Is it just my imaging circle too small or are my calibration frames bad

[ianaiken]

I'm trying all kinds of calibration tests and cannot seem to get a correct master light. This is the image with just flats and no darks on either flats or lights. It does get a bit better with both flatdarks and darks. With darks applied the centre is a bit more balanced, there is a noise type ring where the black ring is here and the outer edges are over corrected but not as much (as seen when using ABE with default settings. I've tried flats ranging with a median from 4k to 20k In fact anything above 18k approx yields a flat with a maximum of 65k (saturation) with a minimum of about 3.5k. Bias min is around 2.2k.

Now, you may say why am I imaging with a full frame sensor on a Hyperstar 3 expecting a 27mm imaging circle to give a good image on a 43mm sensor. Well the star shape is reasonably good to the edges and I wanted to see if it works.

What do you think?

Trying a bit DBE for what it was worth on a dark calibrated master light

 

[ianaiken]

Here is a log file. Looks like calibration was performed correctly.

 

[pfile]

This is the image with just flats and no darks on either flats or lights. It does get a bit better with both flatdarks and darks.

at the very least you need to calibate flats with a master bias or a master flat dark. the flats won't work right otherwise.

you might have to crop out everything that's way outside of the properly illuminated area before calibrating, which is a real chore to do right. the flats in those areas might be so wrong that calibration is impossible.

rob

 

[ianaiken]

at the very least you need to calibate flats with a master bias or a master flat dark. the flats won't work right otherwise.

Hi Rob. Yes, I've tried with darks, bias, and darks and bias. THe result is very similar each time. Maybe too strong a gradient. Not helped by using 2 inch filters. I found some threads with similar problems so trying out a few things.

Cropping out everything doesn't sound found I can see my QHY163M being mounted soon if I don't get this resolved This camera is a QHY367C.

Ian

 

[pfile]

i saw something similar recently posted here from a pretty experienced user (i think it was also hyperstar); i tried to do DBE and just couldn't solve it either. i think if the gradient is too wild the image is impossible to flatten.

with the cropping, since it's an OSC you have to take care to crop the right number of pixels off the top and bottom such that you don't change the bayer pattern. well, i guess even if you did change the bayer pattern it would at least be the same in all cropped images but there'd possibly be some wasted work to discover what the new matrix should be. you can use ImageContainer and Crop to do this...

rob

 

[ianaiken]

i saw something similar recently posted here from a pretty experienced user (i think it was also hyperstar); i tried to do DBE and just couldn't solve it either. i think if the gradient is too wild the image is impossible to flatten.

with the cropping, since it's an OSC you have to take care to crop the right number of pixels off the top and bottom such that you don't change the bayer pattern. well, i guess even if you did change the bayer pattern it would at least be the same in all cropped images but there'd possibly be some wasted work to discover what the new matrix should be. you can use ImageContainer and Crop to do this...

rob

I'm reading now, the other thing people seem to have done with this camera is double the gain and for some this appears to have solved the issues. Of course, this means new data, but worth a try.

I may try ImageContainer too not used that process before so will look into it. Appreciate the help! Maybe this is all justification for RASA11...