problem with gain settings

[Johnp1951]

I have recently bought ZWO ASI1600 cooled pro camera and filters,  after using an ATIK camera for many years, this new camera is quite a challenge, for example the gain settings, how many astronomers have had this challenge, I took some images of Messier 27 last night at gain 175 offset 25,  taking 10 x 180 sec in luminance and R,G,B and HA O3 S2, the LRGB where a lot darker than I expected when blended together in photoshop.

 

has anybody else had this problem.

 

 

[symmetal]

When starting out with a CMOS camera it's best to select the unity gain setting, where 1 photon landing on a sensor pixel increases the pixel output ADU by 1. This gives a good compromise on dynamic range, output signal and exposure time. On the ASI1600 this is at gain 139. The offset you used seems a bit low and would likely lead to black clipping of some pixels on your images which is never a good thing. Offset 50 is a good setting to start with.

You'll need to use dark frames and not bias for calibration, as older CMOS cameras (like the ASI1600) suffer from amp-glow which increases with exposure. Also use flats exposures of at least a second along with corresponding flat darks as bias frames on some older CMOS cameras (including the ASI1600) are different to the bias component of longer exposures.

CMOS cameras have significantly lower read noise than CCD so many shorter exposures are taken rather than fewer longer exposures, particularly with 12 bit cameras (like your ASI1600) as they have less dynamic range than the 16 bit CCD camera you're used to.

Newer and unfortunately more costly CMOS cameras are 16 bit and have no amp glow, nor strange bias behaviour, so behave more like your CCD camera but with much lower read noise.

Fot the ASI1600, L exposures are often around 60 sec, while RG and B exposures are 3 times longer at around 180 sec as they each pass only about 1/3 the light of the L filter.  Narrowband exposures are as long as you can manage without seeing tracking errors, often 10 mins or longer if possible. If you have dark skies you can expose for longer.

The aim is to expose until the noise from the image sky background is three or more times higher than the read noise, and in so, making the read noise contribution to the total image noise insignificant. If this point is reached there is no difference in S/N of stacking say 10 five minute exposures against 5 ten minute exposures. The shorter exposures are preferred as less highlights will be blown in the final stack at the expense of more storage space needed. With narrowband you won't be able to swamp the read noise by the sky background noise, unless you expose for several hours so just make them as long as you can get away with.

You say your images look dark. Your total integration time is fairly low, so will look fairly noisy if stretched too much. I usually start out with 120 L subs of 60 secs (2 hours total) and 20 each of RG and B of 180 secs (1 hour each) and stack and process them to get a good indication of what the target looks like, and then add more subs as necessary to get the noise down. For fainter targets you may need to take 2, 3 or more times as many subs. For narrowband I start off with 10 of each filter of 600s subs, to see how it looks, and then add more as required.

This exposing long enough to swamp the read noise works equally well with CCD cameras but due to a CCD's  higher read noise the exposures would likely be longer than the 10 or 20 min exposures you usually use with CCD.

Hope this helps. 🙂

Alan

 

Edited June 24, 2022 by symmetal

[newbie alert]

15 hours ago, Johnp1951 said:

I have recently bought ZWO ASI1600 cooled pro camera and filters,  after using an ATIK camera for many years, this new camera is quite a challenge, for example the gain settings, how many astronomers have had this challenge, I took some images of Messier 27 last night at gain 175 offset 25,  taking 10 x 180 sec in luminance and R,G,B and HA O3 S2, the LRGB where a lot darker than I expected when blended together in photoshop.

 

has anybody else had this problem.

 

 

Hi John what CCD was you using before?

 

[Johnp1951]

Thank you for the advice Alan, will give it ago when the weather gets better.

[symmetal]

John, I forgot to mention that while unity gain is generally best for LRGB, for narrowband, some imagers use 2 x unity gain, which is gain 200 on the ASI1600. This is because the read noise decreases with higher gain, though at the expense of reduced effective well depth and so dynamic range. The graphs in the ASI1600 manual give the figures.

To check you are not black clipping pixels by having too low an offset, just take a 1 or 2 second dark and check the resulting stretched histogram curve in something like Fits Liberator to see what the minimum pixel values are. The image statistics in your capture program should also give the minimum pixel value too, which is a good indication of clipping.

It's best to only use one offset for all the gain settings you use, rather than changing offset with gain as it saves you having to keep track of different sets of darks for different offsets. Having to have different sets of darks for different gains is complicated enough when calibrating. 😀

Alan

 

[rickwayne]

I'll add that ZWO did some performance testing awhile back and concluded that overall best results were with low gain for high dynamic-range targets where clipping would be a problem, and high gain where it wouldn't.

If I sound as if I'm arguing with Alan, I'm not. Unity gain is an excellent starting place and you won't go wrong using it.

IIRC he's also right about not using bias calibration frames with the 1600, it gets nonlinear at very short exposures. So darks, flats, and dark flats are the way to go. Which means it also helps to settle on a limited number of temperature, gain, and sub exposure-time settings to control the combinatoric explosion of needed calibration frames. Much better to get a "good enough" sub time and vary the number of frames.

Many folk dance on the heads of the sub exposure-time and best-temperature pins and it's really beside the point. Noisy? More subs. No detail in the dark nebulosity? More subs. Tracking issues? Shorter subs and more of them. More integration time forgives any number of sins. 🙂