Join David Lesperance for an in-depth discussion in this video David Lesperance, Digital Environment Design - Film, part of David Lesperance Digital Environment Design.
[David Lesperance]: Whenever I played video games or watch movies, I was always looking at the backgrounds. The sets, the vistas. I've always been really fascinated by huge, grandiose environments. Characters are great, but they're just never my thing. In environments, there's a lot of technical restrictions and limits that exist when you're creating that kind of content, so it works really well on both side of the brain, I think. All my personal art pieces, they're just studies. I'm doing it just to try and learn something new. For me, I'm much more interested in the studies and the processes behind the things than I am necessarily just the outcome. The outcome of your skill set typically is a product that gets shipped, that's your day job, that's what you do, but the personal art and that kind of stuff, that's for you. You get better at the stuff the more you do it, and whether something's finished or not is kind of irrelevant because you can look at something and go, "Oh, this is done," I can guarantee you come back in a few months, you're going to see a million things wrong with it. One of the first things that typically I'll ask for when I see somebody's portfolio, I'm like, "Oh, cool, professional work looks awesome. "It's great. What about your personal work? "What are your ideas?" Like, "Oh, I haven't finished anything." "Let me see the stuff you haven't finished, "let me see the stuff that you're playing around with." You're going through and you're doing an environment for a game, you want to be pretty sure that the art you're making is correct. You don't want to make a really cool-looking sci-fi pillar and realize that there's not enough room for the character to walk by, so for that you start off with a massout. It allows people to see their ideas much earlier. I'll start laying out primitive versions of the scene with a camera in mind. It's all primitive, it's all boxes, and this will stay this way up until I get the composition locked down, in a position where I can start lighting. If you work in pieces and you can instance stuff, this is actually kind of a fun little thing inside of Max. Let's say this is my holdout for this geometry, and I'll have five or six instances of them in the scene. I take one of these guys out and I'll work on in in a separate Max scene. Once I do that and I'm happy with the general shape, I'll come back into 3D Studio and I'll go, "Okay, I have my blockout geo, my new arted asset, "I'll come in, I'll drop an edit poly modifier on it, "move this thing into place where the instance should be, "and I'll attach it to that position." Now, what that't going to do since this is an editable poly in its instance, that geometry's going to propagate to all of those other boxes. Your layout, your initial layout, the initial block out of your scene, if you use it correctly and you leverage it, it's a huge time saver because once the composition's locked down, you're literally saying, "this box is going to be "this awesome mechanical chest", or whatever you're doing with it. I was trying to figure out ways of trying to meld traditional hard-surface kind of CGI modeling with more of a game modeling mindset. Basically, what I'm doing is I'm trying to find pieces of geometry that I can use to build other things with. This floor looks like it's one solid mesh, but it's not, it's just a piece of a mesh. Same thing with this, this looks like its own little mechanical piece, it's broken up into a bunch of small little pieces that I use to go in and I can make bigger assets or flip them around and make it look like it's entirely something different. The other thing that I really like doing when I'm constructing these scenes is, inside of 3ds Max, the nitrous viewport, I can get really good shadows and preview ambient occlusion and that kind of stuff, and the lighting really alters the way a lot of these forms are seen. Repetition can be covered up pretty quickly once you put something into shadow. You can really break up things, you can call things out, you can do a lot with them. Modeling with the lighting in mind is really, really useful for me. Since all the light's going to be pouring in through these windows here, I have to know what the shadow's going to look like, and I have to know what the material's going to look like early on. There's no way for me to get around it, and if I'm sitting here modeling, I might go through and have the window be the entirely wrong size, which is a huge pain, so as I'm doing it, I'm setting up the shaders and doing very basic UV projections on a lot of these assets. Most of the time, if I can avoid doing super-intensive UVs on stuff, especially for environment pieces, I'll do it, because most of what dictates how well something looks is not getting in there and highlighting each of the edges, that's not what makes it look good, it's lighting interaction, shader interaction, and you'd be surprised what you can get away with by throwing a tiling texture onto something and throwing a box projection on the UVs. Nine times out of ten, it'll look pretty good, and then if you need to go in and spot-fix stuff, you can do that as well, but for me, I try and approach things as broad-stroked as I possibly can, if I can do less work on doing UVs, it means I can spend go my time doing more work on something else. When you're doing a painting or a 2D painting you can come in and pull from what is called the z-depth pass. This is also referred to as a fog pass. It makes the space feel like there's actual volume and depth. In games this is used for all sorts of things. It's used for object-calling, it's used for depth of field, it's used for shader effects, it's used for all kinds of things. That's why I tell people to know more than just their discipline. If you just know modeling, you're only knowing a very small focus. There are some pipelines where all you have to do is just high-poly modeling. You don't have to texture, you don't have to do UVing or any of that stuff, but for me, that's kind of narrow. It doesn't leverage the full creation of the scene or the environment, because a set, there's a million things that go on with it. For a game, for example, you're going to have to go deal with collision, you're going to have to deal with gameplay, you're going to have to deal with lighting, you're going to have to deal with runtime performance. All this stuff happens at once. There are so many takeaways from other disciplines and other things that you do that will make your own art a thousand times better. If you don't bring that stuff into it, you get sterile, you get stagnant. A big thing is trying to share as much knowledge as you possibly can, because you get better just by showing your methodology, your analyzing, and when you're teaching something, the process of how you do that is under a little bit of a microscope. This is actually one of the work in progress tutorials I'm doing. I've had a lot of questions about how I go through and how I build out these jet engines, and I'm approaching this as I would for a personal art piece, so there's probably about five times the amount of detail (laughs) but the idea behind this is to show how much detail you can create in an object and what that detail does when you look at it in a render, because the bolts, the reflections, there are certain things that happen when there's inserts or edge hits or even threading that you do pick up on the render, and then it's up to the artist to figure out, "Is this something that is worth doing, "or not doing as a work in production, "or is it just really nice to have a really "cool engine with a bunch of kitbash pieces." The community, I think, is what really drives a lot of quality of the art because every other week, somebody does a cooler robot, a cooler environment, a better effect, and it just keeps getting better and better and it's because there's this competition, but there's this competition and willingness to help each other, which is a huge component of why the art looks the way it does. ArtStation is one of the art forums that are picking up steam, they typically use ZBrush Central pretty heavily, it's a really good positive environment, so finding something that looks cool or is inspiring is never hard. Technology and the tools change so quickly that I'm constantly trying to learn as I'm constantly trying to do art. People think of success in understanding this stuff is suddenly getting to a point where you "get it." I've never seen that. For me, I'm just constantly learning. Art in general, it's so complicated. People have spent their entire lives trying to understand one facet of their craft or whatever they were doing. You're going to spend years trying to understand it, but if you're so worried about trying to just be better, and just be the best, you're not going to really grow. I wouldn't have been where I am without all the tutorials that I've gone through and done, that's the stuff that helped me out, so if I can help some kids do that, or some people who want to try a new field, then I think it's a good thing, and I think the industry as a whole benefits. It's kind of a rewarding thing, to get an email from someone that says, "Hey, you really helped me. I put this in my portfolio," or, "I did it at work, my boss "really likes it, so thank you." That's super awesome.