Robo Ball

This is a little animation using my RoboBall rig! The rig is free to download and use. If you know maya, I hope you'll download it and give it a spin! Okay. Bad pun, but I had a lot of fun making this little guy and I'm really excited to share it.

If you make any videos using this rig, please post them as video responses on youtube. I'd love to see what you do with it!

The rig should be pretty self explanatory, but creators always seem to think that. ;) If you can't read my mind when you open the file, here's a demonstration of the rig's basic functions.

There's some technical insights if you stick around after the jump!

I had been reading about metaNetworks in maya for a while, and its some really interesting stuff. I've also been slowly writing the classes foundation that will become a new auto-rigging tool at some point in the future. But as I was working on this rig, I came to the realization that my typical selection-altering script could be vastly improved with some message links of the sort metaNetworks run on.

With some minor re-writing, the selection helper control objects now intelligently link to the objects they want to help you select. Because the script doesn't have to do any thinking, just grab the objects the control is asking for, it's lightning quick too (and about a quarter the length). And not only that, its renaming proof! Imports, references, and all that good stuff won't break the chain of grabbiness.

This ball originally started with two important scripts running in it. The selection helper which I affectionately have named Grabby, and the script which was doing my calculations to prevent breakage in the ball. In the end I was able to convert the latter into a utility node network. A rather more complicated one than I usually use, but much quicker than what you'd get from a script job or expression. There's a real sense of accomplishment when you can convert an equation into utility nodes, especially when it's on the more intricate side (and bonus cookies when you do it through script so you don't have to try to recreate the 8 identical copies your rig needs by hand).

The other interesting discovery I made while working on this project is control objects which change their appearance intelligently. Parenting multiple shape nodes under one transform node is pretty standard for control objects, but what I'd never noticed before is that shape nodes have their own visibility independent of the transform node. So you can turn on/off different shape nodes to change the physical appearance of the control object.

In this particular case, it was useful to prevent my cap controls from disappearing inside the geometry when they're moved about. It's probably not something I'd use frequently, since it's sort of a frill in most cases, but when you need it, it's quite nice.

In my excitement over rigging techniques, I've sort of glossed over the reason I started this project in the first place: the interesting geometry puzzle it presented.

A fun little piece of swag from LOGIN 2011 turned into a challenge I posed to my modeling friends: can you figure out the correct angles and such in this little contraption that allow it to exist (and perhaps function)? When I remained the only one who hadn't given up, I knew I'd have to rig the sucker as well. In the end I did get it to work, although I think I still messed up on some of the interior angles.

Creating the bone structure for this little guy was an interesting affair. It quickly became apparent that there would need to be lots of intermediary bones simply for the purpose of zeroing out angles and thus pin-pointing the movement of those further down the chain. And I do mean a LOT of intermediary bones. In the end the animation skeleton had 269 bones, while the simplified version to store the final movements only has 45 (and could be further dropped to 21 in a pinch). Overlay on top of that a massive network of constraints and a set-driven animation here and there, and you wind up with a deceptively simple looking result.

But of course, that's much the point in rigging. It looks simple on the surface and can be baked into a simple bone set in the end, no matter how much engineering had to go on under the hood.