Should I included bias if not including darks?

[smr]

Whilst imaging NGC2244 last night I was reading up about calibration frames and from what I gathered from this forum and others, it seems that darks may be more hassle than they are worth. I have a newer sensor Canon 80D with on sensor dark current suppression which is applied to the pixels whilst the light frames are taken. Last night was quite cold, not freezing but not far from it, and I have read that using darks requires the sensor to be pretty much at exactly the same temperature as the lights were which is difficult to ascertain. As well as this I also read that they can introduce even more noise to the final result in such sensors. So with all that read I elected not to take any.

I did however take flats, but I also took some bias frames, I know I need to take flats and apply them, but if I am not using darks is there any point in using the bias frames I took?

[happy-kat]

I'd make sure to take dark flats as well as your flats as I've found they do help with the dslr I use.

[smr]

Just now, happy-kat said:

I'd make sure to take dark flats as well as your flats as I've found they do help with the dslr I use.

Thanks, I'll read up on them, but I am wondering before stacking whether to include the bias frames I took, or just include the lights and flats only?

[happy-kat]

I've found it useful to create a stack with without etc then see what the differences made in processing.

[spillage]

Hi

I would say flats  and dark flats, then dither instead of darks. But It would be worth experimenting with bias and darks to see how it looks. Then try replacing the bias/darks with dark flats.

[jjosefsen]

Hi,

Normally your darks would include the bias pedestal, but since you are not using any darks I would definitely take bias and subtract these from the lights.

You also need to subtract the bias pedestal from your lights, so either with bias again or with darkflats. If you don't you will just reintroduce the bias pedestal as noise into the lights when you calibrate with flats.

My recommendation would be to use darkflats if you shoot narrowband as the flats are typically much longer with a NB filter, like for example Ha. (With unity gain on my CMOS cam my Ha flats are around 40 seconds, enough to introduce dark noise).

[vlaiv]

Depends on sensor really.

You can calibrate with bias only, but these are problems that you can face:

- if your dark current is uneven you can get gradients in the image, like those due to amp glow. These can be calibrated out only with darks.

- if you have excessive dark current and you take long flats - your flats can over correct.

You can also get artifacts if your camera is doing internal "optimization" as well - because you will not control full calibration process.

Because darks are not reliable without temperature control, you can get above artifacts anyway, so most people using DSLR use bias only and no darks, or apply dark optimization - which works with stable bias.

This would be bias only calibration:

- take bias and flats as calibration frames

- shoot lights

- subtract master bias from flat stack (or subtract master bias from each flat sub and do the stack - for simple average stack it is the same, but if you use advanced stacking method then calibrate each flat sub prior to stacking)

- divide (light - master bias) with master flat

 

[smr]

Wow am I glad I took bias frames, thanks for the advice all.

I stacked the image with just flats and lights, and after stretching loads of horrible banding and smudges etc. I was really disappointed. 

Last night I read a few articles saying that bias frames were not necessary at all, well that's a load of rubbish. 

I stacked again but this time with the bias frames added, I took about 50 or more bias frames. What a difference, the image is a lot smoother, no horrible banding or smudges or anything, it looks really nice and clean and I'm able to stretch the data nicely too. 

[jjosefsen]

6 hours ago, smr said:

Wow am I glad I took bias frames, thanks for the advice all.

I stacked the image with just flats and lights, and after stretching loads of horrible banding and smudges etc. I was really disappointed. 

Last night I read a few articles saying that bias frames were not necessary at all, well that's a load of rubbish. 

I stacked again but this time with the bias frames added, I took about 50 or more bias frames. What a difference, the image is a lot smoother, no horrible banding or smudges or anything, it looks really nice and clean and I'm able to stretch the data nicely too. 

You can reuse that bias for at least a year I would say. So go ahead and do a couple of hundred or more of them, and just reuse the master. ?