Unless I'm misunderstanding somewhere a bias frame will have a lot less signal than a light frame though. That being said I tested a light frame of gain and offset 0 and the minimum was only 256.
That's the tricky part to get beyond. You made that decision based on an uncalibrated light frame where in reality the calibration process will subtract the master dark which will cause many pixels to go negative and become clipped. Unfortunately most of that clipping at the brightness level of space. When you try to stretch it in order to eke out faint nebulosity you will rue the day you were born! (How harsh, just kidding.)
When looking at the example histograms there isn't enough resolution to it to see if the left hand side is touching in Fits Liberator.
That is the screen/software struggling to display the detail on the compressed horizontal axis - remember that line is representing 0..65535, so there is not much room for the gap. Look at the MIN value in Image Statistics.
Alnitak-light-frame.jpg (72.27 KiB) Viewed 3879 times
This is from 10 Oct. Agree there is no gap at LHS but Image Statistics Min=1760 which leaves me well clear of 0. Look at the Alnitak result in my Gallery post around 10/11 Oct - its the mega session. I knew from the figures the images would be good even before I started the capture - the figures presented are that reliable.
M31-light-stretched.PNG (185.69 KiB) Viewed 3878 times
This is the same M31 frame stretched with ArcSinh(x) - I use that rather than log(x) as ArcSinh(x) is better behaved around zero. Note the stretch doesn't affect the Image Statistics numbers. The gap is just about there visually now.
Sorry for all the questions
No problem - shows you are thinking by asking all the right questions.
does loading undebayered fits into Fits Liberator skew the results? Would they need to be debayered before loading them in.
No & No. We are dealing with RAW data. Debayering (Demosaic) is sensibly done as one action after master dark & flat have been applied to the light frame.
Yes my bias frame is a lot darker. Thanks guys, I think I'm understanding this a lot better now. I'll have a look at my setting and see if I get better results with the appropriate offset.
Just to clarify - the bias histogram is symmetrical only because of the correct offset adjustment. This can all be tested on the bench on one of those rare cloudy nights.
Spot on. You can be confident that for gain 0 that is the offset to use with that equipment combination. Now experiment with other gain values that you use just to have an idea of the offset settings you might need. Practise this during the day, or when clouded out. That way good sky is not wasted as you are ready to go.
I have a CCTV lens that screws directly into the camera, enabling me to try things during the day without having the clutter of mounts and tripods etc. I have a small tripod about 12" high, this gives the camera stable mounting point. The real learning is done in my study, so a clear night is about imaging not grappling with basics. Another good learning tool is the Test Camera - you could load your own image and try out techniques. Many of the images in the manual were obtained from the Test Cameras.
Thanks for the help guys. I've gone through the gains I'm likely to use and I've noticed the minimum values aren't all the same for a given gain when I capture several bias frames. They can fluctuate by 100-200, especially for higher gains. Is this normal?
One other thing I've noticed is that even with maximum offset I can't get a minimum above 16 when using gain 375:
Untitled3.jpg (330.88 KiB) Viewed 3872 times
The sensor is capped and wrapped up in a dark cloth so I don't think any light is getting in.
They can fluctuate by 100-200, especially for higher gains. Is this normal?
Yes, that is to be expected whenever you look at a single sample of anything in this world.
One other thing I've noticed is that even with maximum offset I can't get a minimum above 16 when using gain 375
Isn't gain 375 the maximum for your camera? Mine is 10000. I would never consider imaging with that setting. I'm not surprised that the maximum offset won't get you higher than 16.