One of the biggest challenges for a small team making indie games is to balance the time spent on the various disciplines. Often the best approach is to use techniques that are well within the team's capabilities so that if unexpected problems arise there is time in the schedule to fix them. This also means that an extra level of polish can be put in to the game near the end of production if the fundamentals are in place early.

With that in mind, for my game Trackless, I wanted an art style that was easy and adaptable. What I came up with is a 3D art style that doesn't use any modeling tools, and avoids many tedious aspects of 3D modeling, like UV editing and export processes.

In-browser 3D version viewable HERE

ARTISTIC BENEFIT/WEAKNESS

While the art style that results from this technique has some obvious shortcomings when compared to more conventional art, I think it is actually better than your typical AAA approach in several key ways that are hard to notice.

For example, because the game doesn't use shadows or mesh normals for shading, it's easier to control the overall look of the game. This means that when viewed from a distance, the picture-plane can be easily arranged to look appealing using colored lights without having to worry about shadow-popping or LODs.

Another benefit of using textures with simple high contrast lines is that the game becomes more readable. I decided to author the textures for objects with their ideal viewing range in mind, so that large building textures are best viewed from long distances, and smaller props have enough detail to look good up close. This signals to players what is important in terms of gameplay and guarantees that there are clear, identifiable objects on screen most of the time.

A third consideration is that this art style is clearly not literal. This sort of simplified art direction leaves room for interpretation which can be filled in via in-game voiceovers, or left to the player's imagination. While this may not always be desirable, leaving room for the player to engage and interpret the art allows the player to be a more active participant in the experience, which I think is a better fit for a more cerebral game like Trackless. 

TECHNICAL ISSUES

Since making the video above, I have made some improvements to some of the more obvious technical issues. I made a custom shader that sharpens the lines in the texture. In the image below the sharpened version is on the top, with the original on the bottom:

I also decided to switch from using vertex lighting to a more standard approach. On the left is the new shading, and on the right, there is visible popping and triangle shading artifacts from the old vertex-based solution:

With the original simple shaders, many scenes didn't require any special optimizations to run at a reasonable speed, but since reducing the draw calls is easy I have been taking some extra steps to do that.

The first step is to use THIS script to convert all the sprites to meshes.

Then use THIS script to group the meshes together based on their material. Because it's more convenient, I usually put the textures for an area and all it's props on a single texture sheet. This means whole areas can be reduced to just a few draw calls.

CONCLUSION

I had originally attempted a more conventional art style, but spent most of my time making art assets and not much time was left to design puzzles or code gameplay systems. This would have been fine except that the art failed to get much of a reaction from players. Not only is this new art process faster and easier to produce but it has several advantages over a conventional workflow. More critically, the art has gotten a very positive reaction from the backers of our Fig crowdfunding campaign, which has resulted in us passing our funding goal.