Short exposures vs long?

[XCalRocketMan]

Since taking a bias frame is relatively easy. I would go that route and just give APLab a bias frame as opposed to finding a work-around.

JT

Agreed. And I have BIAS frames from a few months ago. It's the inability of the flats to get loaded that really keeps me from using it.

[JayTee]

I'm about to do the same thing. In other words, I'm just trying to get APLab to read one of my light frames. So I'm fudging every other frame. Dark, Bias, and Flat. I'm willing to bet that the results are not going to be vastly different than if I had taken all the images during the same imaging session -- I think!

JT

[XCalRocketMan]

I'm about to do the same thing. In other words, I'm just trying to get APLab to read one of my light frames. So I'm fudging every other frame. Dark, Bias, and Flat. I'm willing to bet that the results are not going to be vastly different than if I had taken all the images during the same imaging session -- I think!

JT

I'm finding that APLab seems to have issues with loading files. My last attempt it had a problem loading my light frame. I restarted it, and then it loaded the light but not the darks. Something just ain't right in river city!

[JayTee]

How about doing an uninstall then reinstall? Might work???

[STEVE333]

I used to use APLab a lot and it worked well. However, a while back I had some issues and tried to contact the author. He never got back to me. I'm worried that it isn't being supported. Too bad. it has some nice features when it works.

The lack of support is actually the reason I made my own program for my own use. I know the author (me) and I usually listen to me .

Steve

[XCalRocketMan]

I used to use APLab a lot and it worked well. However, a while back I had some issues and tried to contact the author. He never got back to me. I'm worried that it isn't being supported. Too bad. it has some nice features when it works.

The lack of support is actually the reason I made my own program for my own use. I know the author (me) and I usually listen to me .

Steve

LOL. As a programmer of over 45 years, at NASA and then at government agencies, I could do that - but, alas, the time! Right now I'm working at home doing data mining and statistical analysis using R for the DoD.

It does look like APLab is not being supported. The last build was a number of years ago. I'll try a re-install and see what happens.

Mikey

[Greenman]

From a sensor viewpoint (CMOS or CCD) noise is the limiting factor on exposure time.

Having many short exposures does improve the signal by averaging; so the signal/noise raises. As always there are limiting factors to this, most likely the sampling rate (pragmatically). However, on a low noise (typically cooled) detector [say -30C from ambient] exposure time will increase signal significantly. In this case the argument fades as the noise is low and shot noise is easily removed (dark pixels).

On the most sensitive Raman spectrometer, I sell (Lab not Astro) we can integrate for up to an hour easily [dual stage Peltier on a back-thinned array detector].

[SkyHiker]

From a sensor viewpoint (CMOS or CCD) noise is the limiting factor on exposure time.

The exposure time is usually limited by when the stars blow out - the stars are usually brighter than the noise.

Having many short exposures does improve the signal by averaging; so the signal/noise raises. As always there are limiting factors to this, most likely the sampling rate (pragmatically).

For Poisson noise it does not matter if you have many short exposures or few long exposures so long as the integration time is the same. In that case the SNR is the same. Did you have any other kind of noise in mind?

However, on a low noise (typically cooled) detector [say -30C from ambient] exposure time will increase signal significantly. In this case the argument fades as the noise is low and shot noise is easily removed (dark pixels).

Shot noise is the same as Poisson noise, it has nothing to do with dark pixels.

[Greenman]

From a sensor viewpoint (CMOS or CCD) noise is the limiting factor on exposure time.

The exposure time is usually limited by when the stars blow out - the stars are usually brighter than the noise.

Having many short exposures does improve the signal by averaging; so the signal/noise raises. As always there are limiting factors to this, most likely the sampling rate (pragmatically).

For Poisson noise it does not matter if you have many short exposures or few long exposures so long as the integration time is the same. In that case the SNR is the same. Did you have any other kind of noise in mind?

However, on a low noise (typically cooled) detector [say -30C from ambient] exposure time will increase signal significantly. In this case the argument fades as the noise is low and shot noise is easily removed (dark pixels).

Shot noise is the same as Poisson noise, it has nothing to do with dark pixels.

Hey SkyHiker, thanks for the response. One at a time below as I'm too lazy to intersperse responses.

Good to remember that I am talking about sensors/detectors, and it will be the loss signal to noise that prevents you collecting, if you don’t saturate the detector - as you say blow out the star.

Missed the point, averaging more data improves the statistics, and that in turn averages both the noise and the signal, but will have a greater impact on the signal. I can pull data out from a single acquisition, by simply increasing the number of replicates.

Got me there, I conflated terms. Dark pixels remove electronic noise including some shot noise inherent in the circuit. It will do nothing about photon shot noise at the detector.

When I am working with a detector, I am usually using them for referenced quantitation, not qualitative work. For that reason noise is the main factor as it will affect the limit of detection. In an image it would reveal itself by loss of resolution in the weakest parts of the image.

I did wonder about responding at all as my area does not exactly mesh with astronomy, but I thought what the heck, it may be useful.

[STEVE333]

I used to use APLab a lot and it worked well. However, a while back I had some issues and tried to contact the author. He never got back to me. I'm worried that it isn't being supported. Too bad. it has some nice features when it works.

The lack of support is actually the reason I made my own program for my own use. I know the author (me) and I usually listen to me .

Steve

LOL. As a programmer of over 45 years, at NASA and then at government agencies, I could do that - but, alas, the time! Right now I'm working at home doing data mining and statistical analysis using R for the DoD.

It does look like APLab is not being supported. The last build was a number of years ago. I'll try a re-install and see what happens.

Mikey

Don't know what "R" is, but, sounds like you are very busy. Having the time to work on little projects like this makes me really appreciate being retired.

The worksheet I have made doesn't do everything that APLab does. I take advantage of the nice information already available with the ZWO cameras and use it to create the graphs I showed here. If you would like an analysis for any of your ZWO cameras I would be glad to do it. All I need is one calibrated frame for each channel you want analyzed. I use the calibrated frame to determine the SkyGlow.

Steve

[Larry 1969]

This thread took off in a few directions but do you know what I find most interesting?

The varying opinions of some VERY capable imagers... I've seen your posts and a lot of you have breathtaking images!
I was under the impression that calculating the perfect gain, offset and exposure time was a science and wouldn't allow for opinion...
Boy, was I wrong! :smile:

I suppose there is no substitute for trial and error... See what works best for you in your unique conditions...

I did develop a decent understanding of the electronic process of imaging as far as what is actually going on inside the camera and that helps.

Thank you all!

Larry

[JayTee]

Our hobby is referred to as the "Art" of astrophotography not so much the science of astrophotography!

Artfully,
JT

[Larry 1969]

Our hobby is referred to as the "Art" of astrophotography not so much the science of astrophotography!

Artfully,
JT

That's true I suppose but I always thought of the artistic part being people applying their preferred colors or "starless backgrounds" and such.
Not so much the actual capturing of photons...

This has been an informative thread for me for sure!

Hey Steve!
I may still send you a light and dark frame once I get my filter and a night out if that's OK.

Thanks!

Larry

[STEVE333]

Hey Steve!
I may still send you a light and dark frame once I get my filter and a night out if that's OK.

Thanks!

Larry

That's fine Larry. If you wanted to send along a light and dark frame from your present setup it would let me work out the process for working with OSC data. Then I would be ready to go when you send the new data.

Just a thought. If you want to wait that is fine too.

Steve

[XCalRocketMan]

I used to use APLab a lot and it worked well. However, a while back I had some issues and tried to contact the author. He never got back to me. I'm worried that it isn't being supported. Too bad. it has some nice features when it works.

The lack of support is actually the reason I made my own program for my own use. I know the author (me) and I usually listen to me .

Steve

LOL. As a programmer of over 45 years, at NASA and then at government agencies, I could do that - but, alas, the time! Right now I'm working at home doing data mining and statistical analysis using R for the DoD.

It does look like APLab is not being supported. The last build was a number of years ago. I'll try a re-install and see what happens.

Mikey

Don't know what "R" is, but, sounds like you are very busy. Having the time to work on little projects like this makes me really appreciate being retired.

The worksheet I have made doesn't do everything that APLab does. I take advantage of the nice information already available with the ZWO cameras and use it to create the graphs I showed here. If you would like an analysis for any of your ZWO cameras I would be glad to do it. All I need is one calibrated frame for each channel you want analyzed. I use the calibrated frame to determine the SkyGlow.

Steve

Appreciate the offer. I'll consider it if I can't get APLab to behave.
R is a programming language specifically developed to handle large data sets and analyze them in a statistical way.

[Stuart]

I don't have a CMOS camera, but a CCD, so this answer may not be relevant. I do my exposure time not based upon the sky background glow, but my statistical ADU count. I'll take an image and if there are saturated pixels, say from a bright star, then I decrease the exposure time.

[STEVE333]

I don't have a CMOS camera, but a CCD, so this answer may not be relevant. I do my exposure time not based upon the sky background glow, but my statistical ADU count. I'll take an image and if there are saturated pixels, say from a bright star, then I decrease the exposure time.

Hi Stuart - Your approach makes perfect sense to ensure no saturated stars.

Since I'm doing NB images now, having saturated stars isn't a big concern. My focus is on reducing the noise in the low-level nebulosity and trying to bring out faint details. That is why my focus is on optimizing the Signal-to-Noise.

As always there is a tradeoff. I guess which choice one makes depends on what their final goal is. Wish I knew the answer!!

By the way, the reason I look for the mean ADU in an image (usually determined by the sky background glow) is to determine the Sky Glow current (e-/sec). The noise in an image comes from three contributions: (1) Read Noise (the random noise added to each pixel when the signal level is measured), (2) Dark Current noise (the shot noise in the Dark Current), and (3) Sky Glow noise (the shot noise in the Sky Glow current). One measure of the data quality is the Signal-to-Noise Ratio or SNR. A higher SNR implies better data. In mathematical form the SNR for a single pixel can be expressed as:



The important thing from this equation is that as T (the Exposure Time for a single image) increases
the SNR increases. This improvement continues until the sum of the Dark Current and Sky Glow noises are much greater than the Read Noise. At that point the improvement in SNR is minimal. Knowing the value of the Sky Glow current is thus important to determine how long an Exposure Time is required to "optimize" the SNR.

Hope this makes some sense.

Steve

[UlteriorModem]

Sometimes a saturated star simply cannot be avoided. The horsehead nebula and Altinak comes to mind.

[Stuart]

Hi Steve,
I agree. For narrowband I ended up just deciding on 15 minute subs. I could go longer, but I worry about blowing a sub because of a wind gust or vibration or whatever. I don't get all mathy in my calculations. I just do 15 minutes for narrowband, 10 minutes for broadband if I'm doing a nebula, and 1-3 minutes for star clusters.
Stu

And Tom, I agree. Altinak is out of control. But when I'm doing star clusters I stay short because I want to try to collect as much color as I can in the stars. I also don't do lum when I'm doing star clusters.

[Stuart]

Here is an M13 I did recently as a camera test where I used 1 minute subs.