9 replies to this topic

[Jlex]

How should I manage calibration frames for a 30 day (or longer) DSO project with temperature swings +/- 20 degrees Fahrenheit?

 

For example, I'd have to deal with a few nights with temps in the 70's followed by a few nights with temps in the 50's.

 

 

 

 

 

 

 

 

 

 

 

[bignerdguy]

I would suggest making a stored catalog of calibration frames, specifically dark and Bias frames.  You can usually put your camera in a cooler or some other cool place that can mimic the temps you will be seeing to take dark frames as long as the camera CCD is covered.  Dark and Bias frames can be kept and used for quite a while and do not need to be taken every night if you keep copies for later. Flat frames might need to be taken every time (i recommend it for the dust donut deletion if nothing else) but Darks and Bias shouldn't change that much over short term time frames.  for longer, say a year you might want to redo them as camera responses can change over long time frames, but otherwise keeping a catalog of darks and biases will save you time.  In fact make masters of each as this will also save you space. on your storage media.  Be sure you take them for each time length and ISO you plan to use.  Keep in mind that temps varying by 10 or so degrees can be compensated by using dark frame optimization in software though it isn't perfect.  The software will attempt to stretch the image to match the settings of the light image.

[idclimber]

First let's discuss the flats/bias. These need to match for calibration. 

 

Can you keep the camera and scope/lens together such that flats can be reused? This will help simplify as you can then reuses the Master Flat. When you do take flats, take a set of temperature matched bias frames to create a temperature matched Master Bias. 

 

When you save a Master Flat put the temperature in the file name. 

 

Each time you take flats make sure you either take new bias frames for a new Master Bias or use one that matches within a couple degrees. 

 

Now Darks 

 

First is to estimate how much variation you have in your sessions. Does a session start at 70 and finish at 50 or is the variation typically only 5 deg?? There is no reason to have super precise matching darks if the variability is high. 

 

Let's say the variation is about 5 - 10 deg. You could try to capture a set of darks every 5 deg. So 50, 55, 60 etc. I would do this on nights when it is cloudy so I am not wasting time. 

 

Assuming I start a session tonight and the temperature of my subs was between 55 and 65. I would then try to match that with a Master Dark of 60 deg. If I need to take my the next night and it is 68 deg when I do, I match them with a Master Bias at 70 deg if I have one. 

 

When you finish a session create a log and record the temperature. You can always work on any bias frames you need later. 

 

I would get in the habit of doing a test calibration through integration. Do this even if you do not have the correct temperature matched darks/bias. The goal is to ensure the flats are working not that the noise is lowest. If a new dust spot shows up you know you need a new set of flats and can take action to get them. 

[Spaceman 56]

OK. this is getting complex.

 

Lazy old me would just shoot Darks at the highest themperature and use those, assuming the lower temperature darks would have less noise.

 

is my theory Wrong ?

 

could a lower temperature Dark have more noise than a higher temperature Dark ?

[marvyyk]

For darks, I may add it also depends on the DSLR model...

I do not gain in using darks with my stock D5100, but I do see some difference with my D5300 mod.

[idclimber]

OK. this is getting complex.

 

Lazy old me would just shoot Darks at the highest themperature and use those, assuming the lower temperature darks would have less noise.

 

is my theory Wrong ?

 

could a lower temperature Dark have more noise than a higher temperature Dark ?

The closer the darks are in temperature to the lights the better they will work. If there is a ton of variability in temperatures in your Lights there is little reason to be precise. Just get it as close as practical. In your example, a Master Dark in the middle of your range of temperatures would be better than one at the highest temp. How much difference is  sensor dependent. 

 

You could do the entire month with one Master Dark. Or you could do it with 2 or 4... Just depends on how much work you are willing to do.  I personally would simply add darks as I had time on cloudy nights or other times I can not image. Creating a Master only takes a few minutes, and I don't keep the originals just the masters. 

 

The main thing to understand is your goal is to match the Darks to the Lights and then separately the Bias to the Flats. 

Edited by idclimber, 11 September 2024 - 11:39 PM.

[sharkmelley]

DSLRs don't have set-point cooling and so the sensor temperature will vary during an imaging session and from night to night.  There is a technique called dark optimization which generally works very well with DSLRs. This takes your master dark and scales it to match each light frame individually.  The clever thing is that it doesn't even need to know the sensor temperature because it's a mathematical algorithm that matches the amplitude of the pattern noise in each light frame to the pattern noise in the master dark.

 

This technique is implemented in many processing packages: DeepSkyStacker, PixInsight and Siril call it "Dark Optimization" while AstroPixelProcessor calls it "Dark Frame Scaling".

 

If you choose to use this approach then take your darks when the temperature is highest because the algorithms often only scale the darks down and not upwards.  For instance, living in the UK I take my darks at room temperature because my shooting conditions outside will always be cooler.

[Andros246]

OK. this is getting complex.

Lazy old me would just shoot Darks at the highest themperature and use those, assuming the lower temperature darks would have less noise.

is my theory Wrong ?

could a lower temperature Dark have more noise than a higher temperature Dark ?

this is what I would’ve done

[sharkmelley]

Lazy old me would just shoot Darks at the highest themperature and use those, assuming the lower temperature darks would have less noise.

 

is my theory Wrong ?

Yes, your theory is wrong. 

 

As you rightly say, the amplitude of the fixed pattern in the high temperature darks will be greater than the amplitude of the fixed pattern in the low temperature darks - in other words they are noisier.  However, if you subtract the high temperature master dark from your low temperature lights then the hot and warm pixels in your lights will have too much subtracted and this will create black spots in your calibrated lights.

[Andros246]

In my honest opinion you’ll be trying to chase your tail with a 30 day worth integration trying to match darks to lights till you’re blue in the face.

Unless you are really wanting to go blue in the face I generally think using the same darks for all the sessions is the best course of action. You could easily spend 24 hours just taking darks if trying to match temperatures to the best of your ability.

If you want to try to match them it’s time to buy a cooled camera or don’t do such a long integration.

Going to another point that long of an integration has such high diminishing returns you could just save 15 days of shooting if you bought a camera with a cooler and you would also save yourself from turning blue. I believe that’s the difference with a cooler it’s about 1.5x better SNR. Which would equate to 15 days instead of 30 days of integration.

 

Btw 1.5x SNR is hardly anything at all.

Edited by Andros246, 12 September 2024 - 10:41 AM.