Telescope Reviews: Question about Darks with an Uncooled Camera

Ideally, you want to match the temp, but have a lot more total exposure time in the darks than in your lights.

It doesn't really make sense the say the "tail end of the darks has 5 hours of thermal noise". Each frame has its own exposure length of thermal signal based on its temperature.

I believe you may be making a common misconception confusing signal and noise.

Thermal signal is not noise. So, with 5 hours of darks you don't have 5 hours of noise. You have an average of the thermal signal for the length of each sub exposure, and you have the square root of that in noise. You actually have less noise in 5 hours worth of exposures than you do in 3 hours worth.

Using a dark frame removes thermal signal and actually adds the posisson noise associated with that signal. It is critical to understand this.

Every signal has noise associated with it. It's called photon noise, or shot noise, or poisson noise (they're all the same thing).

Shooting lights and darks are doing exactly the same thing... they are recording signal and its associated noise.

In the case of lights, they are recording object signal, sky signal, and thermal signal. Each has noise associated with their respective signals. There are also other sources of true noise involved, such as readout noise, but I'm not going into that here.

Darks are only recording thermal signal and its associated noise.

The whole purpose of stacking lights and darks is to increase the signal to noise ratio between the signal and its associated noise. You do this by averaging a bunch of frames. In an average, the signal stays the same, but the noise is then reduced by the square root of the total signal.

For example, you shoot 16 x 5min of light, or dark frames.

In a single frame you get x amount of signal and the square root of that amount of noise.

In the average stack of 16 frames, you get the same x amount of signal but only 1/4 the amount of noise (the square root of 16 frames).

If you just subtracted a single light frame, you would be removing the thermal signal, but you would be adding 4x as much true noise to the image compared to using the master dark made of 16 subs.

When you subtract a dark to remove the thermal signal from the light, you are actually adding the noise from the dark to the light. That actually increases the true noise in the light frame after calibration.

By shooting a lot of darks to make a really good dark frame, you reduce the true noise in the master dark, so less is added when the dark is subtracted from the light.

This is why you want to shoot a lot of dark sub frames to create a really good master dark.

6 hours of darks is going to produce a higher signal to noise ratio than 3 hours of darks. Less noise will be added with 6 hours of darks compared to 3 hours of darks.

Now, everything gets complicated by the fact that the internal camera temp is increasing for a couple of hours until it stabilizes, and this is different for each camera, and by the fact that, in your case, the ambient temperature is dropping by a substantial amount at the same time.

I guess you could do a complicated mathematical analysis by actually measuring each frame and then figure out which is the best way to put them together, but that is beyond my skill level, and I just don't worry about it because I have always had excellent results by using really good master darks and master biases with Images Plus automatic dark frame matching.

I don't believe DSS is using the same methodology as IP, so I can not advise you as to what is the best way to proceed in DSS.

Jerry