Go Back   IceInSpace > Equipment > Astrophotography and Imaging Equipment and Discussions

Reply
 
Thread Tools Rate Thread
  #1  
Old 08-01-2019, 11:41 PM
Stonius's Avatar
Stonius (Markus)
Registered User

Stonius is offline
 
Join Date: Mar 2015
Location: Melbourne
Posts: 1,495
Confused about exposure times for L+RGB Binx2

Hi all,

This all came about because I'm using an ASI 1600 on my Esprit 120mm which is fine, but if I stick it on my RC10, I'm oversampled. My understanding is that for a CMOS, you might as well bin in post, not when capturing, as the advantages of CCD's doing a single read of four pixels doesn't apply here. However there are lesser SNR advantages to be had from binning in post. I've also read that Binning at 2x makes your sensor 4 times more sensitive.

Okey Doke.

So if my unbinned luminance exposure is 1 min and I'm planning to bin x2 in post, do I expose the same as I normally would (1 min)? Or do I expose for 15 secs, knowing the binning will bring back the detail that has been lost?

And then in post, you have the choice of true additive binning, or averaging, which has the advantages of SNR without allowing out of range values.

So there are a few questions I'm wrestling with;
  • Am I correct in thinking that binning 2x makes the sensor 4X more sensitive? So the information gained in 1 minute would then be gained in 15 seconds instead?
  • Then you add the filters which in my case increases exposure times by multipliers of G2.7, R3, and B3.7 respectively, turning those 15 second exposures into exposures of G41, R45 and B56 seconds.
  • Now I don't seriously want to compensate for the filter transmission and QE by shooting darks for three different exposure levels on the colour filters. Is it possible to shoot more subs to make up the difference? How many more subs would be required to get the same SNR on all channels?

Any help appreciated

Markus
Reply With Quote
  #2  
Old 09-01-2019, 06:54 AM
Marke's Avatar
Marke (Mark)
Registered User

Marke is offline
 
Join Date: Nov 2009
Location: Sydney
Posts: 1,193
I wouldnt worry to much about over sampling I use a 1600 RC10 for a couple years now and if you use an AP 0.67 reducer it will give you better sampling and a flatter field.
Reply With Quote
  #3  
Old 09-01-2019, 07:37 AM
RickS's Avatar
RickS (Rick)
PI cult recruiter

RickS is offline
 
Join Date: Apr 2010
Location: Brisbane
Posts: 10,584
Hi Markus,

There are two things related to SNR going on with hardware binning.

One that you mentioned is that you may get a benefit in reduced read noise. In a perfect world you'd get the same amount of read noise in a "pixel" whether you're doing x1, x2 or more. In the real world you'll often get less benefit than this. My experience with KAF-8300 and KAF-16803 sensors is that you're doing well if you only get 2x the read noise with x2 binning. The other thing to say about this is that if you're able to take sky limited subs then you don't care about read noise.

The second thing is that you're measuring more photons in a single pixel when binned. Shot noise is SQRT(signal) so binning x2 gives you double the per-pixel NR (at the cost of reduced resolution.) This can be done just as well in software. Another way of skinning this cat is software noise reduction - a controlled blur that loses some resolution in return for improved SNR. I like noise reduction better than binning because you can specifically target the areas where noise is most visible: the low signal areas and small spatial scales.

There's another potential benefit from not binning your RGB. You can use it to create a synthetic luminance or to improve real luminance data.

Summary: IMO, binning is usually a waste of time The only justification I can find for it is when you're chasing dim NB targets and can't do subs that are long enough to be sky limited.

Cheers,
Rick.
Reply With Quote
  #4  
Old 09-01-2019, 11:53 AM
Stonius's Avatar
Stonius (Markus)
Registered User

Stonius is offline
 
Join Date: Mar 2015
Location: Melbourne
Posts: 1,495
Quote:
Originally Posted by RickS View Post
The second thing is that you're measuring more photons in a single pixel when binned.
Maybe that's true for CCD, but that's not what I'm finding with my CMOS. For me, the mean value stays the same - it's the StdDev that changes - which I guess means there is less noise and so the signal can be boosted to the same effect as having captured more photons?

Quote:
Originally Posted by RickS View Post
I like noise reduction better than binning because you can specifically target the areas where noise is most visible: the low signal areas and small spatial scales.
That actually makes a *lot of sense.

Quote:
Originally Posted by RickS View Post
Summary: IMO, binning is usually a waste of time The only justification I can find for it is when you're chasing dim NB targets and can't do subs that are long enough to be sky limited.
I think that's my issue - how do I use binning to save time on the field? This necessitates less integration time per sub and therefore shorter (or less) exposures. Since binning during capture (I hesitate to call it 'hardware binning') doesn't actually increase the mean values of the sub, do I;

1) Halve the exposure, but still shoot at 1x1, then bin in post? Surely some of the binning gains are lost to increased read noise contributions?

2) Halve the exposure and shoot 2x2. The median values will be the same as (1), but there will be less SNR which will allow me to boost the signal 2x to arrive at a binned sub of the same median values as (1).

3) Halve the exposure, live with the noise and use selective masking and noise reduction in post instead of binning.

There's also the difference between integer resample and Pixelmath combinations. If you choose average as the method in Integer Resample, you won't get out of range values, you just reduce the noise (so I'm told, but in my experiments, it seems to have very little effect on noise). Which is different from straight binning 2x2 in PixelMath, which *does reduce noise with the danger of out of range values (clipping), using;

Code:
x2 = 2*x();
y2 = 2*y();
pixel( $T, x2, y2 ) + pixel( $T, x2+1, y2 ) + pixel( $T, x2, y2+1 ) + pixel( $T, x2+1, y2+1 )


Best,

Markus
Attached Thumbnails
Click for full-size image (Binning Tests.JPG)
53.4 KB12 views
Reply With Quote
  #5  
Old 09-01-2019, 06:32 PM
kens (Ken)
Registered User

kens is offline
 
Join Date: Oct 2014
Location: Melbourne, Australia
Posts: 314
Quote:
Originally Posted by Stonius View Post
Maybe that's true for CCD, but that's not what I'm finding with my CMOS. For me, the mean value stays the same - it's the StdDev that changes - which I guess means there is less noise and so the signal can be boosted to the same effect as having captured more photons?
That simply means the software binning is normalising the result. Binning 4 pixels means you are combining the electrons from those 4 pixels
The only benefit to binning in the driver is saving disk space. And some time saving if you are running client/server and have to download the subs over a slow network.
No change to exposure time or integration time.
Reply With Quote
  #6  
Old 12-01-2019, 07:19 AM
RickS's Avatar
RickS (Rick)
PI cult recruiter

RickS is offline
 
Join Date: Apr 2010
Location: Brisbane
Posts: 10,584
Hi Markus,

Apart from minimising the effect of read noise there's no shortcut you can use to reduce time in the field. You're limited by shot noise and the only way to reduce the impact of that is to capture more data.

I'm not sure why you don't see larger numbers for binned pixels but as Ken said some normalisation must be going on.

Using sum or average (mean) gives you the same result in the end and there's no risk of overflow once you're dealing with floating point data in PixInsight (which normalises all values to the range 0..1 anyway.)

Cheers,
Rick.
Reply With Quote
Reply

Bookmarks

Thread Tools
Rate This Thread
Rate This Thread:

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump


All times are GMT +10. The time is now 06:06 AM.

Powered by vBulletin Version 3.8.7 | Copyright ©2000 - 2024, Jelsoft Enterprises Ltd.
Advertisement
Testar
Advertisement
Bintel
Advertisement