The games we make at Preloaded are varied, each of which has unique style championed by the Lead Artist. The beginning stage of a project is devoted to the concepting and fleshing out of the overall look of the game. Once this has been achieved, it is my job as a Games Technical Artist to make this vision real.

Initial direction

I usually keep an eye on concept art development from a distance. I’m always sizing up the ideas to see if they’re achievable, but it’s important at this stage to give space to define something unique without fear of limitations.

Once the style has been approved, I sit down with the lead artist, looking through the concept art and discussing the vision of the game. The final art isn’t always meant to be an exact replica of the concept art, so it’s very important to receive the vision verbally. The lead artists here always have an extremely strong intention for the look of the game, so it’s really important I get in tune with that from the outset.

Choosing an approach

Each game we make here has its own technical requirements and limitations. Being able to create the style we want means adapting how we make our assets, so they work within the confines of the platform and game. For each game different questions arise as to how to make the assets specific to that game.

2D or 3D?

Very often the initial question is whether to make the assets in 2D or 3D.

The game concept and initial art direction usually makes the choice between 2D or 3D clear, but sometimes it isn’t obvious. Perhaps the style should look 2D but have a moving camera – so the assets will actually need to be 3D with a 2D rendered look. Or sometimes the style wants to look 3D but the game and platform won’t be able to handle it. This is a challenge I faced when working on TyrAnt; an RTS (Real Time Strategy) game where you could end up with any number of ants on the screen.

Early concept work for TyrAnt

Although we wanted the ants to have a 3D rendered look, they couldn’t be animated 3D models because things would get crazy pretty quickly. Each character would be a ‘draw call’, so even if we kept the polygons really low the game would have become unplayable as more ants entered the screen. We needed animation, and opting for a 2D ant would have felt out of place in the light baked 3D terrain. To solve this I went for a hybrid approach where I made the ants in 3D and rendered them into sprite sheets viewed from multiple rotations. Since the game was generally played with a fixed camera we could get the feel we wanted without the overhead of actual 3D.

Walking Ant Sprite Sheet, viewed from multiple angles

I wrote a script to render each of the ants out at each of the angles, which definitely sped things up, especially since the number of angles was something that we needed to trial and error.

Ant Swarms in TyrAnt

Making room for experimentation

The right choice on how to make the asset isn’t always obvious so it’s always good to make time for experimentation. Sometimes the art style the Lead Artist has in mind leaves open different possibilities of interpretation. The game Immuno-Defense (TBA) gave me an opportunity to explore different treatments for the environment. It needed to have a “faceted” looking style, with a clearly defined colour palette, but the exact treatment was left open.

Early concept for Immuno-Defense

I started to experiment with a few different methods. Firstly I tried baking lighting textures to a grayscale image, and use this to blend between colours in photoshop. Although this produced the faceted look without requiring real time lighting, I found the range of colours I was getting to feel a bit contrived. I decided to render the scene with subsurface scattering with hazy lighting and then bake it onto the models.

Pre-rendered environment for Immuno-Defense

For a while this appeared in the game, and gave the game a nice misty look. Unfortunately, as the game developed, the levels got larger, and the amount of textures in the scene started take its toll on memory consumption. After discussions with the Lead Artist we came to the conclusion that a sharp clean look would suit the game better and would allow us to remove the need for textures. Taking this forward I developed a script to convert the textures to vertex colors, retaining some of the fidelity and range of colours I had before, but bringing it more in line with the crispness of the characters and GUI. It was a happy solution all round – more optimised and with a cleaner style.

Final look of Immuno-defense

Especially when working on mobile, optimisation is fundamental. It can definitely dictate how I go about making my assets, but I don’t believe it should rule it. Achieving this balance is a big part of my thought process. I’m always looking to push the capabilities of what we can do, but when you’re working within limitations you can’t always let loose.

Where you can let loose

When you optimise and reduce in all the areas that won’t be noticed, that allows you to “let loose” in areas that will. In Immuno Defense as we ended up with all vertex color and no textures, I could let loose with the amount of polygons. (Well, within reason).

Normally detailed shaders are a no-no on mobile, but in the puzzle game Crafty Cut I was able to “let loose” with the complexity of the shaders. This is because all we had was a single object rotating in the middle of the screen. Since the object was low in geometry, I could pull out the bells and whistles on the shaders – something that would be inconceivable on most other types of mobile game.

Green Gem Shader in Crafty Cut

Having a good understanding of the key mechanic/ focus of the game can often steer where I put most of my efforts. With Crafty Cut the focus was on having a tactile dazzling centerpiece. For the turn-based Story Cards (TBA) the need was for all the interactions between the cards to have great feedback, to really reward the player when the cards make their attacks. This led us to make the game completely 2D, so I could put all my efforts into the visual FX. It also freed up the opportunity for me to explore different novel types of FX, as we’re always looking to push the boundaries of what we can do.

In my next post, I’ll take you through our Maya to Unity pipeline; how we try to make the whole process much smoother so we can concentrate solely on the art.

Jamie Niman

Jamie is a Technical Artist at PRELOADED.