Flats across temperatures during session

[BrendanC]

I recently created a library of calibration files across a range of temperatures. It is a thing of rare beauty - dark flats, biases and darks at a range of temperatures to match the subs. So, if I have some subs at 24C and some at 25C, no problem, I just use groups in DSS so I have darks, dark flats and biases for the 24C group, and the same for the 25C group, etc. All good.

However, I realised the other day that I don't know how this works with flats (ie light flats, not dark) when there's a range of temperatures involved (I use a non-cooled DSLR so I can't control the temperature).

I understand that, as opposed to dark frames, they ideally need to be taken with the exact equipment used during the shoot because otherwise orientation, dust motes etc could change. So, that precludes building a library of them. They should be taken during the shoot at some point.

But let's say half of my flats are at 24C, and half at 25C. What do I do then? Do I have to take loads of flats at 24C, and another load at 25C? This might not be too much of a problem but what if I really do get a range, like from 20C to 28C or something? What then - are 8 separate sets of flats needed?

I'm kind of thinking that, assuming there's not too wide a range, I could just use the same flats for each group. So, if I have subs at 25C, 26C and 27C, I apply the relevant calibration frames, but apply the same flats for each group.

Or, is it viable to create a library across a range of temperatures, if I'm very careful to make sure the camera is oriented the same every time? Or is this too risky, as I'd imagine even slight misalignment could cause real problems ie instead of removing artefacts it would in fact double them.

Does this make sense?

[MartinB]

Being relatively short exposures the temperature of your flats isn't critical.  You can use a 24deg flat on subs taken at -10.  The benefit of flat darks is pretty marginal compared with using a simple bias but you can always collect your flat darks at the same time if you feel they make a difference.

[sploo]

I'd assume (with only experience and layman's understanding, rather than any actual evidence) that the sensor temperature shouldn't make much of a difference with light flat frames. The signal-to-noise ratio in a flat light image should be very good; whereas dark and bias frames are capturing the behaviour of the sensor + electronics in terms of noise in the system (which will be much more temperature sensitive). That's not to say temperature won't affect a light flat frame, but I'd assume that only really extreme temperatures should cause problems.

[BrendanC]

Thanks all - and it's a good point about the SNR for flat light images, I guess ideally they'd be at the same temperature but they're really for removing artefacts such as vignetting, dust etc. 

So, same flats across all temps it is, but taken at the shoot rather than in a library.

Thanks.