Should the dark temperature match the light temperature?

[han59]

In this post I tried to answer three questions around the calibration of lights:

 

1) Are darks required?

2) Should the dark temperature match the light temperature?

3) If not can the dark be scaled using a prediction of the dark current?

 

The experiments where done with dark series of an ASI1600MM camera. To get a measurable result two series of stacked darks where subtracted from each other. One represents the lights, the other the darks:

 

First some visual tests:

 

First image, no darks gives a pretty high noise value. The second image based on ∑(12x60sec_+26°C) -  ∑(12x60sec_23°C) gives the lowest noise value. So conclusion is that 1) darks help and 2) darks and lights temperature should match.

 

Next a more extensive test presented in a table:

You can see that the noise value for  ∑(12x60sec_-1°C) -  ∑(12x60sec_-1°C) is as good as ∑(12x60sec_-10°C).  . So conclusion is that 1) darks help and 2) darks and lights temperature should match for lowest noise value.

 

 

Next question 3) . Can darks be scaled? I tried to find the best X factor in subtraction of two dark series. E.g : ∑(12x60sec_+5°C) - X * ∑(100x60sec_0°C)  The experiment was done in a custom adapted version of ASTAP which varied X in small steps to find the lowest noise value:

 

Conclusions:
1) Make at least 50 or more darks to reduce the dark noise. See part 1 and 2 (pdf file).
2) It is possible to correct for a wrong dark temperature. Significant improvement above +10°C. Below +5°C the effect is limited
3) A wrong dark exposure time has only some influence above +10°C. See part 5 (pdf file)

 

Full test report:

 

 

Han

dark_test2.pdf

Edited October 3, 2023 by han59

[knobby]

Thanks Han, very interesting read.

[MalcolmP]

10 hours ago, han59 said:

1) Are darks required?
2) Should the dark temperature match the light temperature?
3) If not can the dark be scaled using a prediction of the dark current?
To get a measurable result two series of stacked darks where subtracted from each other.

Very interesting, thank you Han.
What do you use to form the two stacks, I think maybe make two master darks in ASTAP then treat one as a light ? 
What do you then use to do the subtraction ?
I have put a CR2 dark frame in the Lights tab and another CR2 dark in the Darks tab and tried to Stack, not surprisingly it failed  ( I had expected that several darks in the lights tab would fail on account of no stars to align on, but perhaps with only one frame it would not try to stack lights, just subtract the dark, but it was more AI than me and rejected my little subterfuge   )

I have asked myself the same three questions about my DSLR and the problems of controlling the temperature (and knowing what sensor temp. the temp. of the camera represents !) and posts on the forum suggest that I should not worry and only use bias and no darks with my Canon 60D, but it would be nice to try the experiment on it.
 

Edited October 2, 2023 by MalcolmP

[han59]

Quote

What do you use to form the two stacks, I think maybe make two master darks in ASTAP then treat one as a light ?

Yes just make twice a master dark of 12 darks.  So you need then 24 darks.

 

Quote

What do you then use to do the subtraction ?

In ASTAP, tab Pixelmath2 there is an option Apply file to image in the viewer.  Select a second file and option "subtract file view + 1000". The 1000 is to keep positive values.

 

Quote

I have asked myself the same three questions about my DSLR and the problems of controlling the temperature

Only for the latest Sony sensors with cooling you could consider to skip darks. But with darks the image will always be cleaner & hot pixels are removed. Only with a dark applied the flat correction will be optimal. So I still would go for full calibration with darks, flats and flat-darks.

I think it good to keep a dark library of different temperatures. Old darks from year ago will still work fine.  In ASTAP the correct dark with an corresponding temperature will be selected automatically. So you could keep them all in tab darks.

 

Han

 

[MalcolmP]

4 hours ago, han59 said:

tab Pixelmath2 there is an option Apply file to image in the viewer. , , ,

the correct dark with an corresponding temperature will be selected automatically.

Thanks very much for the guidance Han,
 and it is good to know that ASTAP can sort the matching temperatures for me !

 

[sharkmelley]

I was a bit confused by your expression:

corrected dark:=dark * factor - mean_dark * (1-factor) 

Now I've worked it out

Edited October 4, 2023 by sharkmelley