Darks, bias, flats and lights. Informtion needed for DSLR

[Cumbrianwolf]

I understand that I need to do the above but from what I am seeing and reading says I will need to do them all again for each image I take as the exposure and timings change as in M31 will require different settings than M45 for example. So can I not just allow the camera to cool down to ambient temperature, then take say 75+ of each frame at the same ISO but with a slightly longer exposure time than what I would normally use for the light frame as to be honest I do not know what that will be until I am taking the shot and analysing the histogram? Ideally, I would like to build up a set of masters that I can use for subsequent image processing thereafter?

[TerryMcK]

The problem with DSLR cameras is you have zero control over the temperature of the sensor. The temperature of the sensor is dictated by the ambient temperature around the camera. However the sensor itself does increase in temperature especially with long exposures. I remember seeing temperatures of greater than 30 Celsius on one of my old Canons even though the outside temperature was around zero.

You have to do the darks either immediately before imaging or just after imaging lights so that the temperature of the sensor is roughly the same. You have to make sure the ISO and exposure time is exactly the same.

Bias and flats can be done at any time and are not affected by temperature.

 

Cooled dedicated astro cameras don't have that problem as the temperature can be controlled very well.

 

I think you can get away with 20 to 30 darks and the same for flats with a DSLR.

[Cumbrianwolf]

11 minutes ago, TerryMcK said:

The problem with DSLR cameras is you have zero control over the temperature of the sensor. The temperature of the sensor is dictated by the ambient temperature around the camera. However the sensor itself does increase in temperature especially with long exposures. I remember seeing temperatures of greater than 30 Celsius on one of my old Canons even though the outside temperature was around zero.

You have to do the darks either immediately before imaging or just after imaging lights so that the temperature of the sensor is roughly the same. You have to make sure the ISO and exposure time is exactly the same.

Bias and flats can be done at any time and are not affected by temperature.

 

Cooled dedicated astro cameras don't have that problem as the temperature can be controlled very well.

 

I think you can get away with 20 to 30 darks and the same for flats with a DSLR.

Oh, well I hoped I did not need to do all that work, but so be it if that is what is needed. Tonight I hope is the night for some of my first digital images.

[TerryMcK]

I think as the ambient temperature is fairly low today try getting a few dark frames done in advance. Hopefully there is no light leakage in your rig John. Maybe throw a cover over it just to be on the safe side. Good luck on tonight's captures hopefully it'll be a good one for you.

[paul]

Sadly true, internal temperatures are significantly higher than ambient and for every degree change in ambient you eventually get a degree change in the internal temperature which changes the noise.  The sensors are similar to those used in dedicated astro-cameras but are operating at something like 40+ degrees above a cooled camera sensor. This means thermal noise is significant in the dark frames.

 

Yet knowing all this I don't always take dark frames per session, particularly if the operating conditions are similar on multiple nights. The ambient can vary by several degrees through the night so you are never going to be perfectly matched anyway. 

 

After a while you may choose to build up a dark frame collection good enough to reuse for future sessions. That would be my strategy if the DSLR was my main camera now. 

[Dmack1]

Your darks need to be matched to yourlights in terms of exposure time, ISO and teperature.

 

On both my canons, 1100D and 700D a temperature is displayed in both my software (APT) and in the data held within the image (I've set it up to include it in the file name also).

 

Now while that may not be considered an accurate "sensor temperature" I have taken it as a reasonable indication because if the night is very cold, it will start low and then quickly level out  once images are being taken and the sensor is warmed up. What i find it that on normal winter nights between say -1 and 6c (not summer when ambient is much higher) that temperature nearly always settles at around 15-17c. Sometimes up or down a little but not much.

 

My thinking is that once the indicated temperature of the darks and lights is the same then the sensor temperature should also be much the same, even it the actual sensor temperature is not what is actually shown.

 

So what I have done is to set up a cold box out in my shed with a wireless temperature monitor. I put some ice packs in it and then  it will keep a fairly stable temp of around 3-5 degrees for 6-8 hours.  I then put the camera in, set up a sequence of dark frames at my common capture settings (iso 800, 60s, 120s, 180s) and set it to run. The camera very quickly settles at that 15-17c and I have a dark frame library I can use for most of my winter images. Both APP and PI allow for dark frame scaling which allows the temperatures of lights and darks to be different .

 

The only problem I have with dark scaling is that the 700D has some amp glow and scaling will not work well. But I find that even without scaling if the indicated darks temp and indicated lights temp are within  say +/- 4 c the calibration has a noticeable and positive effect.

 

Some people will say that its better to do no darks with a DSLR because of the potential temperature differences but thats not been my experience when I've done a darks library this way.

 

On the other calibration frames flats and dark-flats need to be matched to each other on exposure and ISO but do not need to be matched to the lights. I tend to do flats and dark flats together in the morning after every imaging session but they can last for some time (i know some people reuse for several weeks) so long as you have not taken apart the imaging train or changes the camera rotation angle.

 

Bias needs to be matched to the lights only in ISO

Edited January 24, 2022 by Dmack1