Image Processing Temperature Compatibility

[groberts]

I appreciate that using a mono camera with temperature control it's best to shoot all the same subs at the same temperature.  However, if you have two or more sets of the images taken at different temperatures (otherwise indentical settings e.g. exposure, gain etc), can they still be processed together - in my case using Deep Sky Stacker?  I appreciate it's not desirable but will it work OK?

I ask because my default temperature I use is -20C but at some of the higher temperatures we've been having recently (+30C) I've often not been able to get to my optimum temperature and can only image at lower temperatures e.g. -15C;  my camera is rated at -40 to -45C so when it's very warm / hot it struggles.  I could, of course just take all the images at a lower temperature but obviously when possible colder = better. 

Graham  

 

[vlaiv]

Just go for it - difference between subs taken at -20C vs those at -15C will be much smaller than for example subs taken in 10pm vs those taken in 1am.

I know this sounds funny - but lot of things change between 10pm and 1am - namely level of light pollution and position of the target in the sky - altitude. This has potential to reduce target signal by as much as 20-30% or more and background LP noise - so SNR will change.

Going from -20C to -15C will minimally change dark current noise in the image and since that component of the noise is very small - it will make almost no difference in the final result.

[Seelive]

I've always thought that the dark current doubles for about every 5C temperature rise, but as you say 2 x 0 = 0.

[groberts]

OK thanks - so the software will not reject because of the differences?

As we move into lower temperatures soon it will not be an issue ❄️ just the summer / late summer period.

Graham 

[vlaiv]

40 minutes ago, Seelive said:

I've always thought that the dark current doubles for about every 5C temperature rise, but as you say 2 x 0 = 0.

It does, or rather closer to 6.

However if you look at ASI1600 temp/dark current graph (and I'm assuming this camera from OP's signature):

we can see that dark current is about 0.0062e/s/px. Let's take average exposure of 2 minutes - 120 seconds to see what sort of dark current we have. That will be 0.744e per exposure of dark current, or ~0.86e of associated noise. For -15C, let's go with double that, so dark current will be 1.488e and associated noise will be ~1.22e.

Both of these are smaller than read noise of camera and hence LP noise (if sub duration is chosen properly). Consequently - total noise difference is very small - almost negligible for stacking.

Stacking works the best when SNR of each sub is equal, but it is never equal due to changing conditions in the sky - LP and target position and hence atmospheric attenuation. We don't really notice that in everyday imaging and software happily stacks our subs regardless. My point was that this temperature difference will make less of disturbance then regular conditions.

[Seelive]

I totally agree. I was just trying to make the point that a 5C change will actually approximately double the dark current noise, but as you say, every other noise source will swamp the dark current noise at those temeratures. Infact, I would suggest that when imaging from a light polluted site, any modern camera noise, even at 'temperate' temperatures, will probably be less than the sky background photon noise?

[vlaiv]

1 hour ago, Seelive said:

Infact, I would suggest that when imaging from a light polluted site, any modern camera noise, even at 'temperate' temperatures, will probably be less than the sky background photon noise?

This really depends on several factors. First is that sensor is never at ambient temperature - it will get warmer and it can only be at close to ambient temperatures when having some sort of active cooling. It does not need to be Peltier cooling - it can be simple fan and radiator cooling like some cameras from Altair Astro and Moravian Instruments (G1-2000C for example).

Let's suppose that sensor will be about 10-15C above ambient if not actively cooled. That means about 35C in the summer - not unreasonable assumption.

If we take ASI1600 and 0.0062e/px/s at -20C, we have 55C of difference and if we divide with 6 we will get roughly 9 times doubling, or 2^9 which is x512 in value.

This is ~3.175e/px/s - not value that can be neglected easily.

LP background rate will depend on few factors - sqm level of LP, sampling rate at which we image and used aperture.

If we image for example at 2"/px with 80mm aperture in SQM19 which is Bortle 6 / beginning of red zone, we will have sky flux of about 2.2e/px/s (for 50% QE sensor).

In this case, dark current beats LP flux by about 50% in intensity so noise from dark current is larger than from LP (and mind you, this is fairly "fast" system in strong LP).

[tomato]

Assuming the moon does not intrude I find the elevation of the target on the sky makes the biggest difference to the quality of the subs.

Alas, I don’t think I would have completed a single image if I restricted my imaging to an hour either side of the meridian.