(Deity) Environmental Art Pipeline Creation

When determining how the art pipeline should work on the art and design side, I first decided to sit down with my Technical Director (Ryan Chew). My goal was to find a way to make the art assets be usable by a designer within the level editor, as well as having the art be supported by the given technology. We went over how the architecture worked and what was feasible within the given constraints. What we discussed was how the entire game was effectively running off a single grid-based system, where each square space equated to approximately one square on a 3ds Max grid. The technology was created to place repeatable textures within a given square space, as well as being able to place textured 3D models anywhere in 3D space.

Prior to having any other artists on the team, I began creating repeatable 2D textures for the grid plane, as well as 3D “walls”, and other various objects that I could place around the level. While this all worked as expected, there were a few issues with the current system. The first, and perhaps the most obvious, was that when 2 separate wall objects were placed next to each other, you could clearly see the seam between the two. Additionally, because all objects could be freely placed in 3D space, it was incredibly difficult to perfectly line up two walls next to each other. The second issue was that creating seamless transitions of the floor textures was insanely time consuming on the art side, and having the same repeated texture was very nostalgic.

Wall Seam Issue

(Simple example of the “wall seam problem”) [Deity – 03 – 20 – 2011]

In an attempt to solve the floor issue, I decided to create organic looking 3D models that I would then place under the grid where I wanted the player to move. Since the game was an isometric game, this method did work, but the “wall seam” issue translated to the floor models as well. While this was not nearly as noticeable, it still didn’t give the most ideal result that I was looking for.

Floor Seam Issue

(Example of the “floor seam” and use of 3D model floors) [Deity – 04 – 24 – 2011]

Eventually, I studied gameplay videos of Diablo 3 and Torchlight and tried to gauge how they were trying to handle their art pipeline. While I knew we probably couldn’t replicate exactly what they were doing, it certainly improved my perspective on how it was being done, and made me think outside the realm of my current thought process. I managed to also take away the idea of having “dropdown” areas to provide a stronger atmosphere.

(One of the many videos I studied from for the art pipeline) [Diablo III – 10 – 23 – 2010]

I ended up going around and asking various game designers, artists, and programmers around the campus on how they would recommend doing the art pipeline. All designers, including myself, agreed to the idea that all of the objects needed to be modular in order to make changes to levels easily. The main issue was how to actually fix some of the visual problems we were having with seams. When I approached artists about how to fix the seams, they suggested just making the entire level one 3D object that you dropped into the editor. While this would definitely fix the seam issue, it destroyed the modular idea, and I knew that dropping one mega model into the engine would not be kind to the technology and frame rate.

However, while I was watching an artist create a 3D scene for a movie clip, I noticed that they were having issues welding and getting the texture unwrap to not act up. After several attempts at fixing it, the artist decided to hide the problem area by covering it with another object in the scene. This gave me the idea of actually hiding the wall seam issue by place another 3D object over where the two walls meet.

After speaking to some programmers about how to increase the precision of lining up the walls, one idea that came out was the idea of object welding within the engine. While this sounded ideal, it would require way too much time to actually implement. Another idea though, which was much more practical, was to snap certain objects, such as walls, to the grid itself and off-set as needed. This way, everything lines up flush with other similar objects.

After researching and collecting my thoughts about how to handle our own art pipeline, I went back to my Technical Director to discuss my findings. After several discussions, we ended up developing, what we referred to as, the “Connector System”. With this new system, we combined the best of 2D texture, 3D model, and the grid technology to provide the best result possible. The way it worked out was to have three main walls, a 1×1, 1×2, and 1×3 (in terms of square spaces). Additionally, a special 1×1 object, known as the connector, would be placed over the seams of the walls.

Connector System

(Example of how the “Connector System” works within the editor) [Deity – 09 – 17 – 2011]

Within the editor, I  could access the connector system and start by placing a connector at the starting point of where I wanted the walls to begin. I could then click and drag along the grid until the desired end point, and the walls and respective connectors would be generated and placed in between those two points. We repeated the use of this system for the “dropdown areas” of the environment as well. We then used seamless, repeatable 2D tiles, that we place in the grid squares, and use the dropdown models to hide the edges of the plane, making the entire environment feeling solid and one entity. We also added a “decal system” that would allow us to place 2D textures over the floors in order to add variety to any given floor texture.

Ultimately, this template became the foundation for our current environmental art pipeline, and ended up yielding better results:

(Improved Pipeline Results) [Deity – 08- 15 – 2011]


~ by rhickman-design on September 17, 2011.