Super taf! ;)
Ted Bundy's car? :D
Senior environment artist Felix Leyendecker talked about the way he approaches the creation of virtual spaces for high-end games.
Hello. My name is Felix Leyendecker, I am from Germany and 35 years old. I always loved gaming and especially first person shooters, so the industry was always something I wanted to get into. I first took up computer science at university, but changed my mind and went to art school instead. I graduated at 24 and had a job at Crytek lined up already, thanks to a graduation project I made in school. With Crytek, I worked on Crysis 1, 2 and 3, and did a little bit of work on other projects as well as some RnD/pipeline work. I’ve worked there for almost 10 years before moving to Texas to work at id in 2015. With id, I’ve worked on DOOM, including its three DLC map packs.
Environment Artist’s Responsibilities
It varies between studios, but generally, it is kind of like Lego. Most of the time you’re given a rough, simplistic blockout by level design which you have to stick to, along with some concept art. Then it’s up to you to decide how to break down the work. What needs to be made from scratch, where can you use tiling textures and where do you need uniquely unwrapped props? Does a modular workflow make sense? How long is it going to take you? How do you delegate the work to other artists?
Then you start hunkering down and get stuff done. It’s definitely more of a slow burn process and you won’t see much progress until the very end. I’ve shipped many projects now and it’s always been the same where everything seems to magically fall into place during the last few months. The reason is that you have to get beyond a critical mass in production, where you have a large enough asset pool to choose from, and enough routine/familiarity with the people and processes. Once you’re beyond that point, progress comes pouring in very quickly and the end result is (hopefully) a pretty game!
Preparing the Assets
It depends on the concept. If you have a very accurate and detailed concept or if you’re working from photo reference, then the 3d process is rather straightforward. If the concept is kind of loose or just doesn’t work in 3D, then you’ll have to do some rather significant changes to proportions or shapes. This tends to be the stage that takes the most time. Adding details or refining shapes later on is comparably fast. How this process differs from 2D work is that the camera perspective isn’t fixed. You can walk around objects and look at them from multiple angles, the composition of an environment changes depending on your vantage point, and so on. Good concept/2D artists are aware of these things and structure their work accordingly, with quite a few of them incorporating 3D modeling into their workflows to tackle precisely these issues.
Many artists tend to overestimate the number of variations that modular assets need. Often, you can just mirror something on an axis, and it becomes a new variation. With modularity in general, it’s worth thinking about if you actually need it or not. Unless you’re very experienced with this sort of thing, coming up with a modular system can easily cost you more time than you end up saving by using it. If you’re making a space that is rather unique and won’t repeat, it’s often better to just go ahead and build it, rather than spending weeks coming up with an elaborate lego-system where everything snaps together perfectly.
Of course, this also depends on the type of game. If you make an open-world game with tons of content, then it’s absolutely imperative to be modular. If you make an old-school linear campaign, maybe less so.
Then there’s subject matter. It’s comparably easy to be modular with organic environments, since for example with caves, you can just jam pieces into each other, they can overlap and it will still look good. Whereas with sci-fi or hard-surface environment, every trim, column and segment will have to snap together perfectly, or you will get ugly gaps and seams.
It’s a tricky subject, because gameplay and the playability of a space is always the highest priority. The only way to ensure that compositions work is to involve an artist in the level design/blockout process early on. There should be a back and forth between designers, 3D artists and concept artists to make a space that both works for gameplay as well as for compositional aesthetics.
Since the player is constantly moving, there’s no such thing as a fixed image composition in an FPS game. You will have to pick a certain amount of positions in a space where composition matters. For example a door that you step through, which gives you the first look into a room. A vantage point or watchtower that gives you an overview. Or a fork in a corridor that presents the player with several choices. You will want to prioritize your composition efforts to such instances.
Playing into large-scale composition is also the next step, the amount of detail and the 70/30 rule. 30% of the space should contain 70% of the detail. This applies to both props, assets or environments. You need some large, empty and calm surfaces to contrast with detailed areas, or otherwise the image becomes noisy, overwhelming and hard to read. It will also be beneficial later when the final touches are being put on decal work and lighting. Surfaces that looked empty and boring during the building process will end up looking interesting and cool later, since they act as canvases to place interesting lighting or decal detail such as dirt, lettering or light gradients from wall-mounted lights. It’s important to keep this in mind, since especially if you’re an inexperienced artist, it’s easy to feel obligated to cram every inch of a space full of detail. But it’s better to show restraint and localize your efforts.
In my latest work “Empyrian”, a DLC multiplayer map for DOOM, we had precisely this issue. The map is circular and shares the same architectural style in all sections. So it was easy to get lost and not realize where you are in the map. So what we ended up doing was dividing the sections up into color themes to work as signposts for the player. One area was the red area, another the yellow area, and we also had a blue/cyan area.
Since everything is made from marble, and we didn’t want colored marble, we introduced color in the form of set-dressing objects, lighting and particle effects. We added glowing stained glass windows that lit up in the color of an area. Then we added flags/banners that matched the area color, and also matched tertiary small details like torch flames, glowing floor elements and even colored ivy growth to introduce more color. We also added large-scale elements like colored lightning effects that shot up into the sky like a lighthouse to represent each “end” of the map, blue and red. This would be visible in the sky, no matter what was in front of you and couldn’t be occluded by anything else.
Finally, we matched the particle effect color of jump pads, teleporters and portals to the area they were leading to, so you could figure out how to get to a certain area just by color. The lower area contained two jump pad ramps, the one with the blue effect lead you to the blue area, and the red effect lead you to the red area.
Lighting has to serve both the aesthetics as well as the level design. Lights can be used as important cues to tell the player where to go and what to do. As with other processes, it’s best to involve lighting artists early on in the design/art process to arrive at a solution that is the best compromise for all disciplines. Lighting can also have a big influence on an artists’ workload. Areas that are highlighted by lights appear more prominent in the composition and thus need to have an appropriate amount of detail. Conversely, areas that are dark or in shadow can be more simplistic and don’t need as much work. Which is why it’s important to nail lighting as well as FX early in the process, since it can save the environment artist a lot of time down the road.
How do you make an environment piece that is usable, not too heavy and versatile enough to work with various materials?
It depends on the subject matter and the engine. If you have an engine with very effective texture streaming and a high streaming pool budget, you can get away with making big assets that are either entirely uniquely textured, or are unique for the most part. These assets have to be used a lot to make such an effort worthwhile. For example, in Argent Breach and Empyrian (DLC multiplayer maps for DOOM) we had a selection of architectural styling elements like column pieces, decorative elements, and cinder blocks (large masonry/stone slabs) that were uniquely baked and textured. It was worthwhile because we got a lot of use out of them, and they also added a lot of detail to the areas that looked good.