what length dark when using different exposures

[iwols]

hi all what length exposure of darks do you use when your imaging exposures differ,ie when i integrate say a series of 10 min subs with 5 min subs what would you use for the darks, cheers

[Cornelius Varley]

The dark exposure length should match the light exposure length. 10mins lights = 10mins darks etc.

[iwols]

thanks peter but when you integrate a 10 min sub and a 5 min sub together what would you use ,cheers

[andrew s]

If you take bias frames you can then used scaled darks taken at the longest exposure.

Regards Andrew 

[wimvb]

33 minutes ago, andrew s said:

If you take bias frames you can then used scaled darks taken at the longest exposure.

Regards Andrew 

For completeness: this depends on the camera used. For a camera that exhibits amp glow (cooled cmos), the darks need to match the lights, and scaling generally won't work. As @Cornelius Varley wrote, each exposure time will need its own darks.

[Anthonyexmouth]

Wouldn't you do them as separate stacks with their own darks then combine them? 

[tomato]

I calibrate the lights with the appropriate flats and darks before integrating them, this is easy in APP as you can enter the data into individual sessions before registration and stacking, but I think DSS can do something similar.

[vlaiv]

Actually I would use shorter exposure darks and dark scaling.

This would only work for cameras with stable/usable bias files - which means basically CCD cameras as most CMOS sensors don't have stable / usable bias files.

It might seem counter intuitive to use shorter darks, but let me explain. Let's suppose that you have 5 minute subs and 10 minute subs, and let's suppose you have 4h dedicated time for taking darks. With most modern CCD cameras read noise is far greater than dark current noise. Difference between short and long sub dark current noise, or rahter dark current noise increase is by factor of ~1.41 (square root of 2). Also, when you stack double number of subs (given that we have limited dark budget time - you will get twice as many half duration subs), read noise decreases by factor of 1.41.

So it's the matter of which one is bigger - and with modern CCD sensors, it is the read noise that dominates in these exposures (order of 6-7e per exposure, while dark current is roughly 0.0001 e/s/px, and for 10 minute exposure that translates into ~0.25e dark current noise).

Which one would you rather have lowered at expense of other being higher?

[vlaiv]

17 minutes ago, vlaiv said:

Actually I would use shorter exposure darks and dark scaling.

This would only work for cameras with stable/usable bias files - which means basically CCD cameras as most CMOS sensors don't have stable / usable bias files.

It might seem counter intuitive to use shorter darks, but let me explain. Let's suppose that you have 5 minute subs and 10 minute subs, and let's suppose you have 4h dedicated time for taking darks. With most modern CCD cameras read noise is far greater than dark current noise. Difference between short and long sub dark current noise, or rahter dark current noise increase is by factor of ~1.41 (square root of 2). Also, when you stack double number of subs (given that we have limited dark budget time - you will get twice as many half duration subs), read noise decreases by factor of 1.41.

So it's the matter of which one is bigger - and with modern CCD sensors, it is the read noise that dominates in these exposures (order of 6-7e per exposure, while dark current is roughly 0.0001 e/s/px, and for 10 minute exposure that translates into ~0.25e dark current noise).

Which one would you rather have lowered at expense of other being higher?

Actually, I'll retract above statement, @andrew s is absolutely right - +1 for the longest darks with scaling.

In above analysis, I made an error - read noise also gets scaled by scaling factor, so both noise sources raise by 1.41 in 5 minute subs vs 10 although there are twice as many of them (5 minute subs).

[ollypenrice]

The wider question would be why you want to blend different exposure times? If it is to cover high dynamic range then you'd want to make a long stack and a short stack and then use a blending method either bought in as an HDR package or do it manually using Photoshop layer masks.

The best way to work on 99% of targets is to find the optimal exposure length and shoot only that. M42 needs multiple exposure lengths but it's very much the exception to the rule.

Olly

[alan potts]

6 minutes ago, ollypenrice said:

The wider question would be why you want to blend different exposure times? If it is to cover high dynamic range then you'd want to make a long stack and a short stack and then use a blending method either bought in as an HDR package or do it manually using Photoshop layer masks.

The best way to work on 99% of targets is to find the optimal exposure length and shoot only that. M42 needs multiple exposure lengths but it's very much the exception to the rule.

Olly

Somewhat away from topic, I have done this with M42 but would M31 not benefit from this technique?

Alan

[ollypenrice]

44 minutes ago, alan potts said:

Somewhat away from topic, I have done this with M42 but would M31 not benefit from this technique?

Alan

Just maybe. The real trick in getting further into the core is extracting the few contrasts that are in there. My own efforts wouldn't bear performing in a public place, they weren't pretty!

Olly