 Okay. Can you hear me all right? Okay. Thanks for coming. So, we just converted all of these slides two minutes ago to PowerPoint, so I have no idea if this is going to work, but let's give it a try. Okay. So, I'm Dave Ferner. I'm Creative Director at Studio RGL in London, and we're making this little documentary up here. It's actually my first Blender conference, and really excited to be here, and excited to tell you about what we're making, and it's very much a we that are making it. Almost all of these talks, it's just one person standing at the front, almost always a dude, but actually Rosie, our producer, runs the company with me, and Una, who's from a live action background and is a filmmaker and editor, is helping out with this project, and the three of us have been working on it for almost a year now in pre-production, and we're being joined by all sorts of talented people and helped out by many others, and a few more starting once we're into full production. So, yeah, this is actually our first nonfiction work, so you can see the problem, right? We actually had to change the logo, so people trusted us more. If anyone can think of what this stands for, let me know. Cool. So, we're going to be talking about a lot of things, look development, art direction, technical direction, but what it all boils down to really is how to make something look cool. So, yeah, this is what you start out with, or rather actually this was the hard bit, the script, but we're not going to be talking about that today. So, you've got all your characters and all your scenes, so how do we figure out what we want it to look like? So, I know all of you are here today for technical tips, and I'm not great with technical tips, but I have one, and this is the only one, so you can go off and get lunch after this, and it's how to make something cool in two easy steps. So, hopefully this is going to work. Figure out what you want to make, and then make it. Okay, so maybe not easy steps, and you could probably break them down into sub steps, but the important thing is the order, right? Figure out what you want to do and then make it. If you start making it before you know what you want to do, it's great for prototyping, great for experimenting, but often you'll find the tool starts to guide you rather than you guiding the tool. So, we'll come back to this. Okay, so what do we want to make this thing look like? So, a lot of the talks have been in this green circle here, and a lot of the work that goes into 3D software, you know, radiosity, subsurface scattering is all focused on making photorealistic stuff, and then there's this little area some of you may be aware of down here, NPR, which if you look at the forums is pretty much just anime, but actually that's not what NPR means, it's just non-photoreal, so really everything outside this circle is NPR, including anime, and there's been some wonderful talks with different areas in this, deduce particularly, I think he's about here, so, and I've been thinking about this a lot actually, so this is from understanding comics by Scott McLeod, and what Scott's tried to do is figure out a space and kind of categorize stuff, he's got three poles from reality to the picture plane to meaning, and you know anime is here or whatever, but even he will admit that this is a massive oversimplification, there's no such thing, it's an infinite space, right? Everything outside of photoreal is up to us. It doesn't really help us in narrowing it down, because we've got so many options, so what do we want? We have our story and we need to come up with a look that supports the mood of the story, and as you may have noticed we're a really tiny team, so we need to come up with something that works with the team size, so obviously when you're doing animation it's nice to just doodle all the time, so you're doing story doodles all the way through, but you also want to know like what's it going to look like when it's finished, right? And I was sort of, so actually we started this thinking about this in lockdown and then I just moved out into this wonderful space in London, which is, it's called Miller's Junction and it's just a shared studio with animators, but also illustrators, graphic designers, photographers, poster makers, everything in between, really fun place to be, and it's actually, it's part of something called Print Club, which you can see at the bottom, it's a silk screen printing place, and first rule of Print Club is you've got to print something, and I think it's really good to get away from the computer every now and again, talk to some people, get your hands dirty, get your clothes dirty, and learn something slightly different, and one of the things you notice when you're screen printing is every color takes time, effort, ink, soap, and so these artists have come up over the century, it's been around, with really efficient ways of using limited color palettes, and kind of using tones and patterns, and I have this really interesting graphic language, which we don't tend to explore very much in 3D, so this just seemed like it would be really cool to explore it. So, yeah, the other thing that's really useful at the start of a project is sort of collecting a mood board, right? And none of these images is going to have everything you want, but they all have some kind of feeling that you want to get across, it's not something you could put into words, because if you could, you wouldn't need a mood board, right? And they don't all need to be the same artist, or the same medium, or even the same century, really, but they all kind of help you crystallize this look, and it's also really good to have something to refer back to as you go through the project to remember what was the spark that you were trying to capture. All right, so this is the sort of figure out what you want to do, and now let's go into more detail about how we actually make it. So the first thing you'll notice about all of these images is they look very unlike what naturally comes out of Blender. They're not, you know, really taking advantage of ray tracing or beautiful lighting, and in fact, there's a super limited color palette, like in the screen printing examples, and also these wonderful textures, which we think of textures as like we're going to UV map it onto something, and it's going to curve realistically, but none of these people are worried about that. So they're doing some really interesting things, and, you know, it's a cool look. So how would we go about creating something like this? Oh, yeah, and then we'd also be doing style frames the whole way through. So these are just quick mockups to kind of get a feel for how some of these techniques might come together. So yeah, once we've got a good idea of, you know, figure out what we want to make, then we're on to the making it. So onto some less arty stuff here. So I used to work with a load of programmers doing video games and stuff like that and sort of try to absorb some of their massive knowledge. And it comes in really handy with shaders, because even though it's nodes, it's programming. I know it's not quite as glamorous as art, but you can be an artist and a programmer, I think. So a lot of their tips come in really handy when you're building something. So plan everything. I didn't do this for the presentation. That's just for me. Label everything, right? If you don't have time to label everything, you definitely don't have time for what happens if you don't label everything. And then keep it simple. So human brains, mine especially, can't understand that much stuff at once. You don't want to be looking at a plate of spaghetti graph. So break it down into blocks. And then also once you've got a block that works, don't copy paste, just reuse it. Don't repeat yourself as the programmer thing. And you can see, for example, this is our palette. And that little 15 means 15 shaders are referencing it. So if we find one color is a bit off, we can just tweak it a little bit. So these are all part of figuring out what you want to do and then doing it. And yeah, so we because we want to do all these graphic things, we sort of built up a little library of basically the functionality you have in Illustrator. So these are all real-time EV shaders doing various bits and bobs. I don't have time to talk about them in detail, but we'll put up a link maybe on the comments for the video, and you can just download them and have a play around. So these are all like little building blocks that we can put together to build these graphic shaders. And so yeah, with this stuff, oh, that's very dark video, isn't it? We quite quickly are able to build up some really interesting looks. We figured out what we want to do. And then making it is actually not that hard. I've seen a few people, particularly in the deduce presentation yesterday, that were doing some really nice half-tone stuff. One of the things we're doing here that was kind of fun is using the object's location in the shader so it kind of sticks to the object while turning to face the camera. And you can do some really fun stuff. These are just us messing around, really. And you can see it's all real-time in the viewport. And you can get these very interesting shapes and patterns from very limited pallets. And then once we've got all of those, we can just sort of slap them on whatever we make. And it's going to look sort of cool. It's the hope, at least. So yeah, once we knew what we wanted to do, this was not that hard to do, to be honest. Okay, so we've got our color and texture. And then the next thing with the look is shape. And this one's a little bit more involved. So we want to go and look at some of these images in a bit more detail. And now some of these. So a few reference images here. That's George Greaves, Gratia Lam, Aisha Frans, of course. And I actually thought this was Fry Alder, but someone told me this is Norwegian for free entry. So if anyone knows who that is, I really want to know. But yeah, so all of these look very unlike what we're getting in our blender viewport. Not just the shading, but something about them. And it's helpful to really get in there and analyze what's going on. So one thing you'll notice is all the horizontal aligns, they're all aligned with the frame. So it looks very neat and graphic design-y. Same with the vertical lines. So we're already thinking if you've done drawing lessons, you might be thinking this is one point perspective, right? And it is. But actually, if you look where that point is, when you take a picture with your phone camera, that one point perspective is going to be in the middle, right? None of these are in the middle. That's way off to the side. Fry Alder's one isn't even in this presentation. It's like up in the sky somewhere. I didn't even know you could do this. So easiest thing to do would be not do this and then just go back to blender and carry on. But we decided what we wanted to make and we're going to make it. So let's figure out what we can do. So I think about this. Actually, remember the conversation I'd had years earlier with a guy me and Rosa used to work with the photographer, James Medcraft. It's one of his pictures. And you'll notice it's not one point perspective, but there is something weird about it. So it's two point perspective. So a little bit off the horizontal. But all the verticals are parallel and the vanishing point is way off in the center. So it's like way over there, below where you'd expect it to be. Now you can do this in post or you can take a big image, crop it off. But Medcraft is a purist and he would never do something like that. And actually the way he'd been doing it is with something called a tilt shift lens, which is designed to do this exact kind of perspective distortion that these artists are doing it intuitively. And once you know what you're looking for, this is actually in blender already. It's just a button you have to drop down. Apply that with the graphic shaders and you know you've got something that looks really interesting and very unlike the stuff that you'd naturally think would be coming out of a 3D program. But the 3D program is actually the best way of doing these and they're just really fun, right? They have this, wow, these are way more zoomed in than they're supposed to be, so you're only seeing quarter of the image, but thanks PowerPoint. But you get the general vibe, like we've already got something that's quite unexpected and very interesting. So we got it. Except we didn't quite because there's still some more really interesting ones. So this is Gracie Lam again with a different perspective, Tally Bear, Dominic Kesterton, and if we look at these, sure, the horizontal is horizontal, vertical is vertical. There is no vanishing point, like completely parallel, which again gives it this weird atmosphere that's really cool. So easiest thing to do would be ignore this, carry on, but we decided what we're going to do, let's do it. So, and actually this goes way back. So this is a thousand years ago, and you can see that this is a Chinese painting, but Japanese painting also does this. Vertical is vertical, and you have all these parallel lines, right? And you're thinking, oh, did they see like this? Did they not get perspective? But actually when you look at that, literally the bigger picture, they're working with a 20 by 1 aspect ratio, so perspective would look terrible, and actually this is much better way of doing it. And it has this weird feeling, like those of you that grew up in the 90s might be getting like a weird nostalgic feeling from this perspective, or maybe a violent feeling, because all of those side-scrolling beat-em-ups did exactly the same thing. And for the same reason, like, you know, 10 centuries later, it's still a very good projection for this exact problem. But yeah, so we're interested in this weird vibe that it gives you, how do we do it? So a little bit of research, a little bit of maths, it becomes obvious that you can do it with kind of an orthographic, but you need to skew it or rather shear it. And you can kind of fudge blender to do a shear by rotating non-uniform scale, rotate back again. And once we know what we're doing, it's actually not that bad to do. So this is really running nicely, it's not even geometry nodes, it's just an empty and when you parent stuff to that, you get this funny little thing. And you can see it's got this horizontal axis aligned, vertical aligned, and it has this weird motion sickness vibe that I absolutely love. Not everyone is going to like that. And you can see that we've thrown on all these graphic shaders from before and we're already getting something that is quite close to what we wanted to make. Oh yeah, so we don't really have time for this, but I'm going to throw it in. So I just didn't want you to think I hated anime. So I used to share a studio with some 2D animators and they do some very interesting things with perspective. And I had to play around with it myself, very amateurishly. And it also made this funny little tool for doing curvilinear perspective grids. It's an open source thing, it's around somewhere, this is before I learned Blender. And so 2D animators use this to kind of, they draw there, they draw there and they pan the camera across. So this one, for example, has two vanishing points opposite each other on the same bit of paper, which of course you can't do with straight lines. And what you do is you draw something cool on that vanishing point and then draw something cool on the other vanishing point and then move the camera and it looks like a 3D scene, but it's really just a flat image with curved perspective. And anime does this incredibly all the time. So that one we made, but the rest of anime, I actually don't know that one, anyone know that? But kill or kill, watch what happens with her elbow there, red line. And this red line example is incredible. So pushes the button, speed boost, and then I don't think there's even a word for this kind of perspective. The whole film is like that, it's amazing. And they're basically like we know with perspective smaller stuff is further away, but you can play with it and there's no reason we can't do that in Blender. Once we know what we're trying to do, we can just go ahead and make it. This is just messing around, but instead of smaller is further away, these are just like smaller is further away and dips down or smaller is further away and tilts to the left. You can do smaller is closer and bigger is further away. It's just in geometry nodes and then a little car in there. And I would love to see this with some anime shaders on. So if Dylan Goo wants to grab that, just throw some cool shaders on there. And yeah, who knows? Could be cool, right? So once you figured out what you want to do, actually doing it is kind of the easy bit back on track. So we've got our little scene. Now I need some people in. So again, we had the mood board and we're sort of going through figuring out what we want and what kind of characters are going to tell our story and fit in the scene. And yeah, so these are really graphic, very simple and it's pretty standard Blender stuff, you know, just build it, rigify, throw the shaders on. One tricky thing though is the faces, because you see the very graphic. And we played around with a couple of standard ways of doing it. We'd love to do it in grease pencil, but it doesn't quite wrap nicely around the head. If we had more budget, we could do it in meshes like they did on sprite fight, which is incredible, but we don't have the budget or the skill to be honest. So we'd been playing around with the shaders and thought, you know, let's do it with shaders. I'll be honest, it wasn't quite as easy as I thought it was going to be. And Rosie was like, Dave, are you okay? And I'm just getting deep into the Blender forums and weird bug reports and finding the limits of like, for example, EV can only take eight inputs from the object info node and really esoteric bugs, which culminated after about a week's work in one new line in the Blender manual, which is my proudest achievement. But eventually we got this working and it's pretty fun. So it's just a shader for doing cartoon eyes. And I got it a little bit carried away because I was like, whoa, this is cool. You can do all kinds of eyes. And we have a lot of different characters. So I was like, we'll make this really flexible shader that we'll do all the eyes we could possibly want. And it's just a texture so you can apply them to everything. So yeah, and same thing for mouths, you know, we already have all those nodes that do, you know, parallax mapping, drawing rectangles, just join them together. And it works great. Got really excited, started applying them to like spheres and cylinders and Suzanne's and, you know, what's coming next, right? This is actually my favorite gift. And then, you know, I really got into it. I was like, this is the way to do it. This is a really cool thing. And like spent the time doing a face rig. These are just like placeholder character designs. Got really into it. And like, by the end of that, and it was like a month's work, probably, had this sort of modular system that could animate different faces and even hooked it up to a really amateur motion capture system. Like, yes, this is cool. All right. So now let's do it properly. So I'm not a character designer and I sort of scribbled some ideas, but I like drawing spaceships. I'm not a poor person. So we went back to the mood boards. And one thing you'll notice here is there's quite a few images from the same artists here. And I know some of you are thinking, oh, we could get an AI to do this, blah, blah, blah. Just message them on Instagram and work with them because they're really cool. So we had Alex and Esther, both incredible character designers, getting involved and throwing some ideas into the mix. So standard thing when you're working with freelancers, you do a little brief. We had like quite a detailed these are the colors we're using. These are the patterns, a few notes. And they had quite a lot of freedom because we could just stick this face on anything. So just don't change the face, right? So instantly they're both coming out with really cool stuff. Did change the faces, but I'm like, oh, we can just replace that later. These are Alex's designs and you can see. So these are just 2D images laid onto the previous ones. And you can see it really fits in the scene and they're really cool designs and just stuff I would have never thought of. And it sort of works, right? And so then we went into more detail. These are some of Esther's designs and they're coming out with these really like different variations and we're sort of homing in on exactly what we want. And like we're normally a remote studio, but actually me and Esther were in the same place for this. So, you know, this is like half an hour's work, just throwing some of the shaders on a placeholder mesh and just figuring out how it's going to work. And of course, because we had those face shaders, there's no rigging. We just apply the shader all good. So by the end of the day, I'm like, yes, this is brilliant. We're going to be finished by lunchtime. And then looking at it, just like something's a little off, you know, like this space kind of works, but look how cool Esther's one is, those almond-shaped tilted eyes and those asymmetrical eyebrows and the diamond mouth. Like, then I'm looking back at all of these faces like, oh, cool eyes, cool eyes. Those mouths, our shader can't really do those, to be honest. So, oh, no. So what you'll notice if you've been paying attention is I did exactly what I've been telling you not to do. I made it before we'd figured out what we wanted to do. So I did it wrong deliberately for educational reasons. As soon as I finish the presentation, I've got to go back and do it again, but different. But yeah, we're getting there. And like as long as you stick to this whole figure out what you want to do and make it thing, you're going to get there eventually. So I just want to add one more step, though, which is something I hadn't noticed, but I think is really crucial. So meet lots of creative people and learn from them because all of these little tips and tricks like if we hadn't been chatting with photographers and programmers and printmakers and illustrators would have never occurred to us. Like, they're just outside learning to keyframe, learning how to do shaders. So yeah, just get away from the computer and kind of surround yourself with these interesting people and you're going to learn stuff that you didn't know you didn't know, which is kind of why we're all here, right? It's why we're all at this conference. And I hope you're going to take advantage of it because I've learned so much already. Thanks for coming.