Picking the correct exposure for Deep Sky

A place to post guides, instructions and HOWTO documents on the use of SharpCap and Astrophotography in general
Forum rules
One tutorial per topic please.
Please only post content that you have created in this form.
Tutorials should be either in PDF format as an attachment or written in the first post of the topic. Links to video tutorials you created are also acceptable.
Borodog
Posts: 97
Joined: Fri Jan 01, 2021 7:25 pm

Re: Picking the correct exposure for Deep Sky

#51

Post by Borodog »

Hmm. How does that work mathematically?

Take for example my ASI183MC. Shooting in might light pollution (B5) at 4/f.8, this difference in optimal exposure at 0 gain and unity gain is about a factor of 3. The difference in DR is about 1.5 stops, about a factor of 2.7. The FWC at 0 gain is 15k; at unity it's just over 4k. So 3 unity gain exposures don't get me quite back to the same capacity.

Not trying to argue; just trying to understand.

SharpCap's Smart Histogram has 2 selectable gain targets, MDR & unity. But many people say that unity is not really a special or magical gain. Why not half unity? Or twice?

A separate consideration for my camera is that it is subject to banding that is worse at lower gain, so that alone might be a reason to choose unity over zero. Although I haven't tested SharpCap's banding suppression for the 2 different gains.
User avatar
admin
Site Admin
Posts: 9137
Joined: Sat Feb 11, 2017 3:52 pm
Location: Vale of the White Horse, UK
Contact:

Re: Picking the correct exposure for Deep Sky

#52

Post by admin »

Hi,

if there is a sharp drop in read noise as the gain is increased then the minimum exposure length needed can drop *a lot* - halve the read noise and the exposure length drops by a factor of 4, so you can take 4x as many frames in the same length of time. Each of them may have a bit less dynamic range than the low gain single frames, but the fact that you have 4x as many can overcome that.

This definitely happens for cameras that have a switchover from LCG (low conversion gain) to HCG (high conversion gain) at a particular gain value, since the step decrease in read noise is usually enough to cause this to happen. For cameras without that step drop in read noise, it all depends on the relative rates of change of the dynamic range and read noise as the gain changes.

chreers,

Robin
barnold84
Posts: 9
Joined: Thu Apr 08, 2021 6:44 am

Re: Picking the correct exposure for Deep Sky

#53

Post by barnold84 »

Hi,

I'm having some confusion here about the read noise. As stated in the beginning, we can safely assume that the read noise follows a Poisson distribution. Say that this distribution has an expectation value of lambda and therefore, the variance is lambda (and standard deviation sqrt(lambda), which is often abbreviated with sigma).

Now, when the vendor states the read noise, I would assume that this would be our lambda? If so, going back to the equations in the first posts, what am I using for sigma_r or sigma_r^2 respectively to compute the exposure time?
My understandng would be that the read noise sigma_r^2 = lambda = stated read noise of camera?

e.g. read noise = 5e-, sky electron rate = 1.5e-/s, allowed noise 5%, hence tmin = 1/(1.05^2-1) * 5/1.5 = 32.5s?
Or is it tmin = 1/(1.05^2-1) * 5^2/1.5 = 162.6s?

Björn
User avatar
admin
Site Admin
Posts: 9137
Joined: Sat Feb 11, 2017 3:52 pm
Location: Vale of the White Horse, UK
Contact:

Re: Picking the correct exposure for Deep Sky

#54

Post by admin »

Hi Björn

you have your noises mixed up I think... It's the shot noise that follows the poisson distribution, with the standard deviation being equal to the square root of the number of electrons measured. This noise is not a property of the camera, it's simply down to the quantization of photons and the electrons they are converted into.

The read noise of the camera is (probably) normally distributed and is a property of the camera, independent of the number of electrons being captured at an individual pixel.

Hope this helps,

Robin
barnold84
Posts: 9
Joined: Thu Apr 08, 2021 6:44 am

Re: Picking the correct exposure for Deep Sky

#55

Post by barnold84 »

Hi Robin,

I guess the question if something is Poisson is not the most important aspect here. The justification would anyways be done if the underlying process is about rare events. In the limit, the Poisson distribution converges into a normal distribution.

My question is indeed what to understand about the term "read noise" or to be precise the underlying process causing this noise. My model was always to see read noise in the way that the reading event is "creating" additional electrons, i.e. read noise which would be modelled by a Poisson process. So from what you're indicating is that the causing process is a different one and if we would always read a pixel with a constant number of electrons (since we prepared a pixel in such a way), we would create a gaussian variation where the standard deviation is the read noise and independent of the actual number of electrons in the pixel?

Björn
User avatar
admin
Site Admin
Posts: 9137
Joined: Sat Feb 11, 2017 3:52 pm
Location: Vale of the White Horse, UK
Contact:

Re: Picking the correct exposure for Deep Sky

#56

Post by admin »

Hi Bjorn,

yes, that's how I understand it - the read noise is a combination of the fluctuations in the analogue electronics and the A->D conversion that eventually turn the pixel brightness from a number of trapped electrons into an ADU number. This is independent of pixel brightness.

The read noise is also uncorrelated with the poisson shot noise, which means that the two noise sources need to be added in quadrature (square each, add, take square root). This allows the read noise to have practically zero impact once the shot noise level exceeds it by a factor a factor of 5 or so.

cheers,

Robin
barnold84
Posts: 9
Joined: Thu Apr 08, 2021 6:44 am

Re: Picking the correct exposure for Deep Sky

#57

Post by barnold84 »

Hi Robin,

Thanks for clarification. I'll try getting a bit more of physical background of read noise.

I don't want to spill this topic with other issues, so please let me know if I should move this to a different thread:
your making the point of swamping the read noise to which I fully agree. I'm still wondering a bit why some people work with second or sub-second exposures for DSOs when the read noise is hightly significant. They never would reach the SNR that long exposures do, or am I missing something? I can't imagine that it's all about detail (w.r.t. to resolution) and a lot of equipment would at least allow 30 or 60 second exposures (guided and some unguided).

Best,
Björn
User avatar
admin
Site Admin
Posts: 9137
Joined: Sat Feb 11, 2017 3:52 pm
Location: Vale of the White Horse, UK
Contact:

Re: Picking the correct exposure for Deep Sky

#58

Post by admin »

Hi Bjorn,

I think you are right that the use of very short exposures will never get good SNR (possibly except on very bright targets with large pixels and fast optics). On the other hand, you can pick up exceptionally sharp images if you can freeze the seeing variations to some extent.

cheers,

Robin
Post Reply