Open Tutorial

Hard-Earned Insights from The AI Sandbox Development

Alex J. Champandard on May 8, 2009

Over the last 8-10 months we've been incrementally building up The AI Sandbox an part of Premium. It's built on top of an open-source stack, but with key parts built from scratch (e.g. the MVC framework) or heavily customized (e.g. animation system, test framework).

We designed and approach most components based on practices used in industry, but also prototyping various ideas we feel have lots of promise. In the process, we've encountered lots of important lessons — some old and some new. In the video below, I go into some of these insights on the following topics:

  • Test Framework — How do you manage and run your unit and functional tests, and your feature samples?

  • Footplant Detection — What approach do you use for automatically extracting footfalls?

  • Animation Mirroring — Which options are available for mirroring, and what are the tradeoffs?

  • Motion Capture — Where can you find good motion capture online and what's the quality of it?

  • Hierarchical Search — What's the catch for hierarchical search in terms of performance and quality?

  • Area Generation — How do you create areas given a lown-level grid or point graph?

Here's the video itself, shown in High Quality. Ideally, you should view it in High Definition to see the text. Press play and let it buffer, then come back and watch it full screen. It's about 10 minutes long.

NOTE: Everything you see here is available as part of Premium. You can sign-up right here, and there are bonuses for signing up this week — whether you're an individual or a studio.

Discussion 0 Comments

If you'd like to add a comment or question on this page, simply log-in to the site. You can create an account from the sign-up page if necessary... It takes less than a minute!