ImageCalibration Question - subtracting biases

[bhwolf]

Background: I'm running a few experiments to determine the effect ambient temp has on the bias of my QSI683 camera.  Despite having regulated cooling, the ambient temp definitely affects the bias pattern.  This is creating vertical banding in the calibrated images.  (There's more here, but keeping it short for brevity.)

As a test, what I'd like to try is having 2 bias masters: one gathered when I captured darks (bias_darks), another when I captured lights/flats (bias_lights).  My thinking is: calibrate the master dark with the bias_darks, enter it into the master dark field of IC.  Do NOT check calibrate.  Enter bias_lights for master bias.  Enter master flat, check calibrate.  Run with lights as the target frames.

Would this be the correct way to run this test?   This way, the dark is calibrated with its bias while the lights/flats are calibrated with their bias.  Not sure if this is making sense.   Thanks for any assistance!

[bhwolf]

I did ultimately get this all figured out -- even though I was troubleshooting a problem with a camera, this really helped me understand and visually confirm what is happening in calibration.  What I ended up doing was creating a number of synthetic light, dark, bias, and flat frames in an image editor.  That way, it became very easy to work with in both PixelMath and IC, and the differences with either visually obvious or easy to compare. 

The test frames are TIFs and too big to attach here, but if anyone wants them for their own tinkering, let me know and happy to upload them.

[mschuster]

On my early build QSI683, the TEC would not hold temp when capturing a long sequence of biases and other very short exposures.

Newer caneras have better firmware.

Adding a 5 second delay between biases solved the problem.

Double check camera temp. It could be drifting without you knowing.

Mike