The noise produced by a Canon 1100D at various ISO settings and temperatures

[Ags]

My Canon 1000D has fairly awful non-random read-noise structure at ISO800 which is not present at ISO1600. I then read somewhere that Canon use different circuitry on the camera to generate ISO1600 than for their lower ISOs - so this might explain the differing noise behaviour.

NigelM

I think davew's link has a more likely explanation - one set of circuitry, but different mathematical hacks on the raw data at different ISOs.

[boardhead]

oddly the biggest problem I had was renaming the files (I kept making stupid mistakes and scrambling the results).

Did you know that exiftool can rename the files for you?:

exiftool "-filename<%f_${ISO}_${Cameratemperature}.%e" -n DIR

- Phil

[Ags]

Yep, I saw that - but i had to read the RAW files and rename the corresponding TIFF files. It should have been easy, what could go wrong :-))

[Gina]

APT names the files for you in the format yyyy-mm-dd-hh-mm-ss _ ISOxxxx_yyys__TT ie. date & time, ISO, secs exposure, EXIF T. THough this can be changed,

[boardhead]

Yep, I saw that - but i had to read the RAW files and rename the corresponding TIFF files. It should have been easy, what could go wrong :-))

Just add this before the "-filename<..." argument in the exiftool command: -ext tiff -tagsfromfile %d%f.cr2

[Ags]

Ah! That will be much easier... :-)

[Gina]

Have set up APT to take a series of dark subs as follows :- (secs ISO) pause of 2 count of 4

1. . 60. .100
   
2. 120. .100
   
3. 240. .100
   
4. 480. .100
   
5. . 60. .200
   
6. 120. .200
   
7. 240. .200
   
8. 480. .200
   
9. . 60. .400
   
10. 120. .400
    
11. 240. .400
    
12. 480. .400
    
13. . 60. .800
    
14. 120. .800
    
15. 240. .800
    
16. 480. .800
    
17. . 60.1600
    
18. 120.1600
    
19. 240.1600
    
20. 480.1600
    
21. . 60.3200
    
22. 120.3200
    
23. 240.3200
    
24. 480.3200
    
25. . 60.6400
    
26. 120.6400
    
27. 240.6400
    
28. 480.6400
    

[Ags]

Great! What temperature is the cooling set to?

[Gina]

1. The above test dark plan will take about 7 hours according to APT - presumably allowing for pauses and time to read the sensor data.
   
2. This plan will be run for a set-point temperature of :- -10, -8, -6, -4, -2, 0, 2, 4, 6, 8, 10, 12 Celsius.
   
3. Unless I find and fix the light leak into my camera box this will be one run per night, making 12 nights. Fortunately this doesn't need clear nights or even fine nights
   

[Ags]

On the contrary - it needs horrible nights where there is no hope of doing real imaging :-)

It will be interesting to feed this data into the app. My guess is that imaging at very high ISOs will gain from very low temperatures.

[Gina]

On the contrary - it needs horrible nights where there is no hope of doing real imaging :-)

Absolutely, and we have plenty of those It's good to have the equipment actually doing something useful

It will be interesting to feed this data into the app. My guess is that imaging at very high ISOs will gain from very low temperatures.

This is what I would expect but when I cooled to -10C with a real image run the noise was still noticeable I have a nasty suspicion that Canon do some adjustments to the data before outputing it to the RAW CR2 file. Either that or the electronics/sensor don't like -10C. I would like to do some signal tests at various temperatures. That will need some thinking about.

I've started tonight's run at 0C for no other reason than that was what I had the set-point temperature set to. I've added some 30s subs to the beginning of the darks run to make sure the EXIF T is running at the desired value. The set-point temperature was designed to be constant to within a few tenths of a degree.

[Gina]

Hi Agnes, I was thinking I may do +10C tomorrow night or I can do any other temperature. I can do runs at whichever temperatures are most useful for your analysis first, and leave less important ones until later.

[Ags]

I think do the extremes of the range, and then work back to fill in the gaps - that way if something goes wrong halfway through the testing, we still have a complete range with a few gaps.

[Gina]

I think do the extremes of the range, and then work back to fill in the gaps - that way if something goes wrong halfway through the testing, we still have a complete range with a few gaps.

Yes, that's what I thought would be best. If we get very cold weather it might not be possible to get the higher temperatures so I'll do +12 tomorrow, then -10, +6, -6, +8, etc. Or something like that.

[Gina]

Darks run as above at 0C EXIF T completed overnight. Ready to do the run again at 12C tonight unless the clouds miraculously clear and I can manage an imaging run

[Gina]

12C run now well under way

[Ags]

Your chip is gonna be sooo analyzed!

[Gina]

Yup! Last night saw a mainly 12C run - a small number of 13s but that's as near as I could get. The reading only has a 1 degree resolution any way. Tonight I plan to do a -10C run.

During the day yesterday I didn't turn it off and when I went to set up last night I found it had picked up a glitch and the MODE was at MAX and temperature was -20C Took quite a while before it warmed up enough for the run I was tempted to do the -10C run but stuck to the schedule

[themos]

Agnes, why don't you follow Craig Stark's method of quantifying noise and dynamic range?

http://www.cloudynights.com/category.php?category_id=181

My tests with my 450D have convinced me to always use ISO 400 from my suburban garden.

[Ags]

Interesting he uses ImageJ - I am planning on moving to that library eventually... There is a lot of good information there which will take a while to sink in, but one thing I did notice that has to be wrong was the advice to cool the camera for fifteen minutes between the dark exposures. Especially when shooting a 10 minute dark, this will result in the exact opposite of stable temperatures. It will lead to maximum thermal instability (with only matching starting temperatures). What is instead required is to have the chip at thermal equilibrium, which means continuous shooting for an extended time at a realistic ambient temperature (certainly not room temperature).

[Gina]

I've started reading that article - I'll carry on later - it's quite heavy going. I agree about the cooling down business - you want a constant temperature. And, of course, as near as possible to the conditions used for the lights.

BTW. My -10C darks run is currebtly well under way.

[Gina]

Completed the -10C run overnight and saved it. Tonight I had a short imaging run but then it clouded up and rained a bit. Since then it's cleared and clouded up and cleared and clouded up again etc. so I've given up and set up the next darks test run at +6C.

[Chris]

Thumbs up Gina:), your always hard at work in the name of Astronomy and tinkering:)

[Gina]

Thanks Chris

[Ags]

Thumbs up Gina:), your always hard at work in the name of Astronomy and tinkering:)

+1!