Robert Hodri shared his extensive experience of modding and sci-fi environment design, talked about the changes that happened in the industry over the past couple of decades and discussed his workflow.
Getting into 3D
I think my path basically started about 20 years ago with the release of Quake 3 Arena which got me addicted as a gamer. After spending hundreds of hours in multiplayer matches I decided to look into the tools and editor that shipped with the game back then. There weren’t that many tutorials or documentation which made it quite difficult to get into it. I was mainly messing around with the editor and creating lots of box maps that weren’t very exciting to look at and no fun at all to play. After a while, I gave up and didn’t do anything with the editor for a very long time.
All that changed when Doom 3 was announced at MacWorld in 2001 and id Software showed the first glimpse of their new engine in a short tech video. The visual quality was quite impressive for the time and got me back to 3D again. The game would come out in 2004 and I was heavily getting into the Doom 3 modding community and making my own single-player levels for it. I spent lots of time in online forums and made friends with like-minded modders who shared the same interest and passion. After a while, I would also get more and more into modeling and texturing so I could give my levels a more unique look, which got me into 3ds Max and Photoshop. Every time I finish something (a level, a 3D model or just a texture) I would post it in the game art and level design community forums to get some feedback on how to improve. Most of those forums aren’t online anymore or barely active.
In order to grow and learn more about levels and game environments, I would play a lot of first-person shooters and analyze every combat space that was fun to play or just amazing to look at. I’d try to understand what made that area fun. I would have tons of screenshots and notes for every game I played. Same for movies, whenever I saw an interesting scene I’d look at it closely and try to see what makes it work. The colors, composition, framing, cinematography, etc. Looking at other people’s work can be both inspiring and demotivating at the same time. You’re eager to improve and get on the same skill level but then you realize it’ll take you a long time to get there. So in order to grow, it’s also important to be patient. Especially at the early stage, it can be frustrating. You’ll often get honest feedback that might hurt your feelings and you won't agree with it because you think you created the best piece of art. But you have to step back and look at it with objective eyes and learn from your mistakes.
Lessons Learned from Modding
I think the most important lesson I learned while doing mod work is to be patient and persistent. Sometimes, you encounter a problem, get stuck or have a question. You can post it in a forum and maybe get an answer after a couple of days or no reply at all. I also didn’t have an internet connection in my apartment so I always had to go to an internet café or the university library to check for responses or ask my questions.
In the early modding days, it wasn’t too uncommon to simply reach out to the developers and ask them for help. Some of them were nice enough to take the time and help you out. Back then, game developers weren’t big business corporations and were more intimate. So it was quite easy to reach out to the right person directly. I remember getting a response email from John Carmack (id Software) explaining to me how to use renderbump in the Doom 3 engine so I could bake down my HP models to an LP mesh. Now, I think, most bigger companies have “community managers” answering all kinds of different questions on their official gaming sites.
Making mods is a very long process which can take years because we all used to do it in our free time besides work or school. Even now, there are lots of Quake/Doom/HL mods still in work with no actual release date. I think, all the single-player campaigns I did for Doom 3 and Prey took me about a year to finish since I was the only person working on them. But the positive feedback I got from the modding community and other designers and artists was quite motivating. Releasing my multiplayer levels was a different case - they only required a couple of weeks to finish but barely got any attention because no one would play them.
When I was approaching the end of my studies at the university, my time was getting more precious. I focused solely on making visually great-looking environment scenes to improve my portfolio so it would be easier to land a 3D environment job somewhere. That’s also a bit of advice I would give to any aspiring environment artist: focus on props and smaller environment scenes to showcase your art skills. No need to make your own mod or sp/mp levels unless you’ve got the time for it.
How Tools and Industry Changed
Of course, tools and software packages have improved a lot over the last two decades. 3ds Max and Maya are still the most used tools but then you also have new ones like Substance Painter/Designer which is replacing Photoshop more and more for texture creation. Sculpting tools like ZBrush are not only used to sculpt organic meshes but also utilized for hard-surface stuff. CAD programs like Fusion 360 seem to be used more often in games now, too.
Game engines and the integrated tools speed up content creation heavily. Material creation, shader setups, vertex painting, terrain tools, lighting/VFX/cinematic/scripting editors, etc. - during the early days all that was done in text files or modeling software.
Those are only the changes in tools but the whole industry has changed as well. Sometimes, games still ship with an editor so you can create content but it’s more limited, of course. I don’t think any company would ever release their source codes ever again. You’ve got DLC season passes and new content created by the main devs now.
If you look at the end credit scenes of any major AAA game, you'd see that back then a level designer would fill a lot of different roles: from early blockouts, combat encounters, gameplay, scripting to art stuff like lighting or general detailing of the level. He/she would be responsible for all those disciplines. Now roles got way more specialized and new positions appeared, like texture and shader artists, terrain/foliage artists or even groomers - people who are responsible for a character’s hair.
Creating a Level in a Team: Workflow
Normally, the scene starts with a blockout from a level designer and a concept to work from. Sometimes, I spend a bit of time gathering some reference images that I feel could help me. When doing so, I avoid using other games or popular movies for reference to keep things unique and fresh.
Workflows and tools have changed and improved over the years but the approach to building environments stayed pretty much the same. Focus on interesting shapes and the silhouette first, do the first texture and material pass. Talk to your designer to make sure not to interfere with the gameplay space or how many changes you can make to it. Rely heavily on the tile textures and make sure the texel density is correct. Don’t scale meshes up too much since that can give you blurry results or simply look wrong proportion-wise. Detail out areas but keep readability and player paths and enemy encounters in mind. Try to reuse assets in a creative way to save time. Whenever you find an area looking too bland after the modeling/texturing pass, you can still rely on stamps/decals to add more details.
Lots of engines support detail normal maps which also helps a lot in making things look crisp from close up. You can always go back and add more modeled-in details and do more texture and material breakups.
For props, I still use normal SubD. I haven’t really gotten into the rounded edge shader from Modo but I’m also not baking down many assets anymore.
You can have millions of triangles per scene for modern games but there's still no reason to waste them. Make sure the polys you add help you to support the shapes of your model. Also, think about vertex painting in advance - you can add more polys in the areas you know you’ll vpaint a lot on. When you know that a certain asset will only be used for mid-distance or far away don’t be too wasteful with little details. Don’t model everything out, use geo decals or trim sheets to add more greeble detail. Check your assets' LODs and make sure to avoid too much popping and all LOD1/2/3 looking nice.
Sometimes you’ll work on a cinematic scene where there’s no combat or gameplay. Then you can go a bit crazier and make sure everything looks smooth and nice.
My workflow for UVs hasn’t really changed that much over the years. I know there are lots of external UVW tools that people like but I always use the UVW editor that comes with 3ds Max and Modo. I used both of those programs for production over the last ten years and they improved and added all the things I need.
I use world space UVs a lot for a quick texturing pass and hotspot textures for environment art. For unique UVs like a weapon, I still pack everything by hand. Texel density checker, projection from view, peeler, sewing/splitting/welding - those are the common tools I use, nothing special.
When working on larger environment scenes in Modo, I make sure to always use .jpgs with at least half the resolution from the original texture. Loading in .tgas can cause extreme slowdowns.
The texturing process can depend on the project and the pipeline. These days, for sci-fi environments, it seems to be quite common to use custom vertex normals and geo decals with fairly simple tile texture materials. There’s a trend to rely less on baked down assets which saves time. But as I said, it depends on the project and the game type. I’m talking mainly about workflows for AAA first/third-person games.
Trim sheets are still relevant because it can be fast to just quickly UVW a low poly trim sheet instead of adding tons of geo decals. I normally start with very clean base materials and add grime/dirt/scratches later on with vpainting and decals.
Advice for Environment Artists
Be familiar with the 3D modeling package and the game engine you're using, know common workflows like HP/LP baking, custom vertex normal, geo decals. Try out new software and implement it into your workflow. The job can be very technical so try to understand how things work. There's no need to get into programming and writing scripts but a technical understanding of what normal maps are, how game engines work, etc. is always useful.
For me personally, it was a huge help having designed levels and gameplay spaces in the past. Getting into traditional art or photography is also great. Studying color and composition will help you a lot with creating interesting environments.