 2 percent full-scale instead of 27%. Hmm. So the lack of color satch is due to something else; 30 minutes' integration time on each of the color channels ought to bring out something.
OK, here's the FITS of the RGB-combined image as it left Astro Pixel Processor:
M81-82 FITS
Attached is a slightly touched-up JPEG of the same (little bit more stretching, turned down the excess blue a bit).
I figured out the main problem, I think: I also have a new filter wheel, and when I went from 5 slots to 8 I didn't map the slots correctly. So the "Blue" channel data actually had no filter at all, and the others were also wrong (probably no R at all, "R" was really green, "G" really blue). I did kinda wonder why the putatively "Blue" channel was several times brighter than the rest...sigh. Embarrassing, but that's my astrophotographical life, mostly. Also I did my sums wrong when I redid the optical train with the new filter wheel and had the wrong backfocus, which is probably why all the stars have such terrible coma. Oh, and the focus sucks too. All in all, a really stunning achievement.