Why are we so obsessed with making game art more and more realistic? Is it really the only path worth pursuing? Why?

I say it all stems from an inferiority complex from our childhood! It originates from games' originating in technologies with very limited capabilities of visualization. For the first several decades, computer and video games had to settle for a cartoony and very unrealistic look, no matter what desires of the artists behind them.

I remember the first major breakthrough in realistic games visuals, which came back in 1986.

Rambo: First Blood 2 published by Ocean, had the most incredible looking loading screen. Even the mainstream was amazed. You'd never seen anything like it! Why, the look practically rivaled that of the original movie poster! You could actually make out the actor's face, muscle striations and what make of weapon he was holding. All of which was accomplished with some 10 different colors used over 320 x 200 pixels.

But if you remembered the original reference and concept, it was quite laughable. And once inside the game, the ”realism” was quickly ruined:

Limited to very few possible colors, the artists had to make do with garish combinations of green, purple, pink, black, and white. The best of the artists of that day strove to conjure up more believable colors by using color theory; placing a color next to another color that would draw out particular qualities from the original, and thus make it seem different from when placed next to a third color. Apparently the only way to tone down the fluorescent green into something somewhat believable in a nature setting was to put it next to the magenta or purple.

That kind of upbringing, with enforced limitations is bound to create a complex of sorts. We, as games artists, generally strive to reach as high a degree of realism as possible, before even contemplating any other artistic styles. Likewise, generally speaking, the game consumer will associate a realistic art style with cutting edge quality, and the stylized (for example cartoony) look with a product, which is either out of date or aimed at minors.

Please tell me, what is cartoony game art really?

I am going to demonstrate to you why the often-used term "cartoony" (in connection with game art) is very limiting, and why we need a more precise framework of terms to describe what we see, play, and make.

To use the term "cartoony" for these graphics is very limiting. This is a huge category where anything non-realistic is lumped into. It consists of several very distinct subcategories, which often have nothing in common, apart from not being realistic in nature.

What is the difference?

Without light there would be no color, no visual forms, no nothing. How light reacts with our world, and how we choose to describe that is the core of it all.

Light describes an object's shape, texture, color and other surface properties such as reflectiveness and translucency.

The difference is in how we use light to describe our world.

Essentially cartoony games graphics is all about what methods we use to describe light's interactions with the game world and its (the light's) properties.

So how do we go about that then? We need some tools...

Enter Scott McCloud's Picture Plane from the excellent book Understanding Comics – recommended by Will Wright himself. According to Scott McCloud ANY visuals can be contained in this triangle.

Here is a triangular space held between 3 vertices: Reality, abstraction and iconic.

So let's try it on for size with a select few games from the cartoony range:

The three vertices are represented with Half Life 2: Lost Coast for near photorealism, Jeff Minter's Unity for almost complete abstraction, and Zork for the ultimate in iconic simplification, which is the written language according to Scott McCloud. However good the picture plane is at giving us a quick and rough estimate of an aesthetic style relative to other styles, it is of limited use when you want to look at what exactly makes a style look the way it does.

Although the cartoony, or non-photorealistic styles take up the majority of the matrix, they still comprise of several distinct categories, which have nothing to do with each other. Furthermore Scott McCloud's Picture Plane does little to explain the role of light and its properties in the game world. So we need to find something else...

In his excellent book Digital Texturing & Painting, Owen Demers tries to lock down artistic output into 6 categories of varying sizes:

Some of the categories like "Stylized" are incredibly roomy, whereas "Hyper-Real" is incredibly narrow. Both hyper-real and graphic are graduations of other categories, respectively realistic and simplified. We this is closer to the practicing artist's point of view, I still want to break down the individual components of a certain look or style, so we can recreate exactly the style we want, easier and more precisely.

Still not satisfied, I went through buckets of visuals, dissecting and describing them as I went along. Patterns and trends emerged, and it seems to me that all the images, from Tetris and Rez to Auto Modellista and Gregory Horror Show, share the same basic visual make up.

Always:

In addition to that, they may or may not have a series of added characteristics.

Optional:

A New Set of Categories

So my findings seem all to fit into one of three categories. I have suggested names for them according to their relationship with realism, which is sort of our ground zero in art.

Enhanced Realism

Let's start with enhanced realism. For this category I found the following to be true:

Here is a prime example of enhanced realism. The proportions of the anatomy, and between the anatomy and the level of details here are well in line with what you would associate with realism. But color wise the work is not realistic. The contrast and the saturation have been enhanced significantly, giving the image its crisp and lively style. That level of saturation in the green skin could never be found in a red-blooded animal, since the light of the underlying blood vessels would cancel out the color of the almost complimentary green color of the skin.

Furthermore, the contrast has been increased as well. In our world where light continuously bounces off reflective surfaces you would never see shadows quite so dark. But apart from those two areas, the texture and proportions are well within what you can expect from realism in terms of proportions and level of detail.

Here is another example of enhanced realism in a much more modest setting, technologically speaking. But still within the limits of the polygon counts and texture resolutions, the proportions are not far from what you will find in nature, and the only major differences between this and realism is again saturated colours and increased contrast – the classic artistic tools of the enhanced realism style.

Simplified Realism

Simplified realism, on the other hand, takes a different approach. Instead of merely furthering the trajectory of realism beyond its bounds, it scales it back. The essence of simplified realism is:

Here is an image from Animal Crossing that illustrates simplified realism fairly well. While it is not the only technique used here, it is dominant. Notice how patterns and colors have been simplified into easily recognizable clichés. Color ranges and gradients are simplified down to very few colors or even one flat shade. The leaves on the trees are shaped approximately like a normal leaf would be in real life, but there are much fewer of them, and their outline is much more simple compared to a real leaf. The grass and the water have been simplified into simple patterns, which together with their placement and color tell us what materials are portrayed here.

Another example, Mojib Ribbon, can seem completely abstract, but at its heart, is essentially another example of simplified realism, albeit at a very extreme level. It uses the artistic tools of simplifying real objects down to their last recognizable denominator, which often is just a simple outline, as well as using color fields with very few tones. And of course. the range between the lows and highs of values gives the images an overall high contrast.

Distorted Realism

Whereas enhanced realism builds on top of realism, and simplified realism scales realism back a few notches, distorted realism takes a different approach.

Psychonauts is a good example of distorted realism. The shape, colors and the textures of the objects are like nothing you would find on similar objects in the real world. This gives the art style a near abstract feel, as well as imbuing the image with a magical and whimsical quality.

The shapes of the character and the environments are elongated beyond anatomical and physical practicality. Essentially they take root in real shapes; The characters don't have 3 elbow joints per arm but the shapes are stretched, shortened, twisted and curled out of proportion.

Although not strictly computer-generated imagery, this scene from Tim Burton's The Nightmare Before Christmas illustrates well another take on distorted realism. The houses have walls, roofs and windows as any normal house would have, but the shapes are out of all proportion. Also notice how details which would be relatively the same size or thickness in real life can exist side by side with one being paper thin and the other incredibly thick and full.

Conclusion...

_____________________________________________________