Long life to Embark studio and its fabulous procedural artists dream team !
truly excellent and inspiring to read. Would have loved to read some on the texturing since that is top-notch.
great environment with a lovely serene sense. Thanks for the write-up!
Phillip Stoltz did a breakdown of his lovely cluttered bunker, showing how he created the assets, materials, shaders, decals, and lighting.
Thanks for having me back! It’s great to be doing another interview with 80.lvl. First, I just wanted to shout out to Arseniy Chebynkin for his incredible concepts! He’s a boss and I highly recommend people to check out his stuff.
I’m currently working as an Environment Artist at Amazon Game Studios. Unfortunately, the work I’m doing professionally is still under NDA so I can’t discuss that currently. But besides that, I’ve been working a lot on personal work like always! That is how this scene came to life.
When I initially looked at Arseniy’s concept of the bunker, I really liked the overall composition, lighting, and mood. It has this eerie vibe to it and made me immediately think of Dying Light and Fallout. So I think my decision to go more realistic was based on my desire to try and make this scene feel like it was in a Dying light or Fallout type of world.
As with any work I do, I always like to set goals for myself and try to achieve them. Some of those goals were: brush up my skills with Unreal Engine by making a small environment, make this scene look like it belongs in Dying Light or Fallout, create a dreary lighting setup that makes you feel a little uncomfortable.
And also I wanted to have the scene feel like someone definitely lived there and worked there for a time and had recently vacated. I did want to leave the story more open-ended and allow the viewers to question, “what happened here?” I took a lot of inspiration from artists like Clinton Crumpler, Corey Hill, and Harley Wilson.
When I finally decided to use this particular concept, reusability of assets was one of the first things I thought about. Just by looking at the concept I could already start making a quick initial list of assets I needed to build. But what really solidified the asset list came during/after the blockout phase. During that phase, all I care about is making everything in the concept with the correct shapes/proportions and making sure that most everything makes sense in 3D and reads well. I believe the total amount of assets I had to make was probably around 50. I think around ⅓ of the assets were reused; especially ones like the cardboard box, radios, lockers, trash, pipes. Reusing assets is always efficient and time-saving but when you’re doing a smaller scene, that can be noticeable pretty quickly. So I had to be cautious about which assets are reused and how many time I was planning on placing them in the scene. For example, I didn’t want to just make 1 or 2 radios because of how many I needed to place to fill space up. So I needed to make a couple more to add variety to that space.
I did the majority of the modeling in Maya and used Marvelous Designer and ZBrush for the cloth assets including the pillow. As far as elements that were complex to model; I think making believable sheets, cloth, wrinkles is always challenging because there is such a specific look that cloth materials have. Like between satin, leather, denim, wool; they each have such a different look and feel and the eye can almost immediately tell if something doesn’t look right. For this scene, in particular, I wasn’t worried too much about optimization because I wasn’t going to take this to a finished playable level. But it was still in the back of my head at all times. One thing I did to help keep things clean and optimized though was to group similar items in the same texture sheet and limit myself to using mostly 2k textures for everything. With the exception of some props, (ie: flags, paper) that only needed a 1k texture. For scale, I went with a default height for humans being 6ft and I would constantly run around my scene using the first person actor in UE4 and refer back to reference to make sure items looked as proportionate as possible. I think from a design standpoint, choosing how to fill out the opposite end of the room that wasn’t in the original concept was difficult. I needed to come up with new elements that enhanced the scene and story without going overboard.
For this scene, the grid was really only used initially to get the scale of the walls, floor, and ceiling. I did make sure that everything snapped the grid and that the walls and ceiling could be used in case I needed to make hallways or other rooms. So there weren’t too many fancy tricks used here. If the scene was more elaborate in size and scale I would’ve paid more attention to the functionality of the grid. However, I needed to make sure that the space that was created would be enough space so a player could actually walk around and move without feeling claustrophobic.
I really only used trim sheets for the walls and ceiling pieces and also the pipes/wires in the scene. Using trim sheets for the pipes was just really fast and efficient because then I didn’t need to spend lots of time baking out new assets and trying to piece it all together. The rest of the assets were grouped together based off of the type of items they were. For example, the radios and cardboard box were all on one sheet and the bed frame, sheets, mattress, and pillow were all on one sheet. Since I primarily use Substance Painter for my texturing, one of the challenges I often face is getting all the color, roughness, and metalness values to display correctly. Luckily now, Painter has a direct export to UE4 which is a great time saver. I also like to use the export option in painter for UE4 which will organize the different textures into a ROM (Roughness/Occlusion/Metalness) format. And the benefit of doing that is that there are fewer textures to keep track of and it is more efficient for the engine. This time around I actually didn’t use a master material file for the environment.
I don’t know if I used any particularly interesting shaders. Every shader was pretty basic even for the emissive and translucent shaders. For the concrete materials, I went with using downloadable textures and some compositing in Photoshop to get the worn look. For the exposed brick that’s on the concrete; that was using vertex painting within UE4. For all the metals I wanted to establish the base roughness value so everything looked even across the scene in terms of how opaque or shiny something looked. That way when I was in Substance Painter I could use my defaulted value and not have to worry if it was going to look correct or not.
Decals are the best! And you are absolutely right, they can make a world of difference. I knew as soon as I got the concrete walls up that I was going to need some good decals to break up the scene and give the viewer some interesting things to look at so the scene didn’t feel flat. I used a lot of destroyed looking decals and broken wall ones since it was relevant to the look/feel I was going for. I also used some grunge floor decals to give some wear and history to the bottoms and tops of the walls.
Lighting is always a fun challenge to tackle. I wanted to stay as close to what the original concept had since that was one of the things I found most compelling and still wanted to capture that as best I could in the scene. One of the goals was to have the lighting be mostly dim in areas where the main light sources weren’t so it always had that eerie feeling. I went through many iterations throughout the lifespan of the project, a lot of it was experimenting with color and intensity to get the right look. I used point lights and spotlights for the scene. Something that I used to get the atmosphere for the scene was an exponential height fog. That really helped to give some depth and make things appear a little further away than they actually are. It also added a nice brown haze over everything as well which gave it a sickly feel to it. At first, I was using dynamic light to help with making quick changes. But once I found a light setup I liked I began to bake the lighting because it had a much more powerful effect that I liked.
I think overall the biggest challenges for me was being able to capture the concept in 3D and still put a little of my own flavor and storytelling elements to make it a little different. And getting compelling lighting in the scene to enhance it and not hurt it was also something that was really fun to try and accomplish. Overall the scene took about 2 ½ months working on the weekends and in my spare time.