DESIGN CRIT 010
COMPOSITING
Compositing techniques are crucial for vfx and animation, but you can use them to glue your designs together into beautiful stuff too!
TRANSCRIPT
Today we’ve got a submission from Dan Skvaril for the 5th challenge in the Visual Design Lab, where you choose from a list of themes and go at that theme full force with everything you’ve practiced in the preceding chapters. No holds barred, imagery, shapes, colors, 2d, 3d, candy sprinkles on a velvet cupcake, oil paints on your mom’s favorite duvet… the point is to convey your theme, whatever it means to you, however you can best get that meaning across. There are a bunch of themes to choose from, and the theme Dan chose was “battle”.
Now most people are gonna be kind of literal about it. First thing that comes to mind: soldiers, fighting, that’s all fair game. Dan interpreted it as more of an internal battle. A mental battle. And he chose to represent that through running. A struggle of endurance. Of stamina. Of mental fortitude in the face of pain and exhaustion and all of that. But his first attempt was kind of all over the place. He’s actually a UI/UX designer and is really new to this sort of thing, so he was trying a lot of things that weren’t really working, as everyone does. He had multiple ideas and wasn’t sure how to make them work together. So I kept helping him focus in on what he actually wanted, stripping out the non-essential, and he ended up with this, which is a much clearer place to start. Just being simpler isn’t necessarily better, but this is more coherent and to the point. He flipped the values to light on dark, but he did that to really focus down on this sort of double exposure image and to highlight the fire. And highlighting specific things makes the visual hierarchy much more apparent, so as viewers, we know what we’re lookin at.
But most of all, I was getting him to start using the same strategy we explored in this episode. Using minimalist principles just to find some clarity on our own ideas so we can be more intentional about what we’re doing. And stripping it down to what’s essential also means we have fewer elements we have to juggle and think about, which makes things easier to handle at the start. And that allowed Dan to focus down on what he was trying to do. What’s important. And for one thing, he realized this image he was using here is pretty passive, and he needed to swap in this image where the guy has some grit. He’s maybe struggling a bit, it’s more of a battle. And that’s more to the point, right?
But stripping it down and reducing doesn’t mean the result has to look like minimalist art by any means. It’s only a strategy for clarifying things, and we can still build some visual richness back into this. Because Dan admits to being pretty new to the kinds of compositing techniques that could push this a lot farther, so I thought we might try some of that out.
Now most people are gonna be kind of literal about it. First thing that comes to mind: soldiers, fighting, that’s all fair game. Dan interpreted it as more of an internal battle. A mental battle. And he chose to represent that through running. A struggle of endurance. Of stamina. Of mental fortitude in the face of pain and exhaustion and all of that. But his first attempt was kind of all over the place. He’s actually a UI/UX designer and is really new to this sort of thing, so he was trying a lot of things that weren’t really working, as everyone does. He had multiple ideas and wasn’t sure how to make them work together. So I kept helping him focus in on what he actually wanted, stripping out the non-essential, and he ended up with this, which is a much clearer place to start. Just being simpler isn’t necessarily better, but this is more coherent and to the point. He flipped the values to light on dark, but he did that to really focus down on this sort of double exposure image and to highlight the fire. And highlighting specific things makes the visual hierarchy much more apparent, so as viewers, we know what we’re lookin at.
But most of all, I was getting him to start using the same strategy we explored in this episode. Using minimalist principles just to find some clarity on our own ideas so we can be more intentional about what we’re doing. And stripping it down to what’s essential also means we have fewer elements we have to juggle and think about, which makes things easier to handle at the start. And that allowed Dan to focus down on what he was trying to do. What’s important. And for one thing, he realized this image he was using here is pretty passive, and he needed to swap in this image where the guy has some grit. He’s maybe struggling a bit, it’s more of a battle. And that’s more to the point, right?
But stripping it down and reducing doesn’t mean the result has to look like minimalist art by any means. It’s only a strategy for clarifying things, and we can still build some visual richness back into this. Because Dan admits to being pretty new to the kinds of compositing techniques that could push this a lot farther, so I thought we might try some of that out.
The first step to doing some solid compositing is to look at what we’re working with. This is one of the original images he was using and it means basically the right things. This guy’s got that look of struggle and determination. That’s good. But visually, we also really have to note that the image itself, the photography, isn’t very dynamic. It’s just shot sorta straight on, and the lighting on him is pretty flat. And that makes sense. He’s in the shade. No directional lighting source, it’s all ambient. For context, look at something like this. You can more easily tell where the light is coming from. Highlights up top, shadows underneath. Or even more extreme, just these sort of rim lights from the sides. Being able to spot this stuff is important because it’s not always immediately apparent, and when we’re looking for other elements, they’ll need fairly similar lighting to each other if they’re gonna feel like they belong together.
If you’re compositing 3d elements, obviously you can control the lighting in the renders. Same if you’re illustrating elements, and you’ll need to, if you want to combine or composite one with the other. But if you’re using photos, unless you’re shooting them yourself, the tough thing is finding shots that fit all your criteria. Determined, struggling runner, maybe in a dynamic position and in lighting that isn’t super flat. You’re gonna spend a while looking for that. And I did. I got lots of stuff to try out. And in that search, I found some things that might not work with what Dan used, but might work well with each other. There are undoubtedly much nicer images out there, better photography, but maybe they don’t tell the story or convey the idea you need to convey. Like this is professionally lit, but this is a fashion ad. This guy’s posing, not struggling. Sometimes you get lucky and find beautifully shot stuff that actually means exactly what you need it to, but often you kinda compromise and see what you can make work. Better photography makes it way easier, but really you find the best you can for your subject matter.
Like, a long time ago, I created this spot for the UFC as a storyboard sequence, and designed this ridiculous training facility, and needed to put a fighter into the environment. This image seemed pretty good, that fist is flying… with the UFC logo on the glove there, which is great. But the photo obviously doesn’t fit the background very well. The lighting is way too bright, but at the very least it’s coming from the same basic direction. It’s from above. So, get it positioned, focused on that fist and the UFC logo, which is now backward, and start using simple photoshop tools like levels and hue and saturation to darken the shadows, sharpen the highlights, and suck some of the color out, just trying to match the general lighting conditions in the environment. I even painted up some of the highlights to get them to pop. And those simple moves work pretty well in this case. Sometimes it takes more, sometimes less. Dan actually comped this guy’s determined face on to a totally different body. That’s a good choice, and sometimes that’s just what you gotta do.
But I want to use these runners and take Dan’s basic idea and do a different take on it. I want to use these two in particular, both because the body language seems to convey that they just gave it their all, but also because visually those poses are much more dynamic. And then I also found this woman who looks to be in some kind of mental anguish and she might help convey that idea of internal struggle.
If you’re compositing 3d elements, obviously you can control the lighting in the renders. Same if you’re illustrating elements, and you’ll need to, if you want to combine or composite one with the other. But if you’re using photos, unless you’re shooting them yourself, the tough thing is finding shots that fit all your criteria. Determined, struggling runner, maybe in a dynamic position and in lighting that isn’t super flat. You’re gonna spend a while looking for that. And I did. I got lots of stuff to try out. And in that search, I found some things that might not work with what Dan used, but might work well with each other. There are undoubtedly much nicer images out there, better photography, but maybe they don’t tell the story or convey the idea you need to convey. Like this is professionally lit, but this is a fashion ad. This guy’s posing, not struggling. Sometimes you get lucky and find beautifully shot stuff that actually means exactly what you need it to, but often you kinda compromise and see what you can make work. Better photography makes it way easier, but really you find the best you can for your subject matter.
Like, a long time ago, I created this spot for the UFC as a storyboard sequence, and designed this ridiculous training facility, and needed to put a fighter into the environment. This image seemed pretty good, that fist is flying… with the UFC logo on the glove there, which is great. But the photo obviously doesn’t fit the background very well. The lighting is way too bright, but at the very least it’s coming from the same basic direction. It’s from above. So, get it positioned, focused on that fist and the UFC logo, which is now backward, and start using simple photoshop tools like levels and hue and saturation to darken the shadows, sharpen the highlights, and suck some of the color out, just trying to match the general lighting conditions in the environment. I even painted up some of the highlights to get them to pop. And those simple moves work pretty well in this case. Sometimes it takes more, sometimes less. Dan actually comped this guy’s determined face on to a totally different body. That’s a good choice, and sometimes that’s just what you gotta do.
But I want to use these runners and take Dan’s basic idea and do a different take on it. I want to use these two in particular, both because the body language seems to convey that they just gave it their all, but also because visually those poses are much more dynamic. And then I also found this woman who looks to be in some kind of mental anguish and she might help convey that idea of internal struggle.
But we’re not necessarily gonna try to reinterpret the idea of a battle or an internal struggle, we’re mostly just gonna try to execute Dan’s idea of the runner on fire more dynamically, which takes us to step 2, where we’ll see how these main elements can fit together. And a great way to do that is to start by blocking it out, like a rough sketch built with real photos. Starting by looking at positioning and size relationships, like where they are in relation to each other, and how that feels, and what those relationships kind of imply as a result. Here they’re kind of cascading from one to the next, so it almost feels like they’re in sequence. It kind of implies that it’s the same woman at different points in time. Maybe that idea is something we can work with.
It’s at least something to think about while going about with production stuff like cutting the images out more cleanly, where it pays to not be sloppy, and if background removal or auto-trace functions aren’t getting it quite perfect, then it really makes a difference to take an extra 10 minutes and do it manually. Same with things like removing this woman’s ring, which I’m doing because I’d like this image to be fairly large in the frame, and at that size, her ring would draw just enough attention to be important, and then it seems like what she’s agonizing over is her marriage or something. This is that issue of visual signifiers and it shifts the meaning toward something I clearly don’t want it to convey. So out it goes. Now we don’t know what she’s crying about, could be anything. And in the context of these runners in their struggle, we’re more willing to interpret it as something related to that instead. I want to remove this text for the same reason. It’s just calling too much attention to itself, and potentially shifting what’s important in the frame.
Now the reason I darkened each of them is twofold: first, I need to adjust their colors and values to get them to look like they kinda belong together, and that’ll take some work. And second, we’re gonna add the fire in, which is just these images from the original on black backgrounds. And I’ll set them to “screen” so the black disappears, but the bright parts that remain won’t be very visible against bright objects, so I really have to set them against darker things. Having a dark background and darkening the runners helps for that particular reason. And I’m gonna knock out parts of that fire so it looks like it’s wrapping around her limbs and coming off of her back.
And that seems to work pretty well, but now I have to figure out how that’s going to affect this second runner. Now there’s a new light source from this fire, and it’s going to need to look like that fire is illuminating parts of her. So there’s gonna be a lot of little tweaks and attempts at doing that.
It’s at least something to think about while going about with production stuff like cutting the images out more cleanly, where it pays to not be sloppy, and if background removal or auto-trace functions aren’t getting it quite perfect, then it really makes a difference to take an extra 10 minutes and do it manually. Same with things like removing this woman’s ring, which I’m doing because I’d like this image to be fairly large in the frame, and at that size, her ring would draw just enough attention to be important, and then it seems like what she’s agonizing over is her marriage or something. This is that issue of visual signifiers and it shifts the meaning toward something I clearly don’t want it to convey. So out it goes. Now we don’t know what she’s crying about, could be anything. And in the context of these runners in their struggle, we’re more willing to interpret it as something related to that instead. I want to remove this text for the same reason. It’s just calling too much attention to itself, and potentially shifting what’s important in the frame.
Now the reason I darkened each of them is twofold: first, I need to adjust their colors and values to get them to look like they kinda belong together, and that’ll take some work. And second, we’re gonna add the fire in, which is just these images from the original on black backgrounds. And I’ll set them to “screen” so the black disappears, but the bright parts that remain won’t be very visible against bright objects, so I really have to set them against darker things. Having a dark background and darkening the runners helps for that particular reason. And I’m gonna knock out parts of that fire so it looks like it’s wrapping around her limbs and coming off of her back.
And that seems to work pretty well, but now I have to figure out how that’s going to affect this second runner. Now there’s a new light source from this fire, and it’s going to need to look like that fire is illuminating parts of her. So there’s gonna be a lot of little tweaks and attempts at doing that.
In the meantime, I want some of the fire or sparks in the background as well, to help silhouette the 2nd runner and the background lady. This is really moving into step 3 where a lot of this work is just about helping the viewer’s eyes easily pick out some of the edges that are most important to discerning the silhouette, so everybody knows what they’re looking at. Bright figures against a bright background makes that hard. It’s hard to see which limbs are hers, or where her arm ends, and it makes the image hard to look at, as a result. There’s not enough value contrast for the eye to easily figure it out. Same here with dark against dark, so i’m using the brighter fire strategically to help with that. Something else I’m doing at the same time is making it look like it’s wrapping around them a bit because that helps indicate to the eye that there’s some depth between each of these things.
This moon is just something I found while searching for details to put in the background, landscapes and stuff, and it doesn’t really have anything to do with Dan’s original image, but it’s a nice point of focus that’s appropriate for this dark scene, and again it does a great job of helping to establish this dark silhouette. Really helps you see this arm and her head.
And here are some of those backgrounds I found. I’m choosing these specifically because I want something to subtly point inward to the runners, to keep the focus on them, and these are also good because they already have lighting indications on them, like we talked about. Even when crushed down and darkened, it helps a ton that they seem like they’re lit by something in this general direction, because that makes sense, now that we’ve got fire and this bright moon casting light from here. You might not think that that’s important for such a subtle background detail, but it really is. No matter what your lighting situation, you want the majority of your elements to agree on it. This lady’s light source is coming from the wrong place, but there’s not a ton I can do about that in this case, other than flip her around, and that’s just not gonna work well compositionally. It would be better if she were lit from the right, but at the same time, she’s maybe a more metaphorical figure, and isn’t meant to be seen as existing in the same space necessarily, so it shouldn’t be a total dealbreaker.
What does need to happen is that her silhouette needs a little more defining, especially in the hand, and I tried doing that with fire, but it draws too much attention, so I’m bringing in this sky to be just a bit brighter than her to help the viewer’s eye see the edge of her hand. And shifting the hue of that a bit means we get some complementary colors, some blues and magentas to go with all of the orange and yellow, which is really nice. But it’s primarily there because I need something a little brighter up against her darker shapes just so we can see their edges, and see her. For something like this, a lot of the work of integrating these things is in making sure that the eye can really see them, and is able to distinguish one element from another, all while still trying to make them feel like they belong together. We want these two completely different photos to feel like they come from the same place. So at a certain point it comes down to tiny details: where the fire seems to wrap around from one figure to the next, and where this arm catches a little bit of light from that fire. Where little highlights might be needed. Originally I couldn’t remove the background from her hair very well, but now she’s got something bright behind her, so I can separate the hair out and use “multiply” to set it against that bright thing. And the reason I’m doing that is so that her pose is a little bit more dynamic, but also so you can more easily see that it’s a girl, and hopefully make the connection that these are all women, maybe the same woman at different moments.
This moon is just something I found while searching for details to put in the background, landscapes and stuff, and it doesn’t really have anything to do with Dan’s original image, but it’s a nice point of focus that’s appropriate for this dark scene, and again it does a great job of helping to establish this dark silhouette. Really helps you see this arm and her head.
And here are some of those backgrounds I found. I’m choosing these specifically because I want something to subtly point inward to the runners, to keep the focus on them, and these are also good because they already have lighting indications on them, like we talked about. Even when crushed down and darkened, it helps a ton that they seem like they’re lit by something in this general direction, because that makes sense, now that we’ve got fire and this bright moon casting light from here. You might not think that that’s important for such a subtle background detail, but it really is. No matter what your lighting situation, you want the majority of your elements to agree on it. This lady’s light source is coming from the wrong place, but there’s not a ton I can do about that in this case, other than flip her around, and that’s just not gonna work well compositionally. It would be better if she were lit from the right, but at the same time, she’s maybe a more metaphorical figure, and isn’t meant to be seen as existing in the same space necessarily, so it shouldn’t be a total dealbreaker.
What does need to happen is that her silhouette needs a little more defining, especially in the hand, and I tried doing that with fire, but it draws too much attention, so I’m bringing in this sky to be just a bit brighter than her to help the viewer’s eye see the edge of her hand. And shifting the hue of that a bit means we get some complementary colors, some blues and magentas to go with all of the orange and yellow, which is really nice. But it’s primarily there because I need something a little brighter up against her darker shapes just so we can see their edges, and see her. For something like this, a lot of the work of integrating these things is in making sure that the eye can really see them, and is able to distinguish one element from another, all while still trying to make them feel like they belong together. We want these two completely different photos to feel like they come from the same place. So at a certain point it comes down to tiny details: where the fire seems to wrap around from one figure to the next, and where this arm catches a little bit of light from that fire. Where little highlights might be needed. Originally I couldn’t remove the background from her hair very well, but now she’s got something bright behind her, so I can separate the hair out and use “multiply” to set it against that bright thing. And the reason I’m doing that is so that her pose is a little bit more dynamic, but also so you can more easily see that it’s a girl, and hopefully make the connection that these are all women, maybe the same woman at different moments.
And that kind of stuff is part of step 4, which is the polishing process. The composition is set, and you’re going around addressing little things. The black fringe on this woman needed some cleanup. The moon could use a little bit more detail, but not so much that it becomes super important. These are often really subtle things. I’m playing with whether to bring a little bit of smoke back into the fire, and deciding it’s a little too bright and too busy. Her shirt is maybe too bright for something that’s in shadow, but then some of it is gonna be iluminated by fire. You address little fixes and over time, you’re slowly kind of cementing things in place, making them work together, bit by bit. And then, at some point, you decide it’s done! And I think it is!
Now, we can debate whether Dan’s basic idea of a runner engulfed in flames is a good strategy for conveying ideas about an internal battle, or whether my execution of his idea is more or less to the point. Undoubtedly, there are ways to execute it that would be more to the point, but I really just wanted to take his basic elements and see if we could do something more dynamic with them. And I think we did. I like this one. I like it a lot.
This probably took 6 hours, with some parts going faster and some going much slower. You have to find images and test them, and if you can’t extract the parts of the images you want from their backgrounds perfectly, it pays to meticulously cut them out, and all of that takes work. But it’s really the finessing where you spend most of your time. Once the basic position and scale and orientation of things is pretty much set, then you go to work on the smaller details. In this case that’s things like where the fire weaves between the figures, and how to use that to keep attention on what’s happening right here. What do I need to highlight and what do I need to draw attention away from so that the focus remains on this idea? And then you keep going finer. Once those details are kind of in place, you get down to the really fine details, which again in this case is like where is the light from that fire illuminating this runner? What parts of her body need to be better delineated so her silhouette is visible? How much is necessary to make this hand legible without making it super dominant in the frame?
And most of that will take some experimenting to see what works best. Even in a technical sense. In some cases, I used levels to darken shadows and boost highlights, and in others I just painted darker colors in with a soft brush, and in some others I did both because neither seemed to work well by itself. You’re gonna try stuff that won’t work, and it’s ok to try again. However easy all of these 60 second design recipes and 5 minute tutorials make it look, to do something that’s really yours almost invariably takes some real work, and you may go down some paths that don’t pan out. And you may still spend hours finessing the details once you’ve found a direction that feels good. In the design lab, we look at examples that took 20 minutes from start to finish, and examples that took 50 hours. And we build the groundwork to let you spend that time well and end up with work that you can really be proud of and enjoy. It takes time and practice, but you can get there, so check that out, and check out the full res image of this below, and thanks again to Dan for letting me take a whack at this one. He’s on the right path and has some other fun stuff up that we might have a go at next time.
Until then, take a crack at compositing photos together like this, it’s fun! Find stuff you think will work together, and see if it does. Arrange it, play with it, but keep it really simple at first. Then clean it up and start working in the secondary details to see how they affect how things read and whether the elements might need to be adjusted, because they almost always will. Draw the viewer’s eye to important areas of interest and establish which silhouettes need to be really clear and visible. And then get down to the super fine details that glue everything together. Lighting, colors, further delineating certain edges, making things more or less visible to the eye. Try some stuff out. You’ll figure out what works and what doesn’t. It’s really exciting when it does, and you may end up with something you really like! Alright, get to it. Have fun! See you soon.
This probably took 6 hours, with some parts going faster and some going much slower. You have to find images and test them, and if you can’t extract the parts of the images you want from their backgrounds perfectly, it pays to meticulously cut them out, and all of that takes work. But it’s really the finessing where you spend most of your time. Once the basic position and scale and orientation of things is pretty much set, then you go to work on the smaller details. In this case that’s things like where the fire weaves between the figures, and how to use that to keep attention on what’s happening right here. What do I need to highlight and what do I need to draw attention away from so that the focus remains on this idea? And then you keep going finer. Once those details are kind of in place, you get down to the really fine details, which again in this case is like where is the light from that fire illuminating this runner? What parts of her body need to be better delineated so her silhouette is visible? How much is necessary to make this hand legible without making it super dominant in the frame?
And most of that will take some experimenting to see what works best. Even in a technical sense. In some cases, I used levels to darken shadows and boost highlights, and in others I just painted darker colors in with a soft brush, and in some others I did both because neither seemed to work well by itself. You’re gonna try stuff that won’t work, and it’s ok to try again. However easy all of these 60 second design recipes and 5 minute tutorials make it look, to do something that’s really yours almost invariably takes some real work, and you may go down some paths that don’t pan out. And you may still spend hours finessing the details once you’ve found a direction that feels good. In the design lab, we look at examples that took 20 minutes from start to finish, and examples that took 50 hours. And we build the groundwork to let you spend that time well and end up with work that you can really be proud of and enjoy. It takes time and practice, but you can get there, so check that out, and check out the full res image of this below, and thanks again to Dan for letting me take a whack at this one. He’s on the right path and has some other fun stuff up that we might have a go at next time.
Until then, take a crack at compositing photos together like this, it’s fun! Find stuff you think will work together, and see if it does. Arrange it, play with it, but keep it really simple at first. Then clean it up and start working in the secondary details to see how they affect how things read and whether the elements might need to be adjusted, because they almost always will. Draw the viewer’s eye to important areas of interest and establish which silhouettes need to be really clear and visible. And then get down to the super fine details that glue everything together. Lighting, colors, further delineating certain edges, making things more or less visible to the eye. Try some stuff out. You’ll figure out what works and what doesn’t. It’s really exciting when it does, and you may end up with something you really like! Alright, get to it. Have fun! See you soon.
MENTIONED IN THIS VIDEO