Which Light is the Best?

[timtrice]

Bit of a cross-post from another forum but was hoping for input from some PI guru's as well...

Equipment: 60D unmodded on an AT8RC

I wanted to see what my best setting is as far as reducing noise. I think one of the main areas I've struggled in is not really getting my signal up enough to save it. I've about 20 hours of one galaxy and over 30 of a dark nebula region but no matter what I do I can't really pull out much detail in the faint nebulous regions.

So I ran two subs; one 10-minute of 800 ISO and one 20-minute at 400 ISO. Running PIxinsight's Noise Evaluation script I get the following output:

Code: [Select]

run --execute-mode=auto "C:/Program Files/PixInsight/src/scripts/NoiseEvaluation.js"

Processing script file: C:/Program Files/PixInsight/src/scripts/NoiseEvaluation.js

LIGHT_20141123_21h07m59s697_800_600_02375_01009__26c_14c_001884_RGB_VNG
Calculating noise standard deviation...

* Channel #0
?R = 8.048e-003, N = 10278802 (57.03%), J = 4

* Channel #1
?G = 4.107e-003, N = 13199060 (73.23%), J = 4

* Channel #2
?B = 2.547e-003, N = 4061714 (22.53%), J = 4


run --execute-mode=auto "C:/Program Files/PixInsight/src/scripts/NoiseEvaluation.js"

Processing script file: C:/Program Files/PixInsight/src/scripts/NoiseEvaluation.js

LIGHT_20141123_21h30m10s877_400_1200_02463_01046__26c_14c_001885_RGB_VNG
Calculating noise standard deviation...

* Channel #0
?R = 6.184e-003, N = 9150966 (50.77%), J = 4

* Channel #1
?G = 3.125e-003, N = 12354799 (68.54%), J = 4

* Channel #2
?B = 1.954e-003, N = 3766801 (20.90%), J = 4
So this suggests to me obviously the 20-minute exposure is better where ?B is maybe 20% or 30% lower in each channel. Or am I reading this wrong?

But going back to the dark nebula obviously I need contrast as well. I tried the ContrastBackgroundNoiseRatio script but I was confused and didn't find anything that explained in layman's terms what it meant.

I think obviously 20-minute 400 is best but I've read that the sweet spot for my camera may even be 100 or 200 ISO which means far longer than 20 minutes; not sure if I'd even want to attempt that.

For those of you more familiar with this, can you run a script on the two below and verify I'm interpreting all of this correctly? I feel like though I'm getting better imaging nowadays I'm not using the best settings during capture which makes me have to go beyond what I know during processing.

https://www.dropbox.com/s/feo6ap0p9g3w8dk/LIGHT_20141123_21h07m59s697_800_600_02375_01009__26c_14c_001884_RGB_VNG.fit?dl=0

https://www.dropbox.com/s/8v1bo9j8b3uvd18/LIGHT_20141123_21h30m10s877_400_1200_02463_01046__26c_14c_001885_RGB_VNG.fit?dl=0

What I found interesting was both subs matched in temperature but I think I know why and not sure it'll always work that way. And I know the longer I'm exposed the more likely I am to lose subs for other issues but if I can get 20-minute subs + get better SNR I'll take my chances; I'd rather a cleaner image to process.

[IanL]

You're reading it right, and it is what you would expect. Double the exposure is going to give you a better SNR overall.

Where you are confused is the relationship between ISO (gain by any other name) and sensitivity. There isn't one so your comparison is invalid. Compare two sub's of the same length at different ISOs and it might help. More here:

http://www.blackwaterskies.co.uk/2014/01/do-high-isos-make-dslrs-more-sensitive.html

Changing the ISO will only have a marginal impact on noise , specifically there may be one setting that give slightly lower read noise than others. You need to balance that against quantisation errors (ISO below unity gain will map different sensor voltages to the same DN, usually 200 or 100 on a Canon) or loss of dynamic range (high ISO causing different high voltages to map to the maximum DN, depends on target brightness, exposure length and camera model).

Most people seem to shoot at 400 or 800 but each model is different and 200 or 1600 may be better. More here:

http://www.cloudynights.com/page/articles/cat/fishing-for-photons/profiling-the-long-exposure-performance-of-a-canon-dslr-r2708

You can forget the EXIF temperature on Canons, it is indicative of nothing useful. Basically the temperature sensor monitors one of the processing chips to stop it frying, and has no useful relationship to the sensor or environmental temperatures.