Darks without setpoint cooling?

[Earl]

Is it viable without setpoint cooling to make darks? I notice a piece of software I use has the option, but the gopro thermals are unpredictable. 

[vlaiv]

That really depends on hardware used and calibration performed.

Ideally - you want hardware with very stable bias that does not use its own internal dark subtraction (like modern DSLR cameras do) and you want to perform dark scaling.

With "well behaving" hardware above should work and give you results that are close to temperature matched darks.

Best way to check if your hardware is going to be well behaved is to test it.

To establish bias stability do the following:

- take set of bias subs, then turn off power to the sensor, power it back on and do another set of bias subs. Measure mean and standard deviation of all subs. They should look the same between subs in each batch and between batches.

- take set of bias subs and set of darks with increasing exposure length. Stack each to their respective stack and measure mean value of each. Plot in spreadsheet - they should form as close to linear function as possible (there will be some distortion because of temperature change as sensor heats up and cools down)

if two above behave well - it should be possible to do dark scaling as part of calibration process. That is algorithm that will automatically "adjust" for different temperature (will select suitable multiplicative constant based on pixel statistics).