To try everything Brilliant has to offer-free-for a full 30 days, visit brilliant.org/cuivlazygeek/. The first 200 of you will get 20% off Brilliant’s annual premium subscription.
The effect of the AI is dramatic, but is it simply comparing your exposures to high grade professional images and blending them with your image? You have no way of telling, unless you have access to the AI code so that you can ensure that it isn't cheating.
Nice to see you going back to basics Cuiv. With seestar and Dwarf taking off there are a lot of new potential astrophotographers out there who are completely unaware of this sort of information. Great work.
Excellent work Cuiv! I never tire of hearing explanations of this topic. Like celestial mechanics, it feels simple once you get it, but it's somehow difficult to explain simply. Great job!
Thanks so much Nico! You are the OG of this discussion on TH-cam and we are all grateful to your explanations that started the ball rolling on the platform for this particular topic!
Your analogies used in your explanations are spot on. I just purchased an S50 as my first. Building my knowledge base from watching you will help me squeeze as much product as possible from the S50 before I move on to bigger, better, faster HW. My journey has just begun. Thanks for your help.
The hole has begun. Besides the scope, I've already added another $250 of trinkets. Not complaining, this is going to be fun. I need to get smarter on back focus. Finding out what it will be exactly without experimenting. If possible. Understanding the effect of filters, but is the 1/3 rule still valid with today's new ones? I'm working on it.
I love your math explaining, so clear and spot on. That's extremely hard to find, even on astrophotography channels: so much confusion, wrong procedures, wrong reasoning, wrong assumptions. Please continue with these nerdy videos, lots to learn from you!! (I have a math degree and struggle a lot with all the poor math floating around these subjects, your channel is an oasis!)
Thanks so much for this feedback! It really means a lot to me :) my more "mathy" videos typically get fewer views and require a lot of preparation and work so it's awesome to see them being useful!
Another excellent video, Cuiv! And something that might be worth mentioning when it comes to exposure time and noise is how the telescope's f ratio is involved. An f/7 refractor is 3 times slower than an f/4 newt. That mean if you take 1 hour of exposures with both scopes the faster newt will have almost twice as much signal to noise as the slower f/7 refractor! (1.73 times as much signal to noise to be more precise since the square root of 3, f/4 being 3 times as fast as f/7, is ~1.73).
Yep, absolutely! Although it's worth mentioning as well that if you had two refractors with the same aperture, but one at F2 and the other at F4, after one hour of exposure you'd of course have 4 times more signal and 2 times more noise, so two times more SNR on the F2 refractor (same camera, etc) BUT I can resample my F4 refractor image to match the pixel scale of the F2 image, and in theory, I'd get the same SNR for étendue... But the F4 image will have a much smaller FOV :) So it can also be accurate to say that the F2 focal ratio only buys me more FOV, while the F4 focal ratio let me trade SNR for more target details on a smaller FOV :) Apologies if I made mistakes, it's early in the morning here :D
Cuiv, you must have been a theater major in college…..your descriptions and expressions are so animated, so energetic. My wife has no interest in this hobby but enjoys hearing you teach…..and so do I!
Great video Cuiv! Drives the point home that you can take shorter exposures to beat light pollution but still need the equivalent integration time. And unfortunately it doubles on itself until it is just ludicrous how much more time you need. Love your enthusiasm in this video, keep it up! Clear Skies!
Aw mate what an excellent video this is, incredibly well done work! :-) You're helping so many people! (thank you for the mention by the way, that is really kind of you!)
As of late, several videos explaining this were made by othets. Although thorough, they didn't stick with me. Your undeniable enthusiasm, and plain and "simple" approach made me realize that more isn't linearly better. Your genuine enthusiasm reminded me of why I love Carl Sagan.
Way to go Cuiv! Yet another super interesting presentation. I love your play acting the frustrated imager at around 4:25. But he need not worry, you give all the answers in the rest of the video!
I just love how you explain everything Cuiv. This analogy was as simplistic as capturing photons can be for anyone starting out, but it was still truly amazing to watch the SNR disappearing with each doubling. You did all the work for us if we were even curious. I saw a question in a forum as to why we don't use one photo and replicate it 100s of times then stack. They need to see this video. Thanks a always for another great video. ❤
Thanks Cuiv, motivation a bit low at moment partly because of high light pollution. Your channel is one of of the ones I go to when I need to be reminded why I want to do this time consuming expensive activity. Have some clear weather… can’t wait for moon to go away for a bit.
I think this topic really highlights the benefit of a faster scope in light polluted areas. His F4 newt or the F2 hyperstar. A nice dark zone an astrophotographer may get away with an F7 or F6 scope.
Great shot cuiv!! Few nights back I have taken, 400 2 second exposure, on the Eskimo Nebula (NGC 2392) came out really impressive even from a white Zone!
Thanks Cuiv, Well presented! I also live in a heavily light polluted city and use Hyperstar which makes a huge difference for collecting short exposures.
I think with Hyperstar the point of diminishing returns may be reached much faster. One other benefit of many short exposures is something that you alluded to early on in your presentation. With a long exposure the wind or some other factor can ruin your image which results in lost time etc. With short exposures, you have the opportunity to discard images that are not of the best quality because discarding these images would not normally negatively impact your processed image. Something like Blink in PixInsight works well but can be tedious when dealing with large data sets. Setting your parameters for what is an acceptable threshold can be valuable and then just Blink the rejected images to see how bad bad is. @@CuivTheLazyGeek
Haha thanks Marvin! I've rejected so many other offers, from Skillshare to Surfshark because I don't really believe in them. At least with Brilliant I can be truthful!
Hi Cuiv, thanks for that great info. Hopefully I'll be out tonight doing a first light test on my new scope tonight, full moon or not, just love imagining.
I ordered a Bahtinov mask and lens cover for my S50 from lukomatico based on your recommendation and experience. I’m looking forward to seeing the difference the mask will make in my images. Thank you for your excellent work and enthusiastic presentation. Salutations du Canada! 😎🇨🇦
Cuiv, Lazy Geek Great video on this process! I have gained a better understanding how to get rid of noise Awesome work! brilliant looks very cool too, looks like it keeps the mind computing. Thanks for sharing! Clear skies Cuiv 🖖
I may have overlooked this detail, but what lens/camera did you use? Am I able to capture this with a 135mm lens on a full frame camera? What focal length do you recommend? Thanks!
I would be interested to see a comparison of each of your stacks to an equivalent single exposure (i.e. compare the _4 exp stack to a single 80s exposure). It would really hit home the effect of light pollution... I would also be interested in this same technique in a less light-polluted sky...
@CuivTheLazyGeek long time viewer, first time commenting, love your videos, thank you for explaining in detail the fundamentals of stacking images! Would it be possible in a follow up video to go over and conpare the different algorithms for stacking at some point? Average, median, sigma-clipped average, etc? Thanks for the tip on AI stacking though, ill definately check that out!
I was watching this, and I began to wonder now with the full moon: is exposure of galaxies with the full moon "worth it"? For example if I have lets say 4 hours of exposure on M81 under a full moon, can I add that to other 60 minutes of exposure with no moon? Will it add any benifit overall? Or am I better to just stack the "dark" 60 minutes and not add the 4 hours of data taken during the full moon, as it may induce more noise than usable signal?
It's a very good question - and I personally always avoid mixing full moon data with the rest, even from Tokyo. PixInsight WBPP can also help to lower the weight allocated to images with worse SNR! Thanks so much for the support by the way!
Thanks Cuiv - gives me hope for non-NB targets from the city. One thing i was keen to find out though is how you stack so many images without crashing your computer, do you do them all at once at leave it overnight, or in batches as I've read online some people do? If you've covered this in another video please let me know, otherwise would love to learn about this in a future vid. Thanks again!
1024/3/60 is about 5.6 hours of exposure. I am interested in how much further is reasonable. I see people reporting 50 hours on the Medula Nebula over 2 months and am jealous. I live in a comparably light polluted city (Bortle 8 - 9) and have run some experiments. I took four nights of exposures for both the Crescent and Elephant Trunk nebulae and found some improvement on the second night, but could not see much after that. I was shooting with Optolong L'eXtreme at 300s at f/5.5. For the Crescent this was (41 + 54 + 62 + 51)/12 = 17:20 total time. (Dither every shot)
This is why I take so many exposures even from dark skies, my last dark sky attempt at Pleiades I took 72 exposures with drizzle per frame (I don't guide so no guide settle time)
Great video! I've always wandered about the real effects of stacking different quantities of picture. It would be good to see a similar video with different exposure times! Instead of quadrupling the number of pictures taken, quadruple the exposure!
Wow, what an interesting video. If I recall correctly, you made a video in the past explaining the stacking process and its benefits using Excel (if I am not mistaken). By the way: you are crazy productive for a "lazy" Geek!
At Marker 28:37 bottom middle image , why do stars have spike flares ? Sorry if wrong wording but that’s the only thing I could think. Thanks for all the information you supply and hard work you put into your videos :-)
You guys in big cities are so lucky! You don't have to think up new targets all the time since you get to image the same areas of the sky time and time again and get a noticeable improvement every time 😅
Hey cuiv! Very well done. I tried capturing the same object yesterday frim Vienna (bortel6) using a newton scope at f3 the traditional way (longer exposures) + the new antlia quad band filter. But no chance that way to get proper reflection nebula out of the stacked & streched image. What equipment did you use in your example with 20sec exposures? Your 750mm newton, no filter i assume?
I used Hyperstar C6 at F2, but the problem isn't there - I don't think the Antlia Quadband filter is suited for reflection nebulae (good for emission nebulae though), so you'd want to try without any filter (just UV/IR cut filter)
It just makes it very clear why I almost never pass 10min when doing EAA. But typically 2-5min as it allows to both observe details and go easily to next target - so in one hour session I can cover 5-10 objects with ease. EAA as observation from light polluted area is just about to get enough data to observe details you can't see
I was able to get this basic nebulosity from Bortle 8 with 50/180 guide scope and 224MC on my Android tablet withing a minute or two using Open Live Stacker Unfortunately CA was too strong, still was interesting experience.
Great video! Great explanation! And super great experiment. Unfortunately, for us citizens (literally speaking) we do not have that much time (bad weather, air pollution, you name it). So, no, I would rather reserve emission nebulae or galaxies for my monthly dark site weekend. Cheers!
Thanks for this informative video. I have given up photographing broadband targets due to the light pollution, but this certainly gives me some encouragement to try! How do you organize and stack multiple nights of shots? Say 1 night I may get 50 shots, and the next day, I might get 30 shots, potentially with different parameters. I feel mixing them up is probably not a good idea. What can I do to maximize the SNR of the final stacked image?
I know that this video demonstrates the difference in nebulosity that stacking reveals. However, I am more impressed by how much the stacking darkens the empty space and reveals the fainter stars in the background! How random is typical light pollution? It seems to me that non-moving light sources would tend to be regular and not susceptible to stacking reduction...
You can consider light pollution as a normal distribution with a mean of LPS (light pollution signal) and a standard deviation of the square root of LPS. That's the part that determines the randomness of LP, a.k.a. the LP shot noise. You can easily remove LPS (effectively gradient subtraction does that), but the randomness remains, and is larger the more LP you have because it is a square root
Now i know that for my “near real time” observation EAA 16 exposures is the sweet spot! Really helpful information! What filter were you using for the Pleiades?
Well it does depend on the length of each sub exposure! 16 subs of 1 second and 16 subs of 30 seconds are not the same :) in fact the latter has 30 times more signal and 5.5 times better SNR :-)
Great job Cuiv! Came up with this simple equation, would you approve? x(a) = x²(sub-frame exposure time) where a is the SNR of one sub-frame (considering only shot noise)
Hey, amazing video. So for me to be able to try such with dslr id need to use about 150mm f5, canon 600d, 20-60sc, bortle 6, iso 1600 on my SKW star adventurer gti. Tracked
One thing that comes up on forums a lot. You CANT improve your SNR by taking more subs IN THE SAME AMOUNT OF TIME. For example, taking 60 one minute subs is not better for SNR than taking 30 2min subs. Both are 1hr total. It's only the total time that works. there are always some eager beginners that get it confused.
Interesting try with the buckets, but the key is not adding more signal (it is indeed an average we get through stacking). The buckets don’t help with explaining the random noise averaging. Very nice idea though. Great video!
They are a great analogy for the difference between single short exposures and long exposures. It would be very informative to do a video on the difference in stacking short exposures and taking long exposures (which actually do sum the signal!)
Erm no that's incorrect..., Averaging and summing are effectively the same thing - there's just an integer factor between them and you can get from one to the other at will - both will represent the exact same SNR. Adding to the large buckets will indeed reduce the impact of the noise compared to the total signal :-)
Perhaps a good way to visualize it is that after you poured all little buckets into a big bucket, you mix the big bucket and pour out exactly one little bucket's worth. That would represent stacking.
thank you for all your efforts............ Question: haven't I seen videos indicating that after abt 20-25 images, the reduction in noise starts to have diminishing returns. Maybe, by Dr. Glover???
I would like to see what the small and cheap Seestar S50 can do with 4 hours of data! so far I've only seen people using it for like 30 min, so maybe with 2 hours or 8 hours we can start seeing amazing stuff even with such a small telescope; because also the main advantage is that it's so easy and automatic to use that you can just leave it on your balcony for like 4 hours each night and collect the data with minimum effort, and that's why I'm surprised no one has done it yet!
Just to add to my confusion based on another video of yours regarding read noise and exposures taken with astro cameras. If I take a bunch of short exposure photos, I have additive read noise that's averaged out from the SNR. However, if I never get signal of super faint data over the read noise, how is that the same as a longer exposure that does have super faint data that exceed the read noise?
This is actually a very interesting topic! If there is signal so faint that the signal itself (and not just its shot noise) is smaller than the read noise (which is effectively a standard deviation), it becomes almost impossible to disentangle them regardless of how many exposures you take. In this case, longer exposures will help get that faint signal over the threshold!
Good explanations, thank you very much, but i have 1 question: when i take 50 frames of 120s, is it exactly the same if I take 120 frames of 50 seconds ? theoretically the to 2 situations produces a stacked image of 6000s (100 minute)
It isn't exactly the same because of the sensor read noise, but in practice it is exactly the same, unless you're in a very very dark area imaging a faint target, so the read noise becomes a large contributor.
Cuiv, with image stacking short exposures like this, how do you deal with walking noise? Is it even a problem? Also, do you change your gain or offset?
So nothing to cut the visible light pollution. Just keep the subs short. With an f6 Refractor, how long would my subs need to be for the same exposure? Would an optolong L-enhance be too aggressive?
I would go ahead and try to double your signal one more time. Look how much time you’ve had involved in some of your past reviews on hardware. I think it would make for an interesting video. We will wait. We all know how long it will take to collect that much more data. Probably an entire season.
Just a random thought that popped into my head regarding SNR... If you have 1000 photos.. instead of stacked them all together, what if you stacked 2 photos, and 2 photos, and 2 photos, 500 times, wouldn't this mean that you doubled the SNR in each of them? And then you could do the same with the 500 stacked photos, stack 2 of them together and have 250 photos with twice the SNR again? And repeat.. I haven't really thought this through yet, but the idea popped into my head because it just sounds wrong that you need to take another 500 photos (if you've already taken 500) just to double the SNR. But then again, I suppose after taking 500 photos, you are doubling the SNR of those (hypothetically) stacked photos.
No no no, in the end only the total exposure time counts - stacking sub exposures is just the tool that lets us reach that goal. Effectively the rule is quadrupling the exposure time doubles the SNR! For subs of the same length it's the same of course, but you can't magically create signal to noise ratio via stacking strategies
By my rough calculations that ends up being 5 1/2 hours of exposure (probably more due to images being discarded). So far I've banked only about 15 minutes worth of exposures (10 seconds each). If the weather ever clears up I plan on getting at least 2 hours of exposures. Is there a way to get the Seestar s50 to restart imaging on another day (i.e. keep stacking from a previous day's imaging run)? I'd love to see what it could do before resorting to Pixinsight. .... or maybe it does that automatically and I just haven't noticed?
I've had the option on the Seestar to keep stacking on the same target during the same night but not across several nights as far as I can remember....
Other than the first exposure, did you dither between images? I suspect that it would probably double your total length to shoot (not the exposure time, but the total time), but it might remove some of the noise. I am not currently set up to do any astrophotography at this point, but I still would be interested in hearing what you think about that.
I dither every 15 frames, which is enough in my case (every 5 minutes effectively). Dithering doesn't help that much with random noise, but it does with fixed pattern noise
Great presentation. 20s is definetly a short exsposure. But what is the criteria for 20s definition. Why not 10s or 30s. Do yo dither ? If so, how long give you after dithering to go with the exposure? Thanks
Dither every 15 frames. 20 seconds because it more than swamps read noise for my level of LP and I didn't want to saturate the bright stars of the cluster too much
Thanks for the details. Unlike you I'm thruly lazy. I use OSC camera and shoot no more than 3-4 hrs per night. My exposure times does from 5 to 10 min to save procesing time. But I see the benefits of the short exposure approach. Thanks for reply!
They're a great example of having fairly bright and faint nebulosity that can't be captured in narrowband - and I had many exposures of these, so why not??
Go and live to Chile, Arizona or else!! Anounce crowd funding to make this move - and I will be first to support you! Because I want it myself - but have so much anchors which prevent me from doing it - or in other words - don't have guts..
Hmmm...for your single image at 20 secs, that's from within Tokyo? Not much light pollution at all. I live in a smaller city then Tokyo and could not get that darkness\blackness at 20 secs. It would be a good washout at 20 secs. Did you use calibration images for your examples?
Oh the individual frames appear washed out until the mean of the light pollution is removed via gradient extraction of course (I'm only showing after that)
@@CuivTheLazyGeek , thanks for the response. Any chance of sharing one of the original single shots. Curious as to what the single frame looked like. How much wash out can you get away with is the million dollar question. Thanks again.
To try everything Brilliant has to offer-free-for a full 30 days, visit brilliant.org/cuivlazygeek/. The first 200 of you will get 20% off Brilliant’s annual premium subscription.
The effect of the AI is dramatic, but is it simply comparing your exposures to high grade professional images and blending them with your image? You have no way of telling, unless you have access to the AI code so that you can ensure that it isn't cheating.
Nice to see you going back to basics Cuiv. With seestar and Dwarf taking off there are a lot of new potential astrophotographers out there who are completely unaware of this sort of information. Great work.
Thanks man! I really hope it can help many!
Excellent work Cuiv! I never tire of hearing explanations of this topic. Like celestial mechanics, it feels simple once you get it, but it's somehow difficult to explain simply. Great job!
Thanks so much Nico! You are the OG of this discussion on TH-cam and we are all grateful to your explanations that started the ball rolling on the platform for this particular topic!
Your analogies used in your explanations are spot on. I just purchased an S50 as my first. Building my knowledge base from watching you will help me squeeze as much product as possible from the S50 before I move on to bigger, better, faster HW. My journey has just begun. Thanks for your help.
Welcome to the hobby!! The black hole runs deep :D
The hole has begun. Besides the scope, I've already added another $250 of trinkets. Not complaining, this is going to be fun. I need to get smarter on back focus. Finding out what it will be exactly without experimenting. If possible. Understanding the effect of filters, but is the 1/3 rule still valid with today's new ones? I'm working on it.
Huh!
I love your math explaining, so clear and spot on. That's extremely hard to find, even on astrophotography channels: so much confusion, wrong procedures, wrong reasoning, wrong assumptions. Please continue with these nerdy videos, lots to learn from you!!
(I have a math degree and struggle a lot with all the poor math floating around these subjects, your channel is an oasis!)
Thanks so much for this feedback! It really means a lot to me :) my more "mathy" videos typically get fewer views and require a lot of preparation and work so it's awesome to see them being useful!
Another excellent video, Cuiv! And something that might be worth mentioning when it comes to exposure time and noise is how the telescope's f ratio is involved. An f/7 refractor is 3 times slower than an f/4 newt. That mean if you take 1 hour of exposures with both scopes the faster newt will have almost twice as much signal to noise as the slower f/7 refractor! (1.73 times as much signal to noise to be more precise since the square root of 3, f/4 being 3 times as fast as f/7, is ~1.73).
Yep, absolutely! Although it's worth mentioning as well that if you had two refractors with the same aperture, but one at F2 and the other at F4, after one hour of exposure you'd of course have 4 times more signal and 2 times more noise, so two times more SNR on the F2 refractor (same camera, etc)
BUT I can resample my F4 refractor image to match the pixel scale of the F2 image, and in theory, I'd get the same SNR for étendue... But the F4 image will have a much smaller FOV :)
So it can also be accurate to say that the F2 focal ratio only buys me more FOV, while the F4 focal ratio let me trade SNR for more target details on a smaller FOV :)
Apologies if I made mistakes, it's early in the morning here :D
@@CuivTheLazyGeek absolutely, but I want to see you review that f2 refractor, and be sure and tell us how much it costs! 😆
Cuiv, you must have been a theater major in college…..your descriptions and expressions are so animated, so energetic. My wife has no interest in this hobby but enjoys hearing you teach…..and so do I!
Hahaha I wish! I was a maths, physics, and engineering major :D
Great video Cuiv! Drives the point home that you can take shorter exposures to beat light pollution but still need the equivalent integration time. And unfortunately it doubles on itself until it is just ludicrous how much more time you need. Love your enthusiasm in this video, keep it up! Clear Skies!
Exactly Dave, and thanks so much for the support as always!
❤❤❤❤❤ just a bit of love for your hardship of living in light polluted Tokyo and your yet visibly undiminished passion for the hobby
Videos like this are why I think you're one of the best AP guys on YT! Thanks for the great content!
Appreciate that, it means a lot :)
Aw mate what an excellent video this is, incredibly well done work! :-) You're helping so many people! (thank you for the mention by the way, that is really kind of you!)
Cheers Luke! Thanks as always!
A fascinating exploration of the mechanics of s/n - I'm unlikely to forget this lesson. Thank you!
Glad it was helpful!
Awesome job Cuiv. Great to see you helping the community as always.
Much appreciated
As of late, several videos explaining this were made by othets. Although thorough, they didn't stick with me. Your undeniable enthusiasm, and plain and "simple" approach made me realize that more isn't linearly better. Your genuine enthusiasm reminded me of why I love Carl Sagan.
That is an incredible compliment, thank you so much, and glad this helped!
Way to go Cuiv! Yet another super interesting presentation. I love your play acting the frustrated imager at around 4:25. But he need not worry, you give all the answers in the rest of the video!
Yep, didn't even have to act, I was that guy a few years ago :)
Thanks!
Woohoo, thank you!!
Great video! I love the rain/bucket analogy.
Glad you liked it!
That’s a great in depth explanation of the doubling math. Thx man.
IRS a really nice shot too. You also got a nice showing of UCG 2838!
Thanks!!
Best explanation I have seen. Happy to see you so enthusiastic, and inspiring other lazy geeks to be slightly less lazy!
Yay, thank you!
Brilliant video with fantastically clear and simple explanations. Love it! 👏🏻👏🏻
Glad you liked it!
Why take about stuff
Thanks!
Thank you!
I just love how you explain everything Cuiv. This analogy was as simplistic as capturing photons can be for anyone starting out, but it was still truly amazing to watch the SNR disappearing with each doubling. You did all the work for us if we were even curious. I saw a question in a forum as to why we don't use one photo and replicate it 100s of times then stack. They need to see this video. Thanks a always for another great video. ❤
Thanks Cuiv, motivation a bit low at moment partly because of high light pollution. Your channel is one of of the ones I go to when I need to be reminded why I want to do this time consuming expensive activity. Have some clear weather… can’t wait for moon to go away for a bit.
Oh yeah man, I know the feeling.... Hang in there!! I'm still imaging with the full moon, because then there is no regret if something goes wrong!
I think this topic really highlights the benefit of a faster scope in light polluted areas. His F4 newt or the F2 hyperstar. A nice dark zone an astrophotographer may get away with an F7 or F6 scope.
this was my conclusion as well
Great shot cuiv!! Few nights back I have taken, 400 2 second exposure, on the Eskimo Nebula (NGC 2392) came out really impressive even from a white Zone!
Thanks Cuiv,
Well presented! I also live in a heavily light polluted city and use Hyperstar which makes a huge difference for collecting short exposures.
Indeed! Hyperstar is awesome :-)
I think with Hyperstar the point of diminishing returns may be reached much faster. One other benefit of many short exposures is something that you alluded to early on in your presentation. With a long exposure the wind or some other factor can ruin your image which results in lost time etc. With short exposures, you have the opportunity to discard images that are not of the best quality because discarding these images would not normally negatively impact your processed image. Something like Blink in PixInsight works well but can be tedious when dealing with large data sets. Setting your parameters for what is an acceptable threshold can be valuable and then just Blink the rejected images to see how bad bad is. @@CuivTheLazyGeek
Another great video and explanation of signal collection. You are also getting pretty good at the seqways into your sponsors ads.
Haha thanks Marvin! I've rejected so many other offers, from Skillshare to Surfshark because I don't really believe in them. At least with Brilliant I can be truthful!
Great job! I was aware of the concept - this helped add in more detail.
Glad it was helpful!
Hi Cuiv, thanks for that great info. Hopefully I'll be out tonight doing a first light test on my new scope tonight, full moon or not, just love imagining.
Congrats on the new scope!!! I hope the first light went well!
Great video, Cuiv. I've tried your approach - I got about 300 x 60 sec subs in the time I had - and it really works in improving SNR!
Looked all through from beggining to end
I didn’t know there was that much to see in this area !
This is an amazing summary of all the salient features and is enjoyable by viewers of all skill levels❤
Thanks!
I ordered a Bahtinov mask and lens cover for my S50 from lukomatico based on your recommendation and experience. I’m looking forward to seeing the difference the mask will make in my images. Thank you for your excellent work and enthusiastic presentation. Salutations du Canada! 😎🇨🇦
Great video, f2 was sooo nice.
Yes it was!
Looking forward to watching this but I'll have to come back later - don't have 40 minutes for a video. Excited to watch though!
Awesome, and take your time :D
Fantastic video. Very revealing and informative. I learned a lot. Many thanks!
Thank you so much! It was a difficult video to prepare so it's really helpful to know :)
Amazing content Cuiv!
Glad you think so!
The Subaru !
Yes!
A great astro lecture.
Great video. Thanks. At least you are not in an "always cloudy city" as I am in Glasgow :(
Sorry to hear that... Hope you do get some clear skies!!
WOW……. I learned so much
So glad it helped!
Excellent explanations, Cuiv. Thank you very much!
Cuiv, Lazy Geek Great video on this process! I have gained a better understanding how to get rid of noise Awesome work! brilliant looks very cool too, looks like it keeps the mind computing. Thanks for sharing! Clear skies Cuiv 🖖
I may have overlooked this detail, but what lens/camera did you use?
Am I able to capture this with a 135mm lens on a full frame camera? What focal length do you recommend?
Thanks!
I would be interested to see a comparison of each of your stacks to an equivalent single exposure (i.e. compare the _4 exp stack to a single 80s exposure). It would really hit home the effect of light pollution...
I would also be interested in this same technique in a less light-polluted sky...
You'd see pretty much the same image though....
@@CuivTheLazyGeek but would you need fewer exposures to be equivalent to your 1024 exposure image?
Hey - thanks for the video. Could you provide some technical informations (camera, focal length, with or without filters)?
Hyperstar C6 300mm F2, ASI533 MC Pro, UV/IR cut only
Amazing explanation!
Glad it was helpful!
@CuivTheLazyGeek long time viewer, first time commenting, love your videos, thank you for explaining in detail the fundamentals of stacking images! Would it be possible in a follow up video to go over and conpare the different algorithms for stacking at some point? Average, median, sigma-clipped average, etc? Thanks for the tip on AI stacking though, ill definately check that out!
That sounds like faaaaar too much work lol
Superb!
Thanks a lot!
Nice done!
Thanks!
I was watching this, and I began to wonder now with the full moon: is exposure of galaxies with the full moon "worth it"? For example if I have lets say 4 hours of exposure on M81 under a full moon, can I add that to other 60 minutes of exposure with no moon? Will it add any benifit overall? Or am I better to just stack the "dark" 60 minutes and not add the 4 hours of data taken during the full moon, as it may induce more noise than usable signal?
It's a very good question - and I personally always avoid mixing full moon data with the rest, even from Tokyo. PixInsight WBPP can also help to lower the weight allocated to images with worse SNR!
Thanks so much for the support by the way!
Thanks Cuiv - gives me hope for non-NB targets from the city. One thing i was keen to find out though is how you stack so many images without crashing your computer, do you do them all at once at leave it overnight, or in batches as I've read online some people do? If you've covered this in another video please let me know, otherwise would love to learn about this in a future vid. Thanks again!
1024/3/60 is about 5.6 hours of exposure. I am interested in how much further is reasonable. I see people reporting 50 hours on the Medula Nebula over 2 months and am jealous. I live in a comparably light polluted city (Bortle 8 - 9) and have run some experiments. I took four nights of exposures for both the Crescent and Elephant Trunk nebulae and found some improvement on the second night, but could not see much after that. I was shooting with Optolong L'eXtreme at 300s at f/5.5. For the Crescent this was (41 + 54 + 62 + 51)/12 = 17:20 total time. (Dither every shot)
This is why I take so many exposures even from dark skies, my last dark sky attempt at Pleiades I took 72 exposures with drizzle per frame (I don't guide so no guide settle time)
Great video! I've always wandered about the real effects of stacking different quantities of picture. It would be good to see a similar video with different exposure times! Instead of quadrupling the number of pictures taken, quadruple the exposure!
Wow, what an interesting video. If I recall correctly, you made a video in the past explaining the stacking process and its benefits using Excel (if I am not mistaken).
By the way: you are crazy productive for a "lazy" Geek!
Oh I don't remember all the videos I made tbh lol
At Marker 28:37 bottom middle image , why do stars have spike flares ? Sorry if wrong wording but that’s the only thing I could think. Thanks for all the information you supply and hard work you put into your videos :-)
You guys in big cities are so lucky! You don't have to think up new targets all the time since you get to image the same areas of the sky time and time again and get a noticeable improvement every time 😅
Hahaha sure! Light pollution is great! :D
Thanks Cuiv!
Cheers!
Hey cuiv! Very well done. I tried capturing the same object yesterday frim Vienna (bortel6) using a newton scope at f3 the traditional way (longer exposures) + the new antlia quad band filter. But no chance that way to get proper reflection nebula out of the stacked & streched image. What equipment did you use in your example with 20sec exposures? Your 750mm newton, no filter i assume?
I used Hyperstar C6 at F2, but the problem isn't there - I don't think the Antlia Quadband filter is suited for reflection nebulae (good for emission nebulae though), so you'd want to try without any filter (just UV/IR cut filter)
It just makes it very clear why I almost never pass 10min when doing EAA. But typically 2-5min as it allows to both observe details and go easily to next target - so in one hour session I can cover 5-10 objects with ease.
EAA as observation from light polluted area is just about to get enough data to observe details you can't see
I was able to get this basic nebulosity from Bortle 8 with 50/180 guide scope and 224MC on my Android tablet withing a minute or two using Open Live Stacker
Unfortunately CA was too strong, still was interesting experience.
Great video! Great explanation! And super great experiment. Unfortunately, for us citizens (literally speaking) we do not have that much time (bad weather, air pollution, you name it). So, no, I would rather reserve emission nebulae or galaxies for my monthly dark site weekend. Cheers!
Hope you do get better weather :)
Thanks for this informative video. I have given up photographing broadband targets due to the light pollution, but this certainly gives me some encouragement to try!
How do you organize and stack multiple nights of shots? Say 1 night I may get 50 shots, and the next day, I might get 30 shots, potentially with different parameters. I feel mixing them up is probably not a good idea. What can I do to maximize the SNR of the final stacked image?
I basically rely on the image weighting that's done by the PixInsight Weighted Batch Pre Processing Script!
I know that this video demonstrates the difference in nebulosity that stacking reveals. However, I am more impressed by how much the stacking darkens the empty space and reveals the fainter stars in the background! How random is typical light pollution? It seems to me that non-moving light sources would tend to be regular and not susceptible to stacking reduction...
You can consider light pollution as a normal distribution with a mean of LPS (light pollution signal) and a standard deviation of the square root of LPS. That's the part that determines the randomness of LP, a.k.a. the LP shot noise.
You can easily remove LPS (effectively gradient subtraction does that), but the randomness remains, and is larger the more LP you have because it is a square root
Now i know that for my “near real time” observation EAA 16 exposures is the sweet spot! Really helpful information! What filter were you using for the Pleiades?
Well it does depend on the length of each sub exposure! 16 subs of 1 second and 16 subs of 30 seconds are not the same :) in fact the latter has 30 times more signal and 5.5 times better SNR :-)
Yes, but if, for example I am taking 10 second subs as standard, I know not to stop at 15, but do 16.@@CuivTheLazyGeek
Generally in my area, if I can collect 4 hours data I know its going to be good. 2 hours can still be noisy depending of target magnitude.
You must be living in a relatively dark area, I'm jealous!
@@CuivTheLazyGeek yeah nah, its still B7 mate :(
What program you used for stacking? In DSS I get few stars for alignment.
Great job Cuiv! Came up with this simple equation, would you approve? x(a) = x²(sub-frame exposure time) where a is the SNR of one sub-frame (considering only shot noise)
Hey Cuiv, when did you capture this image and what equipment did you use?
Yes, Celestron C6 Hyperstar, ASI533MC Pro, UV/IR cut filter only, last year
Hey, amazing video. So for me to be able to try such with dslr id need to use about 150mm f5, canon 600d, 20-60sc, bortle 6, iso 1600 on my SKW star adventurer gti. Tracked
You probably mentioned it, and I overlooked it, but did you live stack or stack after? Did you use Sharpcap?
I stacked afterwards - I didn't use sharpcap
One thing that comes up on forums a lot. You CANT improve your SNR by taking more subs IN THE SAME AMOUNT OF TIME. For example, taking 60 one minute subs is not better for SNR than taking 30 2min subs. Both are 1hr total.
It's only the total time that works. there are always some eager beginners that get it confused.
Absolutely correct - I should have mentioned this point explicitly in the video...
What is the best way to work out the best sub exposure length for a target, Cuiv? Great video!
Interesting try with the buckets, but the key is not adding more signal (it is indeed an average we get through stacking). The buckets don’t help with explaining the random noise averaging. Very nice idea though. Great video!
They are a great analogy for the difference between single short exposures and long exposures. It would be very informative to do a video on the difference in stacking short exposures and taking long exposures (which actually do sum the signal!)
Erm no that's incorrect..., Averaging and summing are effectively the same thing - there's just an integer factor between them and you can get from one to the other at will - both will represent the exact same SNR. Adding to the large buckets will indeed reduce the impact of the noise compared to the total signal :-)
Perhaps a good way to visualize it is that after you poured all little buckets into a big bucket, you mix the big bucket and pour out exactly one little bucket's worth. That would represent stacking.
thank you for all your efforts............ Question: haven't I seen videos indicating that after abt 20-25 images, the reduction in noise starts to have diminishing returns. Maybe, by Dr. Glover???
I would like to see what the small and cheap Seestar S50 can do with 4 hours of data! so far I've only seen people using it for like 30 min, so maybe with 2 hours or 8 hours we can start seeing amazing stuff even with such a small telescope; because also the main advantage is that it's so easy and automatic to use that you can just leave it on your balcony for like 4 hours each night and collect the data with minimum effort, and that's why I'm surprised no one has done it yet!
Check some other of my videos, since I have done longer exposures in narrowband with the Seestar!
Just to add to my confusion based on another video of yours regarding read noise and exposures taken with astro cameras. If I take a bunch of short exposure photos, I have additive read noise that's averaged out from the SNR. However, if I never get signal of super faint data over the read noise, how is that the same as a longer exposure that does have super faint data that exceed the read noise?
This is actually a very interesting topic! If there is signal so faint that the signal itself (and not just its shot noise) is smaller than the read noise (which is effectively a standard deviation), it becomes almost impossible to disentangle them regardless of how many exposures you take. In this case, longer exposures will help get that faint signal over the threshold!
Yup.
can You plz list the rig details?
Thanks
C6 Hyperstar, CEM60, ASI533MC Pro, UV/IR cut filter
So what do you do if you have an alt/az mount?
You have to accept the loss of FOV - it's still possible to take very long exposures, but the FOV will basically be halved in both dimensions
What problems do you have shooting from the roof of a 3 story buildings?
Wind, wind, wind...
@@CuivTheLazyGeek Do you have vibration problems?
Good explanations, thank you very much, but i have 1 question: when i take 50 frames of 120s, is it exactly the same if I take 120 frames of 50 seconds ? theoretically the to 2 situations produces a stacked image of 6000s (100 minute)
It isn't exactly the same because of the sensor read noise, but in practice it is exactly the same, unless you're in a very very dark area imaging a faint target, so the read noise becomes a large contributor.
Did you dither these short exposures? Are short exposures not subject to crawling noise?
Of course I dither! Every 15 frames if I remember correctly
Great! Thanks for the reply.
hello! did you use any kind of filter to get this target or do you take your picture only in the new moon week? best regards
Cuiv, with image stacking short exposures like this, how do you deal with walking noise? Is it even a problem? Also, do you change your gain or offset?
What filter did you use?
UV/IR cut
So nothing to cut the visible light pollution. Just keep the subs short. With an f6 Refractor, how long would my subs need to be for the same exposure? Would an optolong L-enhance be too aggressive?
What bortle level are you under?
What device did you use to take the pictures? S50? D2"? or something higher?
This was Hyperstar C6
Nice Antila Filter jacket
I would go ahead and try to double your signal one more time. Look how much time you’ve had involved in some of your past reviews on hardware. I think it would make for an interesting video. We will wait. We all know how long it will take to collect that much more data. Probably an entire season.
Just a random thought that popped into my head regarding SNR...
If you have 1000 photos.. instead of stacked them all together, what if you stacked 2 photos, and 2 photos, and 2 photos, 500 times, wouldn't this mean that you doubled the SNR in each of them?
And then you could do the same with the 500 stacked photos, stack 2 of them together and have 250 photos with twice the SNR again? And repeat..
I haven't really thought this through yet, but the idea popped into my head because it just sounds wrong that you need to take another 500 photos (if you've already taken 500) just to double the SNR. But then again, I suppose after taking 500 photos, you are doubling the SNR of those (hypothetically) stacked photos.
No no no, in the end only the total exposure time counts - stacking sub exposures is just the tool that lets us reach that goal. Effectively the rule is quadrupling the exposure time doubles the SNR! For subs of the same length it's the same of course, but you can't magically create signal to noise ratio via stacking strategies
@@CuivTheLazyGeek Yea, I think I already knew that.. but sometimes my brain just won't shup up when it thinks it has a new idea.. 🤣
By my rough calculations that ends up being 5 1/2 hours of exposure (probably more due to images being discarded). So far I've banked only about 15 minutes worth of exposures (10 seconds each). If the weather ever clears up I plan on getting at least 2 hours of exposures.
Is there a way to get the Seestar s50 to restart imaging on another day (i.e. keep stacking from a previous day's imaging run)? I'd love to see what it could do before resorting to Pixinsight. .... or maybe it does that automatically and I just haven't noticed?
I've had the option on the Seestar to keep stacking on the same target during the same night but not across several nights as far as I can remember....
Can you try to eke out IC 349?
Not at that focal length!
Other than the first exposure, did you dither between images? I suspect that it would probably double your total length to shoot (not the exposure time, but the total time), but it might remove some of the noise. I am not currently set up to do any astrophotography at this point, but I still would be interested in hearing what you think about that.
I dither every 15 frames, which is enough in my case (every 5 minutes effectively). Dithering doesn't help that much with random noise, but it does with fixed pattern noise
Great presentation. 20s is definetly a short exsposure. But what is the criteria for 20s definition. Why not 10s or 30s.
Do yo dither ? If so, how long give you after dithering to go with the exposure? Thanks
Dither every 15 frames. 20 seconds because it more than swamps read noise for my level of LP and I didn't want to saturate the bright stars of the cluster too much
Thanks for the details. Unlike you I'm thruly lazy. I use OSC camera and shoot no more than 3-4 hrs per night. My exposure times does from 5 to 10 min to save procesing time. But I see the benefits of the short exposure approach.
Thanks for reply!
SUBARU スバル❤❤❤🎉
Thanks Cuiv. Why Pleiades? Was there a reason? Or am I pushing the boundaries of laziness here?
They're a great example of having fairly bright and faint nebulosity that can't be captured in narrowband - and I had many exposures of these, so why not??
Oh and btw: doubling my video time on you quadruples my understanding of the topic 😎
Thanks so much!
Go and live to Chile, Arizona or else!! Anounce crowd funding to make this move - and I will be first to support you! Because I want it myself - but have so much anchors which prevent me from doing it - or in other words - don't have guts..
Thanks so much!
Hmmm...for your single image at 20 secs, that's from within Tokyo? Not much light pollution at all. I live in a smaller city then Tokyo and could not get that darkness\blackness at 20 secs. It would be a good washout at 20 secs. Did you use calibration images for your examples?
Oh the individual frames appear washed out until the mean of the light pollution is removed via gradient extraction of course (I'm only showing after that)
@@CuivTheLazyGeek , thanks for the response. Any chance of sharing one of the original single shots. Curious as to what the single frame looked like. How much wash out can you get away with is the million dollar question. Thanks again.