Black Golem

During the last days I was sketching and modeling new vehicles for Nordenfelt. The following sequence is the asset pipeline each new entity has to go through:

1. analysis (aka steal ideas from other games)
2. design (aka change stolen ideas so nobody knows where they come from)
3. modeling
4. texturing
5. animation
6. lightning
7. rendering
8. game integration

Iterative game development tells us to run in circles until everything is great. This means we have to run through the whole development cycle as fast as possible to make an initial prototype. Then we (and maybe our audience) judge if it's OK. If not we go over it again and again until it works. That's the state of the art.

The same scheme should be applied to all parts of the game, especially assets. First make a design, model a draft, render it and integrate it into the game (if your visuals are prerendered 3D models - you get it). Don't add details like textures, lights, particle effects or animations. Judge the asset prototype within the game first. Silhouette, size and style already can be analysed.

Take a look at this effort/quality diagram:

First you'll achieve much with little effort. As you go on the quality gain per work hour will decrease. That's the Pareto principle (80-20 rule) in action. So try to not waste work beyond the 20% mark.

In software development it's common knowledge that early fixed bugs cost less than their cousins creeping out later. Imagine modeling a tank, texturing it and animating the tracks. In 3D it looks great but its shape seen from above simply does not work for the game. You have to adapt it or leave it out at all. Texturing and animation was for the birds because the raw 3D model already would have made this clear.

There is another reason for asset prototyping. What if the design does not fit the game's style? What if you need several attempts to find the right shape? Cool design on paper may not work as rendered sprite. Check this out:

In most cases you won't start with the best design and therefore are on the wrong trip. Don't complete a design with all its details before you judged its draft in the game. Invest the minimum in drafting and bring the asset to the game as soon as possible. When it does not meet expectations you wasted the least amount of time and money. This will make you happy and you'll be willing to cancel not working ideas more easily.

Finally this asset prototyping is the same thing preached for game prototyping but on a lower scale. When something works reuse it where possible.

Cheers,
Thomas