Joe Perez talked about the way he creates fantastic 3d environments for iconic adventure games.
Joe Perez talked about the way he creates fantastic 3d environments for iconic adventure games.
Hi, my name is Joe Perez and I’m a 3D artist that has been working in the industry for about 7 years. I’ve worked my way up from making small but very profitable Facebook and mobile games to a 3D MOBA made in Unity with a studio called Kixeye. When I saw that one of my favorite companies, Cyan, came out with a Kickstarter for Obduction, a spiritual successor to Myst, I became overwhelmingly excited!
As a kid, I played the first Myst game and up until then had not seen anything so beautiful, immersive, and groundbreaking in a video game. The main character was the environment, and it caused my mind to explode in creative wonder. Because of the game, I knew I wanted to try my hand at 3D art, so I spent a lot of time noodling around with a few programs such as KPT Bryce and Poser in order to make my own environments — eventually creating a small game in high school in the same vein as Myst. It was then that I realized I wanted to spend the rest of my life making video games. Once Cyan shut down development of big titles in the late 2000’s, I was saddened and felt regret that I would not get an opportunity to work there as an adult in the game industry.
So, during Cyan’s Kickstarter revitalization, I knew I had to throw my hat in the ring! I set out to learn Unreal Engine 4, and create a new environment based on concept art by Stephan Martiniere so that I would be noticed. Luckily, they liked my work and I was brought to meet the team and eventually got hired–accomplishing one of biggest goals in life.
When you are creating an environment for a Cyan game, you know that the game is slow-paced and that gamers are going to be taking their time looking closely at every nook and cranny. This is because the world is the main draw and it is needed to convey the story and laws of the universe in a logical way. You also come back to many of the areas you’ve seen before but with more understanding. Things clicking in your mind upon better understanding your environment is one of the main reward systems that Cyan concentrates on rather than getting experience points, for example. Many games today are fast-paced and gamers tend to run into things quickly because that is what game-flow encourages. In a fast game, it may not be quite as important to spend a lot of time on a specific section when the player will run by it and never see it again. In a Cyan game, artists are encouraged to create details that may not be noticed by most players but that will create fascination for those who are patient and search for the depth. Creating this kind of environmental depth was the best thing I learned while working there, and it was intensely valuable.
One of the ways I approach finding the most important areas of the environment is by putting myself in the players’ shoes and exploring a blocked out area in VR. I try to imagine where I will be looking and how I’ll be moving through the environment. In a sense, it’s like a small movie I play in my head that has an order to how the environment is revealed and occluded–I not only go for readability and spacing in these environments, but I also want to convey certain emotions whilst keeping the sense of wonder in the player. When an area has multiple ways in and out, it becomes more difficult to plan out, but it’s usually more fun that way. I generally try to create areas that are presented in an original way when compared to what games have done in the past and that works well with the puzzles. Although, there are exceptions in the way of Myst homages we have put in Obduction.
Lighting is also very important to guide the player and to let the shaders really fire on all cylinders. It also helps to convey the mood I’m going for.
When I first came on, the entire game was blocked out but it was very rough. In the end, much of the game changed, because over time there was an evolution of the game in terms of goals and game flow. So, I did do a good amount of blocking before entering production in the environment, though after making the preliminary blocking, I would go in and create higher level details so that I could get a better sense of how the lighting and silhouettes would work in the area. There weren’t any textures in this phase, just very rough and quickly assembled models that would be replaced later as well as some preliminary lighting. By creating this sort of “mid-level detail,” it helped in planning out modular pieces I may be able to use or not use in the given area as well as plan out the composition and generate new ideas in order to improve it. Those at Cyan are used to wearing many hats and helping to plan the environment from an early stage to final polish. This workflow worked for me personally because I liked having a high amount of control over the work I was doing. I felt like I could put more of myself into the game and feeling a higher sense of ownership.
When planning areas out, I get an initial verbal breakdown of what the goals are for the player in a particular area given by my art director, Eric Anderson and the CEO/game designer, Rand Miller. These breakdowns are usually accompanied by drawings on a whiteboard. We would also get in group discussions often that included the concept artist, Derrick Robinson, to see if any new ideas floated up to the top. In some cases, blockouts were made by the concept artist for me to work from if it was something he had a clear vision for. In terms of the following image, I knew that there were two draws–the girl hunched over the desk, and secondarily, the door behind her. I tried to focus the environment in a way that would show these two important things in exactly that order.
Everyone had a say, and that resulted in we ended up coming up with some stellar ideas that I would not have been able to come up with myself had we not huddled together. I am a firm believer that no one has all the answers, and the best art often comes from a group effort. The main thing we were fighting against was the time it would take to create these areas, so we had to make concessions at times, but I think we did well, considering the time and money that was dedicated to the project.
As for the small props littered everywhere, that was mainly done by other artists in order to help me concentrate on building macro scale areas and how they would transition into other areas. I am very thankful for the support in terms of props because I had a ton on my plate just thinking about how these areas would coincide!
The Art Director has been with Cyan for many years, he understands Myst greatly and always had his door open to us at any point in order to help guide us with well-explained and creative ways to solve problems and to better direct the player.
Reusability is definitely different depending on the type of game you’re making. In Obduction, we also had a lot of modular pieces, but we probably have a much higher amount of custom pieces due to the nature of the areas and puzzles in the game. In Obduction, the puzzles need to be integrated into the environment in a believable way, which means that not only do the puzzles need to be original and one-offs but that the surrounding area also needs to accommodate the puzzle and story. Secondly, we have sphere cuts: these are areas that can be swapped into other worlds either as a gameplay mechanism, or story mechanism, so each of these had to be custom made. The areas in these sphere cuts were of all sizes, and since they were swapped from other worlds (such as the train station), they needed to have original assets that would live only inside of that sphere cut.
Lastly, an area like the central tower in Hunrath had multiple ways in and out, and also included an elevator that went both above the structure and below ground. This area was very complex to plan and I had to create a specific sculpt of the giant rock in order to accommodate each of these areas that coincided with the exits, the bridge, and the tower elevator. This area has some of the least amounts of reusable assets because of its complexity. That said, I was still able to get away with using a decent amount of modular assets when it came to rocks inside the tower, wood pieces, corrugated metal, and stone sets.
We had a texture library and some great master shaders that were written by my art director. This is more proof that we had to wear many hats considering the Art Director was also our Tech Artist. The texture library was largely built from large photos that were modified and 3D sculpts that were repeatable and created inside Zbrush. Our workflow was that we would sculpt a mesh, say a rock, and we would export mask maps along with a normal and an albedo map with roughness information in the alpha slot. The masks would have information such as depth information and edgewear, and then we could use vertex blending and shader controls to paint on edgewear amount, wetness, and other details such as strata that is world aligned, that would help make every rock look unique.
We also had a kit that we called “stretchables,” that would allow us to make many types of metallic and wood pieces that would all have normal and edge maps. The shader would help us make each wood or metal piece have slightly different tiling and tints, we could also blend things in like grime that would adhere to the ambient occlusion or edgemap in order to keep the detail in controlled areas. The shaders, in general, were our biggest asset in the most rapid development of Obduction, and it helped us out immensely.
When it came time to work on our next project after Obduction, I taught myself how to use Substance Designer and Painter. So, most of the textures in our game ‘Firmament,’ are done with Substance. It was difficult to start learning Designer, but I’m glad I did because now I can get better quality in decent time that looks very close to something sculpted. In the next image, you can see how I’ve done exactly that in these rocks.
With 2D games scale of things is important, but in VR it is of much higher importance. The Oculus and the Vive headsets allow you to modify your interpupillary distance, which means that the scale in-game can match the scale of the real world as long as you set it correctly. So, if things are not to in scale in the game, then you can tell instantly in VR–whereas in 2D, you can actually get away with it at times.
In VR, another thing that is strengthened is your sense of the foreground, middle ground, and background, so I tried to keep that in mind when making my environments. In this DLC area I worked on called the Crash Site, I intentionally made the walls very close to the player so that it felt like you had to squeeze through in order to get revealing shots of the airplane on your left, and subsequently the bar to your right. In Resident Evil 7 VR, I noticed that squeezing through tight areas inside of the walls felt very immersive and visceral, so I wanted to bring some of that feeling to Obduction in this example.
Lastly, effects in VR look amazing. So I asked to have things like floating particles that followed the player in the prairie, or dust particles in other areas. These things are subtle but very powerful to give you a better sense of place.
One of the biggest challenges in my work stems from wanting to do the best and coolest things while balancing time frame because there wasn’t a time when we didn’t feel like we were behind. Imagining what an area will look like given a specific time frame, rather than what your wildest imagination can come up with was a definite test–It was often enough that I’d ask my art director for more time in order to try something with graphics, and he would do his best to give me that time.
I think playing other games is important when learning about game development because you can see how they have tackled many similar problems you may also have. I see proportional problems in games more often than I’d like, and when certain games are adapted to stereo 3D and/or VR, these problems stand out further.
Another theme I see in many games is lack of world building and allowing the environment to help tell a story instead of leaning on something that looks shiny or glowy in order to build interest. When games don’t put enough focus into this aspect, it leads to video game levels that seem like they are too much of a slave to level design and they lose a personal connection with the player. Though, I do think some games have more challenge in this area than others–it’s more difficult to have great world building in a game like Super Monkey Ball for instance.
These are exceptions, and I do think we are in a great place with a great trajectory in terms of game environments today. I’m just glad to be a part of it!