Different gain settings for flats?

[bluesilver]

Hi, Just been doing some reading up on flats and i have kind of come back with two very different options and was hoping I might be able to get a bit of advise on here on what may or may not be the correct way.

I have a asi2600mc and a Skywatcher Esprit 150 scope

I image at Gain 100
So I would of thought that when taking flats, I should also be keeping the gain at the same as it is for the Lights, so gain 100.
But I have now recently been reading that some are saying that you should always take your flats at a different gain to the lights, so gain 0 for the flats.
So the big question is, which one is correct?
I know that the darks need to be the same gain as the lights, But what about Bias and Flat Darks?
Do Bias need to be the same gain as the lights and the same question of r the Dark Flats.

Any advice on tis would be appreciated.
Thanks.

[admin]

Hi,

the gain itself doesn't really make any difference to flats, since what you are trying to measure is the relative amounts of light falling on different parts of the sensor. What matters is how many photons you capture per pixel in all your individual flat frames added together.

Let me explain that...

Suppose you take 10 flat frames and each (on average) captures 1000 photons - that means when you add them up, you have 10000 photons. The rules of statistics mean that the random error in capturing 10000 photons is a standard deviation of +/- 100 photons, so the noise level of your master flat is 1% of the signal.

Now, run through that again, but lower the gain and increase the exposure, so you are now capturing say 25000 photons per frame - total photons going into the master flat is ~250,000, error level on that is 500 photons (square root of number of photons), noise level is 0.2%

So, to keep the noise levels down in your master flat, you want to the total exposure time to actually be relatively long to let a good number of photons get detected. That's possible at high gain, but you would get a *lot* of very short individual flat frames to average together - much easier to have lower gain, longer exposure flats and less of them.

Ok, we can see that the deal is that the data processing is less work when using lower gains/longer exposures for flats, but it doesn't actually give any better or worse final results.

One more thing we need to cover though...

To make a decent master flat, you have to subtract from it either dark flat frames or bias frames - these remove the part of the signal that is the camera background and leave behind only that part that comes from light falling on your sensor during flat capture. Those dark flat or bias frames must be taken at the same offset/gain/etc as the flat frames. Dark flats need the same exposure, bias would use very short exposure.

Hope this helps,

Robin

[bluesilver]

Hi, Thanks for the very detail reply there, beginning to make more sense.
So if i am reading and understanding this correctly, For Flats I am better off using gain0 then, even if my lights are taken with gain100?
Using gain 0 will give me longer exposure with less noise?
If i want to take Dar Flats, these need to be exactly the same as the flats, Basically just put the cap back on the telescope and re run them?
For Bias, I would need to use the same gain as the Flats?
Now Darks, these need to be exactly the same as the lights just with the cap on the telescope?

So in summary, Flats, Dark Flats and bias, their gain must be the same as each other, but can be different from the lights and darks?

The offset though has to be the same for everything, Light Darks, Flats Dark Flats and Bias.

Dose that sound pretty close to be correct?

[admin]

Hi,

all correct, except that you have to be a bit careful with bias frames, depending on how you are using them.

If you are capturing flats in SharpCap then you can choose the option to use bias frames and it will all happen automatically - same gain, etc.

If you are processing in some other application, be sure how it uses bias frames. If you have dark flats and darks, and the settings for both of those are identical to the settings for flats and lights respectively then bias frames are not needed, so you are best off not giving any to the application, just in case it uses them when it shouldn't.

The only times you would need bias are

1) if you have darks a different exposure length to lights or (equivalently) dark flats a different exposure length to flats
or
2) Using them as a lazy-but-almost-as-good replacement for dark flats as SharpCap does

cheers,

Robin