 Hello there, if you clicked on this video, you're probably wondering why I took 1,024 photos of the Orion Nebula. In astrophotography, we use a technique called image stacking where those 1,024 photos are stacked together into just one photo in software. And this stacking process will bring our image from this, this is just one two second photo, to this, this is 1,024 two second photos stacked together representing a total exposure time of over 30 minutes. And if you've heard the term stacking before and found explanations for it, just completely bewildering, I really think this video is gonna clear it up for you if you really watch it all the way from the beginning to the end. To give you though upfront here my one sentence explanation for image stacking. Here it is, in this video, instead of taking one super long 30 minute exposure of the Nebula, we're gonna break that long exposure up into many shorter sub exposures of two seconds each taken one after another, just keeping the Nebula roughly centered and add up all of those sub exposures together so that in the end we've captured the same amount of light from the Nebula but just broken up into smaller pieces. And that's it, that's stacking in one sentence. And so now you might be thinking, okay, I get that conceptually. We break up the longer exposure into smaller pieces, add them back together to get our final photo. And that's roughly equivalent to just taking one long photo. But you probably still have some practical questions that this video will answer. Like, well, how much of a difference does stacking really make? How does it really work exactly? How many photos of the Nebula should you take? Is 1,024 really necessary? Or could we stop at 64 or 256? What's the difference? Well, to find out, let's start by going back all the way to the beginning, back to my very first picture of the night of Orion. And we'll take a look at the whole process from beginning to end. So this is my setup. I was shooting from a Bortl4 sky, meaning away from the city lights. I drove a couple hours away from Boston. And my setup was the Canon EOS RA camera, a Canon 85 telephoto lens. It was wide open at f1.2. It's on a tripod. I'm not using a Star Tracker. So I'm limiting my sub exposures to two seconds each to avoid star trailing. And I was shooting at ISO 6400. And I just manually focused the lens. I'm just looking at the stars on live view. I didn't even use a mask. I just made the stars as small as possible. And now I'm taking my first two seconds sub exposure of the night. And just like that, it's done. So let's pop it into Photoshop and take a look at it here. So it's okay. It's just a little boring. One thing that I wanna note is that if you take a picture, a two second picture of Orion Nebula, and it doesn't look exactly like this, don't worry too much, because you'll see as we stack, things change dramatically the more photons you collect and how it looks in a single exposure doesn't determine how it'll look at the end. Now, this looks pretty good, I think for a single two second exposure, partly because I was using very nice gear. I was using an expensive lens that opens up very wide to f1.2. And then I was also shooting this from a very dark location where I didn't have too much glow from city lights. But if I zoom in here, you can see there's a lot of noise in the sky background especially, but also in the transition areas between the Nebulae and the sky. And it's really hard to get rid of that. Even if I were to just apply a curve adjustment here and really crush the blacks, you can see we still have that problem of the noise and the transition from the brighter center of the Orion Nebula out here into the sky. And so the best way to get rid of all of this sort of random noise out here in the sky as we'll see is stacking. And so that's why I'm really excited to show you this video, to show you this transformation as we go. But I wanted to just show you what a single two second exposure with this gear looks like. And again, from a nice dark sky. And you can see the way that I framed this was I put the Orion's belt right in the middle. I put the Brightstar Rigel down over here on the right side and the Brightstar Beetlejuice over here on the left side. And that was just the way that I kept trying to re-center it so that I got the most out of the composition. And I think that it looks good with Orion to have those two stars in frame, which is possible with this combination of gear. Again, a full frame camera and an 85 millimeter telephoto lens. So now I'm taking the next three photos for four total and I'm taking them one after another. I'm not re-centering yet because it's just going so fast. But that's where we run into our first problem with stacking is that we have an alignment issue, right? If I take four photos and stack them on top of each other with Photoshop, look what happens. We get star trailing, which is exactly what we want to avoid, but we can expect because we didn't align the photos yet. And so this is the first key thing to keep in mind about stacking is you always have to first align the photos based on the stars. So in this next tab, I've already aligned them based on the star patterns with some free software I'm going to show later in the video. But when the photos are aligned like this, it's easier to see that each individual two second exposure is very noisy. But you can see that that noise is moving around when I blink between them like this. And that's actually exactly what we can expect from how light works. With the particle nature of light, photons arrive truly randomly. So let's say at an hour, this little spot in the sky where my most cursor is right now, should have a true brightness value of 80 photons in one hour. In just two seconds, we may get zero photons or one photon or three because it's pretty random how they're arriving. And so that uncertainty of what the true answer is for the brightness of this spot compared to all the other spots in the picture is noise. That's all noise or technically shot noise is. It's just the uncertainty of measuring light with your camera. And the cool thing is the longer that we expose, the more certain we become because we've captured more of those photons. So let's get back to stacking. First, let me show you simple stacking in Photoshop before we move on to better methods later in the video. I will select these four pictures that are already aligned and turn them into a smart object from the layers menu and then tell Photoshop to stack the layers in the smart object. And there are a few different choices here for how to stack them. I'm gonna choose mean. And if you remember from school, mean is one way to average. So let's first look at what the stacking of four photos did in terms of image quality compared to just a single image. So here on the left is two seconds versus eight seconds total on the right, four images stacked. And I'm just gonna put a small S curve on here just to see the difference a bit better. And wow, look at the horse head nebula here. You can really see that it's just sort of a featureless blob due to the noise in just one photo. But in just four photos stacked, the horse head shape is starting to come out. And you can also see this down here, this dimmer part of the Orion nebula. It's riddled with noise in two seconds, but by eight seconds exposures, it's starting to firm up and look pretty good. And then of course the background sky looks a lot smoother in a stack of four compared to just a single two. So this is the practical difference between one image and four images stacked, two seconds versus eight seconds total. But let's now look at what actually happened at the pixel level. So I'm gonna turn this curve back off. Let's zoom way in here all the way in. I'm gonna sample this pixel in the corner here. And if I look at how bright this pixel is, just in the red channel, we can measure it at 145. That's an eight bit scale. So zero would be completely black, two five five would be completely white. So let's note that down in the stacked image, this pixel has a value of 145 in the red channel. Let's look now at each of the four pictures that went into that stack for that pixel. The first one here has a value in red of 143. The next one has a value of 148. Next one's 144. And the last one is also 144. And remember, we picked just a simple mean for the stacking method. So if I add all these together, we get 579. We then divide that by four. It gives us 144.75. Round that to the host number that gives us 145. So in a simplified manner, that's how stacking works. The brightness of each pixel for each image in the stack is averaged together. And that has the effect of smoothing out the random noise as we're getting closer to the true nature of the sky, the more photons that we collect from it. Now you might have wondered why so far in this video I went from one photo to four photos and now we're about to see a stack of 16 photos. So we're clearly quadrupling each time. And the reason I did it this way is each time we quadruple the total exposure time, we're doubling what's called the signal to noise ratio. And I know a lot of big words there. So let me break it down. Signal means the actual light we're capturing and it always has this random noise associated with it. That's what we call shot noise or photon noise. And this type of noise is equivalent to the square root of the signal. So let's say your first photo you look at had at a certain pixel and you measure it and it measures 25. And you know that the noise for that pixel in just one photo then is five because the square root of 25 is five. So the true signal value for that pixel could be expressed like 25 plus minus five. That's the amount of variation we can expect in terms of the noise. So now let's do a stack of four photos and instead of averaging them together we can just add them together. So now that pixel measures 100. And again to find the noise we take the square root we find that the noise is 10. So while the overall signal increased by four times 25 to 100, the noise only increased by two times five to 10. And that means there's more noise and more signal but what changed is their proportion. So proportionally there's less noise in the photo because the signal went up faster than the noise. So we can say that the signal to noise ratio is five with one photo taken and doubles to 10 with four photos taken. And basically every time we can double that signal to noise ratio you're making the photo twice as clean because you're burying the noise with the signal. Okay, so enough theory. Let's see what actually happened by going from four photos stacked to 16 photos stacked. And I should say these images are stacked and processed by me. I applied the exact same processing to each stack to keep this comparison as fair as possible. So this is four exposures on the left and 16 total exposures on the right and zoomed out like this. They look pretty similar but just like we're gonna see when we do any of these comparisons a lot of the details are gonna be different once we zoom in. So let's go ahead and zoom in on both images. And I think here you can see that while we can see the shape of the horse head nebula there it's a lot more defined here and the nebula, the H-elpha nebula around the horse head shape also looks a lot more filled in and less grainy, noisy like that. Same thing with all of these details. The flame, let's go down here to the running man. And you can see it's still a bit noisy here but we're starting to see some of the bright dust in the 16 sub exposure. And here there's a lot of noise in that part while this is starting to get cleaned up a little bit the transitions are looking nicer. If we look at the sky in general I see that there's a lot of red and blue noise still in this four exposure stack in the 16 exposure stack we've lost a lot of the red and the blue noise which I feel like makes the green noise stand out a little bit more. There's two green pixels for every one red and one blue pixel on a DSLR. So that sort of makes sense. If we look at the witch head nebula it's not very prominent in either photo but I would say in the stack of four exposures it's, if you didn't know it was right there you probably wouldn't see it. In the stack of 16 exposures I think if you knew where it was sort of overall in the image you could pretty, it pops out a little bit there. I can see the shape starting to come into focus there. Okay at this point we need to re-center and with a star tracker this isn't necessary because the tracker follows the movement of the stars but when shooting on a static tripod we need to manually re-center every so often about every few minutes. And if we don't the nebula is gonna move right out of the frame. This isn't hard with a constellation like Orion since it's very easy just to use the bright stars as guideposts for framing and centering. But even with manually re-centering every so often if you look at each different sub-exposure you'll notice the stars are in slightly different places. And so I don't really recommend aligning or trying to align in something like Photoshop manually. It's much better to use an astronomy specific software for stacking like Deep Sky Stacker or serial. What these programs will do is they'll take in your raw photos align them all together based on the star patterns stack them with often slightly more advanced stacking parameters you still average but they do some other useful things like throw out outliers with something called clipping. And so if you think of a standard deviation anything that's an outlier just gets eliminated. So this is things like satellites, airplanes that'll often photo bomb your pictures and clipping just means throwing out those bright pixel values not including them in the stack. When you're done stacking in one of these astro specific programs one thing that's a little bit confusing is why the photo is black. So when I look at individual photos on my camera or in Photoshop they look normal. So why does it come out all black after stacking? And it actually is what we want but it is counterintuitive. So let me explain what's going on. A raw file taken with a digital camera collects all the information in what's called a linear space. So it's just the raw data about how each bright each pixel is but it's all compressed down. And if we look at a raw photo taken with a digital camera without anything done do it it'll look black like this. And you can think of this as you know the way that the digital camera sees it makes sense to the camera or a computer program but it isn't of much use to the human eye. The thing that most people don't know is even if you shoot in just raw when you look at that raw photo and playback on the camera or open it up in Lightroom or Adobe Camera Raw all of those places the camera, the Adobe software are applying processing transformations to your raw files automatically. So before you even touch any sliders or do anything to it the software has already debared your file meaning interpolating the right colors based on the color filter array in your camera and stretch the raw data using a standard curve called a gamma correction which is related just to how the human eye perceives light. So now you might be wondering okay well why don't we just do that with astrophotography apply the standard gamma correction so that it looks right. Well astrophotography processing is very demanding and so having control over the stretch and even more importantly when you apply the stretch from what we call linear space to a non-linear space is part of astrophotography processing. So while it's not typically part of normal daytime photography processing it is important to astrophotography. So we could just let a program apply a gamma correction automatically to your stacked image and then it would not be dark but that's a permanent burned in chain. So instead what most astrophotographic software does is it leaves that up to you when did apply the stretch if you want it to be an automatic stretch or a manual one. Anyways let's now look at 64 sub exposures stacked and processed and compare that to just 16. Okay going from 16 to 64 exposure stacks you could see this is a pretty dramatic difference actually zoomed out because the Barnard's loop area this red semicircle down here and the dust in between the Barnard's loop and these nebulae is really starting to pop in the 64 stack while in this one it's still pretty dim. If we zoom in a bit we can see though that that dust which I'm now able in this 64 stack to bring out a little bit looks okay zoomed out but once you zoom in you can see that dust is pretty noisy still. So I raised the exposure on this stack of 64 to bring it out but that has the consequence of when you zoom in the stack of 64 almost looks noisier than the stack of 16 because I felt confident that I could start bringing out the dust. So there's always trade-offs in processing that you make. I'm one who's sort of a, I think sometimes a maximalist processor I like to bring out as much detail that I can that I can even if it's a bit noisy. Same thing here, this is more restrained it actually probably looks better in terms of noise than this but you can see at 64 exposures was the first time I felt sort of confident about making the picture brighter. Now just as an experiment let's try making this picture as bright as this one so we can really compare them a little bit better. So I would say that's about similar brightness and you can see the noise actually is much worse over here but it's just that I pushed this one a bit more than this one in processing not really on purpose but just because I think I got excited about the details I was seeing. And the main thing that I noticed is when I push the brightness on this exposure it looks a lot, there's a lot more green color noise than there is in this one. Let's look at Orion now. And you can also see that there's certain details that are starting to come out like this finger feature in the dust right here in the stack of 64 that here in the stack of 16 are still completely lost in the noise. Let's look at our old favorite the witch head and you can see there's enough contrast here now in this one, let me zoom out a little bit so you can clearly make out the witch head there while in this one when I made it as bright as this one it really is getting sort of lost in the noise again. Okay, now we're really getting into the law of large numbers where we should start seeing some of that really dim detail coming out from the noise like dark nebulae and dust and things that I really like. So let's compare 256 now to just 64. All right, we have a stack of 64 exposures on the left and a stack of 256 on the right. And I think in this wide view, one thing that's interesting is this up here the angel fish nebulae next to beetle juice. The shape of it is starting to be clearer here. Here you could almost see it as a mistake because it's so dim. Barnard's loop is also firming up and we're also seeing more of the brown dust. We also see that this interesting outcropping of HA from Barnard's loop is coming out a lot stronger than this one on the right, the stack of 256 compared to the one on the left. Let's zoom in now to look at some details. So the main thing that I see here in the one on the right is just that everything is a little bit brighter and a little bit cleaner, right? We're getting into more subtle differences now but if you look closely in some of these areas you're seeing that everything, the gradations from bright to dim look a lot cleaner. They don't look as modeled as broken up and grainy but I would say that we're definitely starting to hit some diminishing returns, right? It's not, this isn't as a big differences from 16 to 64, I don't think. But where you do see bigger differences is on the really dim stuff, right? And that's sort of to be expected because here this dim object is sort of down in the noise floor and you can see it's pretty riddled with noise. Well here it still is pretty noisy but it's starting to really come into its own. We can really see how the nose is connected to the rest of the witch head there. Well in this one it almost looks disconnected from all the noise in the image. Well we've reached the end here, we've quadrupled again to get to 1,024. And usually around here 1,000 sub exposures is my practical limit for untracked astrophotography. Of course what really matters is just the total time. So this is 34 minutes total and astrophotographers will usually call this total integration. And that's basically synonymous with total exposure but by saying total integration I think the point is that it immediately makes it clear you're talking about stacking several sub exposures together. It's not just a single long exposure. So this image has a total integration of 34 minutes and a little bit. And the last one we looked at had a total integration of eight and a half minutes. So that's a pretty big jump from eight and a half to 34. But let's see if that extra time investment was worth it. And now we have 256 on the left and 1,024 photos stacked on the right and zoomed out like this I don't see a huge difference. It looks like the color balance is a little bit different. I don't know, for some reason it looks a little bluish on this one. And then I'm also actually seeing some more dust details up here in the upper right. Let's move this around a little bit and we'll take a look at that. Oh yeah, that's definitely some interesting bright dust up there. And I also see some dark nebula features near beetle juice here. But I think the bigger differences are gonna be when we zoom in on some of these dim things. So let's go ahead and do that. So yeah, definitely we're seeing some interesting dust features that I don't think I've ever captured before here around beetle juice. And the boogeyman nebula, we're seeing some nice bright dust and actually seeing the gradations from the darkest part of the dark nebula to a more brown gray dust around it, which I think is pretty cool in this 1024 stack. Okay, and then this is getting quite accomplished, I think. It's looking really good, the M78 reflection nebula here. Here we still have a bit of color model. We do get sort of the brown color of the dust around it, but it looks a lot more refined here and we're seeing more nicely gradients from a grayish brown to a darker brown to the black parts of the dark nebula. And again, this is just a difference between pretty good, 75% there to pretty much finished. I think that looks really nice, especially considering this is an untracked stack. Same thing with Orion nebula. Down here, if I was gonna nitpick, this is still a little bit noisy, especially if I zoom in a little bit more, you can see there's still a fair amount of sort of red color noise in there, some broken up parts where you can see a little bit of model left. But if we look at some things like the finger here, this finger of dust, there I can see it has a lot of green model left in at 256 and by 1024, that looks like a pretty solid, dusty brown color. So it looks a lot more finished. And then now let's look at one of my favorites, the witch head. Here's some interesting dust near the witch head. Okay, and there's the witch head itself. And you can see again, this looks nearly finished. This looks like maybe 90% of the way there. Well, this looks only half baked, right? It looks a lot more broken up when we see it compared to this one now. So I'm gonna zoom all the way back out. And yeah, I just want you to look at this because I mean, when you're looking at them zoomed out, you really don't see all these differences. But even in the witch head here, you know, but when you really zoom in and take a look at these differences, you can see that they are striking in some places, especially in the little fainter details that are sort of hard to see when we're zoomed out like this. But when we zoom in, we really can start to examine them. And that's something I love about S2 Photography is in a wide field image like this getting both the grand perspective, but then also being able to zoom in and look at smaller details and not be too put off by how noisy it looks. And now let's look at just some crops of each jump from one sun exposure up to 1024. I always find these kinds of comparisons can really drive home what's happening as we go from one to the next. Okay, so let's take a tour through the image here, just looking at some crops. And we're starting with the Boogeyman Nebula. This is a favorite of mine. It's not really the best object in a wide shot like this because it doesn't have a huge amount of impact wide, but it is starting to get interesting, I think, in the stack of 1024. It's definitely defined in this stack of 256. By the time we get to the stack of 64, it's barely visible. You could mistake that for just sort of weird noise structures. And then it's invisible in these three on the left. Okay, next up here, this is Barnard's Loop or part of it. And you can see Barnard's Loop by the stack of four. You're starting to see something there by the stack of 16. You can definitely see it. Stack of 64, it's looking okay, but there's still enough noise that it looks a little, the transitions aren't great. By the stack of 256, I think that looks pretty good. There's just a bit of red noise that's distracting to me. And then the stack of 1024, it's taking on a little bit of a lighter magenta color that might have been processing kind of thing. But I think that you really can only get that nicer color of it with the nicer gradations when you really have enough data to play around with. I think you're also starting to see some dust playing and intermingling with Barnard's Loop. Okay, and then we have the M78 Reflection Nebula. And I think, while this one is visible in all six, it's bright enough in that little core area to see a little bit of reflection, even in a single shot, it doesn't really look accomplished until we get all the way over here to 1024. I'd say in 256, it's passable. You get that nice bridge of dark nebula right there that's really coming across, but it's even coming across in this one, the stack of 64. But to really see all of the interesting structures that make up the dark structures around the reflection nebula, you don't really get that until over here. Okay, there's the flame nebula. The flame nebula is bright enough that I think even in this stack of 16, it looks pretty good. You can see that, I like to think of this dark part in the flame as being sort of like a tree, a tree trunk with some branches coming out. But I also see the idea of it being a flame. And I'd say that while it looks fine here in the 16, it really looks best by the time we get to 256 and then there's really not too much improvement over here. It's just a little bit cleaner in the transition areas. There's also this little reflection nebula right underneath, which doesn't, it's so small in this image that it's hard to really make much of it. But I think that that starts coming across at 256. The horse head, I think already looks pretty good at least at this scale in the 64 image stack, but definitely looks more accomplished in these two. We have another reflection nebula up here. I think this is a sharpness object, but I don't remember. It looks pretty good in the final two. Okay, here's the running man nebula. I think that the big issue with the running man nebula is it almost looks better in the stack of 16. And then we have to skip over the stack of 64 because in the stack of 64, I'm trying to start bringing out this brown dust all around it, but it's really not ready to do that yet until we get to the stack of 256. So that's just sort of probably a processing mistake on my end. I wish to kept this one darker like this one, but I just got overeager. But anyways, I think that even in the single one, we're already starting to see some of that detail in the running man. It definitely starts looking quite accomplished though here at the stack of 256. And then by the stack of 1024, we're just sort of refining the transition to the dust. The Ryan nebula. So the Ryan nebula is a really interesting one because I've blown the trapezium region even in a two second exposure at ISO 6400 F1.2. So the trapezium region, this little core area is so bright that it's so easy to blow out that it's really hard to recover. So I usually just don't really even try. I just know that I'm gonna blow it out, but you can see that by the time we get to 1024, that blown out core has extended quite a bit than from where it was here. And I didn't do any really kind of HDR effect on this to try to recover the core details. I could if I wanted to, but I didn't really try. So you can see that the core gets, you know, successively more and more blown out as we get over here. But I think that that looks natural in a wide field like this. I don't try to really, I like seeing the brightness relations of the different objects, I guess is my point. Around the Ryan nebula, the brown dust doesn't, I think look really accomplished until we get to the final one. If we move up here, this is some interesting colors that really are just look like noise here at the stack of 64. They look like something, but really still pretty noisy in the stack of 256. And I think here you can tell they are something, but they're still even at 1024 pictures really not defined yet like they could be. So that's an area where we could still, I think use more exposure time, more integration. This is interesting. I haven't really noted this that the star halo on bright stars also gets bigger and bigger, the more total integration you put in. And I think in a wide field shot like this, that looks good to have the bright stars have a nice big colorful halo. And so this is the bright star Rigel. And I really like how it has a nice bright blue halo so you can see that's a blue star. Okay, and then here's the chin of the witch head. Let's go to the nose. There's the nose and the eye. And really I think looking at it now, I think really it really only comes across well in the stack of 1024, so 34 minutes. In this eight and a half minute shot at 256, it's definitely there, we can see the structures, but it's still very noisy. And in this one, it's basically a battle between the noise and any signal we have. And then I'd say it's nearly invisible at 16 exposures and below. Hopefully at this point, I've answered how does stacking work? But another question I said I'd address is, when do you stop? How do you know how many photos is enough? How many to take? And I think that really has to do with your goal. After 64 photos, the Ryan Nebula looked pretty good already, but we've discovered as we've gone down that path that there's so much more that can be revealed in the scene with more integration. So if your goal is to bring out the dust clouds or M78 or the Witchhead Nebula, those are all much dimmer and that's gonna require putting in more time. And of course, adding a Star Tracker can really help in this respect because instead of having to take over 1,002 second shots, I could have just taken 34 one minute shots and saved a lot of hard drive space by just having 34 versus 1,000. It also saves the cameras shutter a bit. But that's a whole other topic, Stacker versus No Stacker. I've already done a video on that actually and I'll link it right up here. This video is a lot of work and it wouldn't be possible without financial support. So at this point, I wanna thank all of my wonderful patrons over on patreon.com slash Nebula Photos. And if you're interested in seeing your name in the credits of any of my long videos, any of the videos over 20 minutes, you can head over to this link and support me there. The support of the channel starts at just $1 per month and then I also have a $3 and a $7 tier. And there are a number of benefits included for signing up on my Patreon. These include a vibrant Discord community where people post their photos, get photo critiques, help each other with gear and processing and everything related to astrophotography. And we also have a monthly challenge object that's a friendly competition with some prizes. We have a group project, if you wanna work on something collaboratively with others. And finally, there are some exclusive videos including a monthly live Q and A that's always recorded and shared right on Patreon. So if you sign up, you can actually already watch 10 of those monthly chats. And then lastly, like I said, you get your name in the credits of any video longer than 20 minutes. So the link again is patreon.com slash Nebula Photos. I hope to see you there. Till next time, this has been Nico Carver, Clear Skies.