Thank you for your videos. I appreciate the science, math, and technicality in your videos as opposed to amateurs publishing their experiences. This is what I love about this channel.
Is it? Robert Glovers video is 3 times longer, and 100 times better. So, what is actually noise? It is light pollution. Rest doesn’t matter. Bottom line, watch Robert Glover and you will learn more.
@@anata5127 Sorry you didn't like the video. Just two things: 1) The goal of Dr. Glover's presentation was to show people how Sharpcap can estimate imaging time/exposure length etc. He was explaining his product and the variables you need to know to use it. The goal of the three videos I produced was to show that people are often too rigid in their beliefs about what stacking does. Really, they're saying the same thing: stacking increases SNR. How you define the randomness of the photons entering your camera (and both ways are equally valid) determines your perspective. People often speak past each other without trying to figure out the other's perspective. They'll go onto cloudynights or RUclips comments and argue about it endlessly calling the other argument "bad" or whatever without taking the time to understand how others are defining important concepts. 2) Did Dr. Glover actually say that nothing else matters besides light pollution? I don't think that was the point of his presentation. He was trying to say that for most of us, the largest source of noise is light pollution. Figuring out how much you have is really important for estimating the "ideal exposure length." But other sources of noise do matter, especially in dark skies.
@@deepskydetail He presented theory that you told in video. The same theory. I am not surprised that you watched his video. Yes, nothing is matter besides light pollution. It is dominant noise. But it is no brainer. Right? It was point of his presentation. Exposure is dictated by noise. Otherwise, we could get Hubble level photos.
@@anata5127 Hmmm... I'm sorry but I'm not quite following your point. The way Dr. Glover presented his information in the talk was the second way of thinking about signal/noise. In it he talks about signal increases linearly, and noise by the square root. He talks about noise as a percentage of total signal by exposure length, but that is still different than the first argument in my video. Specifically, I don't see how he framed it as "noise decreasing with stacking, and signal remaining constant" or maybe we're watching a different video. I agree that for most of us, light pollution is the largest source of noise, but this video really isn't about light pollution. It's about how we conceptualize stacking in general. Again, sorry your found this video unsatisfying! Clear skies!
lol I love these videos. Remember that noise is in a quantum state of being there and simultaneously not being there until you look at it. Maybe we shouldn't look at it at all .... 🤔
Great videos! I love them, even though I sometimes get lost and need to rewatch them. I have a question which has been on my mind for ages, but have never really found the true answer, but I feel in these videos you are hinting at it. So my question is: say I take 100 three minute exposures and stack them. The result isn’t like having a 300 minute exposure, but is really just a three minute exposure with all the noise taken out - is this correct? Because in the stacking an average is taken, if a pixel say had a true value of 50, in my stack it would be pretty well bang on 50, but in a 300 minute exposure there would be a variance. Am I correct in my thinking then, basically a stack of images is what a very low noise single image would look like? (As you said in this video when stacking you don’t increase brightness).
Thanks! I think you've understood the first explanation pretty well. One thing to keep in mind too is that if you have an accurate enough measurement of how many photons hit a pixel every 6 minutes, you can just multiply that by 50 to get a very good estimate of a 5 hour exposure. In that case, enough "noise" in your measurement has been removed to make a reliable prediction. Or you can think of it as just adding up photons/electrons to get a "cumulative exposure." In that case, you have to define noise differently. But either way gets the same answer! So whether you think of it as a very accurate average or cumulative exposure doesn't really matter (that much; the other sources of noise like read noise can add up though and so in very dark skies a longer exposure will usually improve SNR faster; but it usually doesn't matter because of light pollution).
I have only one issue. The word "Average" shouldn't be used. Averages tell you nothing. If I tell you that the average temperature of a room is 22 Deg. C then you could conclude thats a nice comfortable room to be in? However, the room is only ever at two temperature -12 and +32 Deg. C. You should be using the word "Mean" in statistics. This might seem pedantic but a mean is associated with a shape or distribution and a std.Deviation or variance. Finally, you have made a lot of work for yourself with these three video's and one of them could have been skipped because the scientific method states that when two things cannot be distinguished from each other, they are in fact the same thing!!! 🙂 Personally, I'd like to see more on the lucky-imaging of DSO's. At least you used my magic number, 16x sub-frame(time) = 1/4 of the noise.
Oh, I completely agree that "mean" is by far the more accurate term. I decided against it for a couple reasons: (1) I didn't want people to get confused with "mean" as in sum(observations)/n and "mean" as in "meaning" and (2) a lot of stacking software uses "average" as "mean" (e.g., in Siril it is "Average with rejection") so I thought keeping it consistent would help with understanding. Wrt making three videos, originally I was only going to do two. I was going to put this stuff into the last video, but decided it would make the 2nd video too complicated. Plus, if they only watched the 1st half of that video, they would come out thinking that the first video was just completely wrong without getting to the good stuff in video three.
It's just square roots and fractions... And the video needs this mathematics to drive home the point that both perspectives are the same and produce the same calculation for SNR.
The outcome is SNR is really the only thing that matters. You can treat stacking as increasing signal or decreasing noise (not really both). Either way gets you the exact same answer (i.e., signal starts to overpower the noise and you get a better image).
Thank you for your videos. I appreciate the science, math, and technicality in your videos as opposed to amateurs publishing their experiences. This is what I love about this channel.
Glad you like them! Thanks :)
I absolutely Love The Math and Science Behind This!! Impressive video!
Thank you!
Is it? Robert Glovers video is 3 times longer, and 100 times better.
So, what is actually noise? It is light pollution. Rest doesn’t matter.
Bottom line, watch Robert Glover and you will learn more.
@@anata5127 Sorry you didn't like the video. Just two things:
1) The goal of Dr. Glover's presentation was to show people how Sharpcap can estimate imaging time/exposure length etc. He was explaining his product and the variables you need to know to use it. The goal of the three videos I produced was to show that people are often too rigid in their beliefs about what stacking does. Really, they're saying the same thing: stacking increases SNR. How you define the randomness of the photons entering your camera (and both ways are equally valid) determines your perspective. People often speak past each other without trying to figure out the other's perspective. They'll go onto cloudynights or RUclips comments and argue about it endlessly calling the other argument "bad" or whatever without taking the time to understand how others are defining important concepts.
2) Did Dr. Glover actually say that nothing else matters besides light pollution? I don't think that was the point of his presentation. He was trying to say that for most of us, the largest source of noise is light pollution. Figuring out how much you have is really important for estimating the "ideal exposure length." But other sources of noise do matter, especially in dark skies.
@@deepskydetail He presented theory that you told in video. The same theory. I am not surprised that you watched his video.
Yes, nothing is matter besides light pollution. It is dominant noise. But it is no brainer. Right? It was point of his presentation. Exposure is dictated by noise. Otherwise, we could get Hubble level photos.
@@anata5127 Hmmm... I'm sorry but I'm not quite following your point. The way Dr. Glover presented his information in the talk was the second way of thinking about signal/noise. In it he talks about signal increases linearly, and noise by the square root. He talks about noise as a percentage of total signal by exposure length, but that is still different than the first argument in my video. Specifically, I don't see how he framed it as "noise decreasing with stacking, and signal remaining constant" or maybe we're watching a different video. I agree that for most of us, light pollution is the largest source of noise, but this video really isn't about light pollution. It's about how we conceptualize stacking in general.
Again, sorry your found this video unsatisfying! Clear skies!
Gloriously simple in its duplicity...
It is!
My comments in the last video stirred up a hornets nest! Lol. You have a nack for explaining these topics 👍🏻
Thank you! :)
lol I love these videos. Remember that noise is in a quantum state of being there and simultaneously not being there until you look at it. Maybe we shouldn't look at it at all .... 🤔
Maybe! Then we'd never have to worry about noise again ;)
Impressive as usual
Thanks!
Great videos! I love them, even though I sometimes get lost and need to rewatch them.
I have a question which has been on my mind for ages, but have never really found the true answer, but I feel in these videos you are hinting at it. So my question is: say I take 100 three minute exposures and stack them. The result isn’t like having a 300 minute exposure, but is really just a three minute exposure with all the noise taken out - is this correct? Because in the stacking an average is taken, if a pixel say had a true value of 50, in my stack it would be pretty well bang on 50, but in a 300 minute exposure there would be a variance. Am I correct in my thinking then, basically a stack of images is what a very low noise single image would look like? (As you said in this video when stacking you don’t increase brightness).
Thanks! I think you've understood the first explanation pretty well. One thing to keep in mind too is that if you have an accurate enough measurement of how many photons hit a pixel every 6 minutes, you can just multiply that by 50 to get a very good estimate of a 5 hour exposure. In that case, enough "noise" in your measurement has been removed to make a reliable prediction. Or you can think of it as just adding up photons/electrons to get a "cumulative exposure." In that case, you have to define noise differently. But either way gets the same answer!
So whether you think of it as a very accurate average or cumulative exposure doesn't really matter (that much; the other sources of noise like read noise can add up though and so in very dark skies a longer exposure will usually improve SNR faster; but it usually doesn't matter because of light pollution).
Wish I could get some signal. All I get is sensor noise, static noise and walking noise.
Oh no! I hope you can get some signal soon!
Ok so with a Canon 700d and a es 72 on a heq5 mount how does that relate to video? To measure out time of exposure?
You and James lamb channels are very similar in math and experimentation, you should check him out
I have only one issue. The word "Average" shouldn't be used. Averages tell you nothing. If I tell you that the average temperature of a room is 22 Deg. C then you could conclude thats a nice comfortable room to be in? However, the room is only ever at two temperature -12 and +32 Deg. C. You should be using the word "Mean" in statistics. This might seem pedantic but a mean is associated with a shape or distribution and a std.Deviation or variance.
Finally, you have made a lot of work for yourself with these three video's and one of them could have been skipped because the scientific method states that when two things cannot be distinguished from each other, they are in fact the same thing!!! 🙂 Personally, I'd like to see more on the lucky-imaging of DSO's. At least you used my magic number, 16x sub-frame(time) = 1/4 of the noise.
Oh, I completely agree that "mean" is by far the more accurate term. I decided against it for a couple reasons: (1) I didn't want people to get confused with "mean" as in sum(observations)/n and "mean" as in "meaning" and (2) a lot of stacking software uses "average" as "mean" (e.g., in Siril it is "Average with rejection") so I thought keeping it consistent would help with understanding.
Wrt making three videos, originally I was only going to do two. I was going to put this stuff into the last video, but decided it would make the 2nd video too complicated. Plus, if they only watched the 1st half of that video, they would come out thinking that the first video was just completely wrong without getting to the good stuff in video three.
nah, both wrong. stacking decreases signal and increases noise. trust me
😂 I should have seen it all along!
Way too mathematical so in the end what is the outcome as I got all bleary eyed and tuned out as soon as you got all mathematical
It's just square roots and fractions...
And the video needs this mathematics to drive home the point that both perspectives are the same and produce the same calculation for SNR.
The outcome is SNR is really the only thing that matters. You can treat stacking as increasing signal or decreasing noise (not really both). Either way gets you the exact same answer (i.e., signal starts to overpower the noise and you get a better image).
@@scottrobinson4611 speak for yourself mate it's all Greek to me