Re: Exposure comparisons with CCD cameras


Benoit Schillings
 

The best guess then is that your camera has significant readout noise
or quantization noise.

-- benoit

On Wed, Aug 19, 2020 at 9:43 AM uncarollo2 <chris1011@aol.com> via
groups.io <chris1011=aol.com@groups.io> wrote:


Roland, you should try averaging the six ten minute exposures. The S/N will be higher, and the result should look closer to the single 60-minute image.

Actually I did that. I tried combining 3 different ways median, average and sum. None of them came close to the sig/noise of the single exposure.

Roland


-----Original Message-----
From: Ray Gralak <groups3@gralak.com>
To: main@ap-gto.groups.io
Sent: Wed, Aug 19, 2020 9:17 am
Subject: Re: [ap-gto] Exposure comparisons with CCD cameras

Median combined can be an issue for very slight background variation.
How confident are you of the software auto-scaling... I would do a
simple mean for the comparaison, even if you end up with a few hot
pixels
Benoit is right.

Roland, you should try averaging the six ten minute exposures. The S/N will be higher, and the result should look closer to the single 60-minute image.

-Ray Gralak
Author of APCC (Astro-Physics Command Center): https://www.astro-physics.com/apcc-pro
Author of PEMPro V3: https://www.ccdware.com
Author of Astro-Physics V2 ASCOM Driver: https://www.siriusimaging.com/apdriver


-----Original Message-----
From: main@ap-gto.groups.io [mailto:main@ap-gto.groups.io] On Behalf Of Benoit Schillings
Sent: Tuesday, August 18, 2020 7:10 PM
To: main@ap-gto.groups.io
Subject: Re: [ap-gto] Exposure comparisons with CCD cameras

Median combined can be an issue for very slight background variation.
How confident are you of the software auto-scaling... I would do a
simple mean for the comparaison, even if you end up with a few hot
pixels

-- benoit

On Tue, Aug 18, 2020 at 6:07 PM W Hilmo <y.groups@hilmo.net> wrote:

“I'm not shooting RGB, I'm shooting NB where the 10 min object signal is at 700 or so out of 65,000.”



I realize that you weren’t shooting RGB, and I know that you can’t do a single 60 minute broad band exposure
without saturating the sensor. I am saying that if you were shooting RGB (and if the sensor had sufficient well
depth to deal with that), I would expect the two results to be closer to each other. The reason for that is that
that the dominant noise would be shot noise from a much higher signal in the background sky. Since you are
shooting narrow band, both the signal and shot noise are very small. That makes the read noise relatively
more significant.



In terms of S/N, the difference between 1x3600 seconds and 6x600 seconds is the read noise of the camera.
You may also be giving up something by using median combine instead of a combination of statistical rejection
and mean combine – but the specifics of that are less clear to me.



From: main@ap-gto.groups.io <main@ap-gto.groups.io> On Behalf Of uncarollo2 <chris1011@aol.com> via
groups.io
Sent: Tuesday, August 18, 2020 2:53 PM
To: main@ap-gto.groups.io
Subject: Re: [ap-gto] Exposure comparisons with CCD cameras





If you were shooting RGB, or if you were using a camera with very low read noise, I’m guessing that the final
comparison between the two results would be much closer.

I'm not shooting RGB, I'm shooting NB where the 10 min object signal is at 700 or so out of 65,000. And the
faintest detail in the thin wisps next to the brightest part are on the order of 160 in that exposure time. In the 1
hour exposure they are 6 times higher, which is still low compared to the max level before saturation. I could go
10 hours before reaching saturation of the brightest part of the nebula.



I was not trying to compare RGB, only Narrowband for extremely faint objects. I cannot do RGB here due to
massive light pollution, so it's a moot point. I'm simply saying that for very faint stuff, long exposure NB has an
advantage. It's not only me, but others have shown similar results. The noise in the CCD is the same for both
exposures - that is, the noise produced by the camera is high for both images. The download noise may be the
limiting factor, so therefore it behooves to increase the exposure time and thus minimize the number of
downloads if you want to capture the faintest possible detail.



By the way, I have changed the image posted on AstroBin with some more info.

https://www.astrobin.com/916uf7/C/



Rolando





-----Original Message-----
From: W Hilmo <y.groups@hilmo.net>
To: main@ap-gto.groups.io
Sent: Tue, Aug 18, 2020 4:06 pm
Subject: Re: [ap-gto] Exposure comparisons with CCD cameras

I think that to understand what’s going on, you need to separate read noise from shot noise.



In the single 60 minute image, there is 60 minutes worth of signal, 60 minutes worth of shot noise and 1
instance of read noise. In the 6 x 10 image, there is 60 minutes worth of signal, 60 minutes worth of shot noise
and 6 instances of read noise. In theory, I believe that the math suggests that with a zero read noise camera,
there would be no difference in S/N between the two final images. If I remember correctly, you are using a
camera with a KAF-8300 CCD, which has pretty high read noise. Also, you are imaging a narrow band object,
so the signal level in the faint areas is very low.



If you were shooting RGB, or if you were using a camera with very low read noise, I’m guessing that the final
comparison between the two results would be much closer.



From: main@ap-gto.groups.io <main@ap-gto.groups.io> On Behalf Of uncarollo2 <chris1011@aol.com> via
groups.io
Sent: Tuesday, August 18, 2020 1:56 PM
To: main@ap-gto.groups.io
Subject: Re: [ap-gto] Exposure comparisons with CCD cameras



Aha, thank you for the explanation.



Rolando







-----Original Message-----
From: sbasprez via groups.io <beneckerus=aol.com@groups.io>
To: main@ap-gto.groups.io
Sent: Tue, Aug 18, 2020 3:52 pm
Subject: Re: [ap-gto] Exposure comparisons with CCD cameras

The lower background noise floor in the 60 minute image is easily explained mathematically. Noise in a sum
of stacked of images is increased by the square root of the number of subs stacked. So summing the stack of
10 subs results in 3.16 time the noise of a single sub. The signal on the other hand adds linearly. So the
signal in the summed stack is equal to the 60 minute exposure, but the noise floor in the stack is higher than
the 60 minute single frame.



Join main@ap-gto.groups.io to automatically receive all group messages.