diminishing returns

[Anthonyexmouth]

is there an easy way to work out how much time you should spend on a target before you see diminishing returns?

just wondering as I think i had some good data from the pelican last night, 3.5 hours and wondering how much more time i should spend on it, if any. 

 

 

[DaveS]

I remember reading on here, I think it was either @gorann or @wimvb who commented on one of my images, that to see any real improvement you have to double the data you already have. Though of course, any extra data will help.

[JamesF]

This was the basis for Robin Glover's talk at PAS this year (eg here: https://stargazerslounge.com/topic/332167-talk-by-robin-glover-sharpcap/)

I think a reasonable summary might be that the point at which returns are diminished to nothing of value really depends on the individual situation -- camera, location and conditions: if you have a very dark sky and/or a camera with higher read noise then longer subs may become more worthwhile whereas from somewhere more light polluted or you have a low-noise camera then you reach the point where greater sub length doesn't improve things sooner.

He's done a load of work in Sharpcap to try to estimate optimal sub length, though it might only be available in Sharpcap Pro.

James

[Anthonyexmouth]

5 minutes ago, JamesF said:

This was the basis for Robin Glover's talk at PAS this year (eg here: https://stargazerslounge.com/topic/332167-talk-by-robin-glover-sharpcap/)

I think a reasonable summary might be that the point at which returns are diminished to nothing of value really depends on the individual situation -- camera, location and conditions: if you have a very dark sky and/or a camera with higher read noise then longer subs may become more worthwhile whereas from somewhere more light polluted or you have a low-noise camera then you reach the point where greater sub length doesn't improve things sooner.

He's done a load of work in Sharpcap to try to estimate optimal sub length, though it might only be available in Sharpcap Pro.

James

whats the process you use in sharpcap? i've only ever used it for PA but i would like to try and get more out of it. 

[JamesF]

5 minutes ago, Anthonyexmouth said:

whats the process you use in sharpcap? i've only ever used it for PA but i would like to try and get more out of it. 

There you have me, I'm afraid, as I'm not actually a SharpCap user :)

Someone else will fill in the details though, I'm sure.

James

[don4l]

20 hours ago, Anthonyexmouth said:

is there an easy way to work out how much time you should spend on a target before you see diminishing returns?

just wondering as I think i had some good data from the pelican last night, 3.5 hours and wondering how much more time i should spend on it, if any. 

 

 

I think that you need to be a mathematical genius to get a proper answer to your question.

 

I've read that doubling your exposure time improves your signal to noise ratio by 1.4(ish).  If I combine 2 subs and compare the result to a stack of 4 subs, I can usually see what this difference looks like.

However, I recently had a go at the Squid in SH2-129, which is a notoriously weak OIII target.  After 2 hours of data I had a result which looked promising.  I then got another 6 hours of data before having another at processing.  I was horrified to see that the 8 hours showed absolutely no improvment at all over the 2 hours.

The subs in the extra 6 hours all looked to be of similar quality to the first two hours.

I don't understand what the problem is.  I think that it is either related to read noise, or sky background or something else.

I have just received a new camera which is much more sensitive to OIII and it has lower read noise.  If you see a post from me in the next few weeks about SH2-129, then have a look, because I will definitely comment on this issue.

 

 

[Ouroboros]

Assuiming all things being equal then statistically  the signal to noise ratio will double when you take 4 times as many subs. In reality contributions to noise and image quality can change over an image session - thermal noise, light pollution, seeing, focus and so on. 

One of the advantages of long imaging sessions is that you can afford (possibly) to throw away the poorer subs. More is not necessarily better. But more better is definitely better, if that makes sense. 🙂 

[wimvb]

This article may be of interest. 

https://www.skyandtelescope.com/astronomy-blogs/astrophotography-benefits-dark-skies/

[wimvb]

On 18/09/2019 at 15:59, Anthonyexmouth said:

is there an easy way to work out how much time you should spend on a target before you see diminishing returns?

just wondering as I think i had some good data from the pelican last night, 3.5 hours and wondering how much more time i should spend on it, if any. 

 

 

Imo there is no point of diminishing returns. The more time you spend on a target, the better results you should expect in terms of signal to noise. If 4 times as much integration time gives you double the signal to noise ratio, it will always do that, no matter what SNR you start with. That's as far as theory goes.

I think that in the end it's the astrophotographer's patience and perseverance that determine the point of "this is it for me" . Despite Jerry Lodriguss article, the time people spend on a target does not depend much on the Bortle value of their skies. I could be wrong of course.

In other words, spend as much time on a target as you need to get to a result you're pleased with, and forget about theory. 

Edited September 19, 2019 by wimvb