Page 3 of 3

Re: Understanding Binning

Posted: Fri Feb 21, 2020 2:58 am
by joeastro
HI.. very good.. but you missed the point.

when do you bin?

as the raw image cmos is read out to a full frame image? do you bin that first?????
then apply dark flat bias???

or

do you bin last? what isthe noise after you bin after you calibrate the image!!!!

nosie should be slightly better as your not workign with the raw image.

now the problem with sharpcap...
when you select a frame size, and then click a binning.. the software runs
out and checks the flat and dark.. and pops up a warning that th flat and dark
do not match the FRAME SIZE??? but you dont say if the frame is fULLLLLLLL size
when i check 5000x3000 pixels and 2x2 binning..

or

the 1x1 flats and darks i took dont match the final image 2x2 binned ..

which means i have to take dark and flats all over again???

will you tell us what the software order of binning and flats and darks...etc...DO!
in what order and why you cant scale 1x1 flats and darks so we dont have to do over do over do over.

thanks
joe
trying to use sharpcap

Re: Understanding Binning

Posted: Fri Feb 21, 2020 9:28 pm
by admin
Hi,

The golden rule is to always treat all the different types of frames in the same way – if you're going to capture your light frames 2 x 2 binning then you must capture your dark, flat and bias frames at the same 2 x 2 binning setting. It's not worth the risk of trying to reuse unbinned calibration frames – you can certainly load them into some piece of software and apply some sort of digital binning to them, but there's always the chance that it won't be done exactly the same way as the binning on the camera which would then lead to problems with either dark subtraction or flat correction. As I say, not worth the risk.

Cheers, Robin

Re: Understanding Binning

Posted: Tue Jan 17, 2023 11:11 pm
by han59
I know this is a very old topic, but it is a good start for me to discuss the S/N for the dominant sky noise case:
1) binning increases the signal-to-noise-ratio (SNR) of an image
2) CMOS style software binning increases the SNR in proportion to the N (of NxN binning), so 2x2 doubles the SNR, 3x3 triples it, etc
3) CCD style hardware binning gives an even bigger improvement, but only because the unbinned case has such a low SNR due to the high read noise
4) Adding and Averaging software binning both give the same SNR improvement
I assume 1) this is valid for the sensor read noise but for modern sensors the sky noise/glow is dominant after maybe 10 seconds. For the case the sky glow is dominant compared with the read noise then noise of a star or deepsky object is only influenced by the sky glow behind the object surface and equal to the sqrt of the number of electrons received. It doesn't matter if the star/deepsky object is captured with 10 or 20 or 100 pixels. The sky noise is the sqrt of the number of electrons received by sky glow for the object surface.

In other words for dominant sky noise the SNR of a star is not depending on the binning. It is only depending on the FWHM and exposure duration.
That is also what I measure and observe.

The explanation can also be found in the formula at 5.2 at

https://www.eso.org/~ohainaut/ccd/sn.html

S/N = s.t / sqrt ( n_pix . sky . t)

if the size of the pixel decreases and therefore the number of pixel to capture the star/object increases then the sky signal for each pixel decreases and there is no difference. Only exposure duration and sharper stars (less background area) so better seeing/focus will increase the S/N.

Please tell me if this assumption is correct.
Han

Re: Understanding Binning

Posted: Wed Jan 18, 2023 8:39 am
by jolo
I have once come to the same conclusion.
Hardware binning reduces read noise, but hardware binning is not available in CMOS sensors.
And anyway, the read noise in CMOS cameras is low compared to other noise sources (or at least we should adjust other parameters to keep read noise insignificant).

So the only purpose of binning in CMOS sensors for me is to have output files smaller if the image is oversampled for example.

Lucas

Re: Understanding Binning

Posted: Wed Jan 18, 2023 3:35 pm
by admin
@Han,

I think you are on the right track when you are thinking of 'noise' as being the variation in brightness of a star due to random fluctuations. Actually, in that case the noise level in a star will be determined by the total number of photons captured over the area considered to be part of the star. It doesn't matter what pixel size you use (and as you say, read noise is likely to be small). The total brightness of the star is obviously proportional to the number of photons collected (really the number of electrons they produce) - the noise in that brightness measurement is the square root of that number, so the SNR is also the square root of the number of electrons collected. Hence the only way to reduce the noise when thinking of it that way is to increase the number of electrons collected - more aperture or longer exposures.

However, we use the word 'noise' for a lot of things. One other thing we use it for is the grainy effect we see in the background areas of images when exposures are 'short' by some measure. Applying binning (CMOS or CCD) will reduce this effect. If the signal that you are trying to spot amongst the noise is faint nebulosity, then this reduction of the background noise level may be enough to let the signal appear.

cheers,

Robin