Graham Smith from Inigmas Studios discussed the way he approaches the creation of beautiful stylized environments.
My name is Graham Smith and I’m the Studio Director at Inigmas Studios. While I reside in Apex, North Carolina, Leo Lutecki, my partner, is based in Gothenburg, Sweden. On our projects, I handle the pre-production process, which usually entails aligning to the client’s vision so that it’s smooth sailing during the development process. It also includes ascertaining the scope of the client’s art requests.
After the pre-production phase concludes, my team and I are ready to begin development. During this step of the production, I focus on modeling, unwrapping, rigging, skinning, animating, creating Level of Detail (LOD) models and all the Unity prep work so that the project is in a game-ready state for the client. I have a fairly streamlined pipeline that splits off usually after I’ve finished modeling and unwrapping—at this point, I’ll hand over the asset to Leo who will then texture it. More specifically, he focuses on diffuse, ambient occlusion, normals, specularity, emissive and alpha (usually inside diffuse nowadays) maps. While he textures, I’ll focus on rigging, skinning, animating and creating LOD models.
I also guide the art direction of our projects to maintain a certain level of aesthetic consistency across our work. When we first started out, this was difficult because we were still figuring out our art style; however, Leo and I have developed a tight working relationship and now have a firm understanding of our artistic strengths.
Hand-Painted Forest Pack
When we began working on the Hand-Painted Forest Pack, I led a team with the purpose of creating a turn-based, JRPG-like game called Obelisk; however it, like most unfunded teams, eventually fell apart. Yet, from the ashes of that team, Leo and I rose above our failures and decided to finish our art and publish it to the Unity Asset Store, thus Inigmas Studios was born!
All the art packs we make go through a pre-production cycle where Leo and I discuss the new theme of the pack, such as a desert or village environment, and then focus on ascertaining its scope. We then collect reference art from Google to further define the vision of our project.
We’ll usually have at least three to five variations of what we’re going to model with the intent of adding our own personal touch to the final product.
I personally create all of our assets in 3ds Max 2015. While modeling, I can sense and feel out how the model should fit into our art direction. We use a sort of “thick-to-thin” procedure to show subtle hints of flare and character to our assets. Check out our Coffin from our Halloween pack to see what I’m talking about!
I had the art style planned out for that turn-based JRPG while forming the team behind the project. I met Leo on Polycount, where I reached out to him and proposed the idea of the game and its art style to him. We eventually had a formal talk over Skype and began working as a team.
The art style pays homage to games I grew up playing, such as Secret of Mana, Seiken Densetsu 3 and titles from the Legend of Zelda series. I fell in love with the way Square exaggerated the color palette of those games—it really set the mood and atmosphere within those titles. The artists behind those projects were effectively able to convey different emotions to the audience by defined shading in their work. Cool colors were used primarily in shadows while warm colors contrasted nicely in highlights. All in all, this was standard practice back in the day.
However, those artists truly took it to the next level by exaggerating the hues in their shadows and highlights, which challenged the player’s imagination in fun and exciting ways—this is what I want to achieve with our assets! We start with a base color, then add warm highlights and finally follow up with cool shadows, which helps us sense a strong contrast and decide upon the best way to invoke a particular feeling with the audience. We handle all of this in Photoshop.
While our forest pack was the first to introduce a certain level of modularity through rocks and cliff sides, our village pack greatly furthered that level of modularity. When I planned the scope of this, I wanted level designers to have as much creative freedom as possible. We closely examined fantasy and realistic villages from the medieval and renaissance eras to scope out which parts we wanted to model for the pack.
The grass was handled via single textures that followed 64×64 or 128×128 resolutions. I created a few planes with subtle edge loops and exported them to Unity, wherein I applied a vegetation shader as well as the texture to that plane. The vegetation shader controlled all the aspects of the grass. Well, that’s what we did back then at least (I believe Unity now lets artists simply select the texture and slot it in as usable grass).
We sort of messed up with the trees in our first pack. Those trees consisted of two models—a trunk and a canopy—and we would apply a single material to both models and then set up a vegetation shader on the canopy model. We followed this two-model procedure because (we didn’t know at the time!) that one could isolate parts of a model’s polygons and set them to a different material ID. The problem with having two models instead of one is that they didn’t play well with the Unity terrain editor.
Nowadays, we integrate animations and LOD models into our trees, which can be quite daunting and complicated. However, the end result offers an immersive depth of realism and quality. I’ll model, unwrap, rig, skin and animate a tree. Then I’ll duplicate the tree and save out the weight data of the rig. After that, I’ll break down the entire rig from the model. I’ll go back into the modeling phase and carefully remove edge loop detail, lowering the triangle count and ultimately creating the first LOD model.
I’m cautious while removing triangles, otherwise the animation will show an immersion-breaking pop. After the new LOD model is created, I’ll reset the rig and import the weight data, ultimately using the same rig as the prior tree; the animation is confined to the rig, so it copies over without a problem. I’ll usually do this for at least one other LOD model, too.
Upon exporting our assets out of 3ds Max, we make sure pivots are centered—each model is centered to 0x0x0 on the grid, and both scale and axis orientation are all correct. When these requirements are met, we’ll then export our assets to Unity and set them up so they’re game ready.
Working at Epic Games for awhile, I learned to create a demonstration level that shows off key selling points through level design, as well as a grid level that shows the viewer exactly what he or she is buying on an organized-grid scene. We then create animation splines that showcase both levels in a video format, as well as create selling-point screenshots for Unity’s submission gallery. Having an understanding of lighting in game engines is a really important skill to master as both a 3D artist as well as a level designer. Lighting helps further the art direction and the feeling you’re trying to invoke from the audience.
We actually incorporate additional assets instead of usually cutting them out, and we always account for this during the pre-production process when judging the scope of the entire project. When we’ve hit our scope, we’ll generally take an objective look at what we’ve got and see if we need to add anything. The village set we created was a prime example of this as we had planned out the buildings and everything within it. However, once it was all modeled and textured, it simply looked vapid and monotonous. We then decided to add more texture variations to the rooms and it really breathed life into the entire pack. With each client, we give 110% effort, always trying to go above and beyond their expectations. Working on these art packs and with clients isn’t just a job—it’s a passion!