Yet another calibration question

[Alestrom]

Hi everyone

I wonder if you could confirm something for me. So far I have not got round to shooting flat frames, but I have been using darks and bias. I think I have just realised that bias frames are only required if using flats, as the darks contain the bias. Using DSS, a master bias is created then subtracted from the dark, which are then combined into a master and then both subtracted from the lights. Common sense now tells me that I would get the same result using darks only, and the process would be quicker (the darks are not being scaled by the way, I understand that you do need bias frames if the darks are a different length to the lights).

Can someone a bit more knowledgeable clear this up for me?

Someday I will get round to using flats (my only attempts so far did more harm than good!) and then I will need bias or dark flats, correct?

Cheers, and sorry for the length of this post

[blinky]

I think BIAS Frames help to calibrate the dark frames but the good news is that you only need 1 for each ISO - you dont need to retake them each time you are out. I would say that Dark and Flat frames are the most important ones to take. Dark's take out a lot of noise and Flat's get rid of vignetting as well as shadow's caused by dust bunnies.

[Psychobilly]

Perhaps thes diagrams will help they are taken straight from the DSS Online Help Files section on Taking Better Pictures....

Peter...

[Alestrom]

Yes I've looked at those diagrams but the problem is they only consider cases where you have flats. I know you need the biases or dark flats if you are taking flats, but I am not.

What I mean is similar to what Blinky said, do the bias frames help calibrate the darks? I thought you only need them if the darks are being scaled because they are a different length (which will never be an issue I don't think).

If the darks already contain the bias then why would taking them separately help (if you don't have flats)? I will start taking flats as I appreciate the value of them, but I haven't done so as yet.

Cheers

[Psychobilly]

If you ignore the FLATS for now and look at the diagrams again... then you effectively rule out the NO BIAS (DARK FLAT option) so your left with the NO DARK FLAT option which shows that you need both DARKS and BIAS (you not using FLATS)

Thats how I would "read" the calibration flows... Until I started taking Flats and Dark Flats I just used to use Light,Darks and Bias when stacking in DSS

If your getting vertical or horizontal banding in your image normally more obvious in the RED then you need to use BIAS frames

The Calibration frames that you need to use will depend to a large extent on how Noisy your camera is and how far you want to stretch the data.

What Camera are you using... I have quite a few and they all behave differently

Peter..

[Alestrom]

Fair enough. I am just trying to work out what the advantage of a separate bias is when it's already in the dark anyway, so that I understand how important it is to include them.

My camera is a Canon 1000D

[SteveP]

Alestrom

You're correct in saying that the Dark already incorporates the Bias data. If you are taking lights and using darks at the same exposure, then that would be sufficient. Bias frames come into their own (so far as darks are concerned) when you don't have a dark at the same exposure as the light and the software attempts to scale the dark to the lights' exposure. Because the noise for want of a better word in the dark is not strictly linear, ie there is an amount which is 'fixed' irrespective of temp, the scaling process in this situation is flawed. It's the bias frame which captures this 'fixed' element and by removing it allows the darks to be scaled correctly. For me, whether you needs bias frames is dependent on how close is the exposure of the dark and the light. In many cases, you can get away with it. However bias frames are relatively easy to take and therefore don't add much time to the overall workflow

HTH

Steve

[Alestrom]

That's pretty much what I thought Steve. Cheers for the imput everyone.

[themos]

One thing I only appreciated recently is that it is very important to shoot many flats. The reason is that the flats contain quite strong signals (of the order of the saturation level). Consequently, shot noise (which goes with the square root of the signal) is also high, much higher than the noise in a bias or dark frame. HAIP work through an example where the variation from bias frame to bias frame is 32 ADU but the variation from flat to flat frame is 248 ADU. You need to bring that 248 down by averaging LOTS of flats otherwise flat-fielding will add way too much noise in your image. The 32 does not need that much bringing down, in comparison.

[themos]

Well, I have probably overstated the case for many flats. What matters is the S/N ratio of each calibration shot. You want to make sure that none of your calibration shots have a S/N ratio anywhere near the lights' S/N ratio. But it is fairly easy to get a S/N ratio of 500 when shooting flats (and combining something like 10-20) and it's very unlikely that your lights are going to have a S/N of over a 100, so flats is not the bottleneck.

[dph1nm]

Strictly speaking you should have a S/N in your master calibation frames which is greater than the S/N in the sum of ALL your lights. I suspect most people (myself included) never get anywhere near this.

NigelM