Some principles of CG Supervision

In leading 3D teams for animation, video games and product visualisaton I've come to realise that there are some principles on which a successful CG project stands. These principles ensure that the project - usually a complicated mix of 3D technology, creative vision, demanding clients and last-minute changes - will be undertaken and delivered as smoothly as possible.

The principles are work ethics that - as a CG Supervisor, Lead Artist or Art Director - you should implant onto your CG team. Some of your artists will already be wired to work this way; for them these principles will be second nature. But for one or two artists these principles will be utterly alien. These artists will be secretive individuals. You've probably met their type already. They are guarded about their work, defensive about its quality, unwilling to share until "it's finished", perfectionists in the wrong areas and at the wrong times, and hold high opinions about their abilities.

As a CG Supervisor your project management focus will be dominated by these individuals. Their approach to work is a risk to your project. The following principles are designed to mitigate their effect on a project:

  1. Breadth is better than depth
  2. Share everything
  3. Get work "in engine" or "down pipeline" as soon as possible
  4. Everyone's a trainer

1. Breadth is better than depth 

It's always better to work in breadth rather than depth. What does this mean? Imagine that Alice and Bob - two artists - are painting the Mona Lisa.

Alice works in breadth. With each day her painting improves as a whole. She starts by sketching the image, then colours it, then refines the whole painting over a few iterations. Alice takes 12 days to finish her painting.

Bob, on the other hand, works in depth. Each day he takes a section of the painting and completes it to a high standard before moving onto the next section. Bob takes 12 days to finish his scene.

Alice and Bob both take the same amount of time to finish their scene and both ultimately produce the same outcome, but Alice's approach has far less risk than Bob's. She is also more able to adapt, change and tweak her image than Bob. Also, if push came to shove, she'd be able to publish her painting at any time and it would arguably be in a better state than Bob's. Alice's approach is better for contingency and, ultimately, better appreciated by everyone downstream in the project.

© Leonardo DaVinci, all rights reserved. Visit his DeviantArt page.

© Leonardo DaVinci, all rights reserved. Visit his DeviantArt page.

Animators understand this approach already. They've been drilled to lay out, block then finalise whole scenes. With each phase of a project the assets should - as far as possible - iterate as a whole.

2. Share everything

Do not allow your team members to work in secret. The creative process benefits from artists sharing their work and ideas with others. Encourage them to print screenshots, give demos or show-n-tells and attend dailies. For CG projects that have an added technical component then the sharing of work is utterly vital for technical problem-solving.

Efficiencies, ideas, complaints, advice and innovation all come from people who have seen the work of others. It is crucial that every approved sketch, concept piece, model, material or animation is shown to others.

The tricky part is insulating your team from what could be a lot of noisy and irrelevant detail. It's actually quite dangerous for unapproved or rejected assets to influence the downstream thinking in a CG pipeline. Your challenge, then, is to devise a sharing mechanism where people can understand which ideas have been approved and which have not. 

3. Get work "in engine" or "down pipeline" as soon as possible

All aspects of CG work involve a huge and immobile piece of technology. For animation it's usually the render farm; for video games it's the game engine. In any case I'm saying that the problems associated with a 3D or 2D asset will not become apparent until that assets is pushed down the pipeline or put in-engine as soon as possible.

Implore your team to push assets down the pipeline or into the game engine immediately and frequently. Give them tools to do that. Ask them to export and test assets as often as possible. Then you'll know that, when crunch time appears, you have assets that will glide easily through the rest of the pipeline.

In practical terms this means that your tools programmer or technical director should create a one-button export tool that simplifies this process as much as possible because artists hate manual exporting tasks. Many artists can't stand the accurate renaming, filing, checking and supervision which a laborious exporting process demands.

4. Everyone's a trainer

Some projects unfold at breakneck speed. In the chaos it's important for the experts pass on their experience to the juniors. Your seasoned artists will have tricks and tools that they should share for the long-term benefit of the team. In my experience it's best to avoid having only one expert in your team. You see that one guy who knows how to export rigged car models? Well, what happens if he's ill? Avoid specialisms as far as possible. If you must develop specialists in your team then make time for training and de-brief when the project's finished. It's always better to have two specialists. So, keep a careful lookout for the person who wants to guard their expertise and maintain an empire of knowledge. Ultimately they'll trip you up.

There are more of these

And I'll get around to writing them up soon.

Comments, arguments and advice welcome below.