Wednesday, September 16, 2009

Kyle Balda Animation Tutorial

Finally found this with sound. Kyle Balda (once of Pixar, teaching summer school at Gobelins).

*little bit later

Had a chance to watch it. Interesting, it reminds me of the splinecast with Doug Sweetland (I think) where he talked also of starting just with the root musically to the dialogue. Makes me think that was the common technique at Pixar as they built there way up through those listerine and lifesaver commercials, when CG was still new. Animation Mentor really pushes the pose to pose style, which is very strong and can give good reliable results (probably easier to teach also), it's interesting to see this different style at obviously an equally high level.

Had me smiling that he was talking bout the same stuff I had just been pondering, namely that the fact that the audience is going to be watching the eyes and face, so get into it as quick as you can so you don't overwork the body.

Also funny when he says "sometimes you need a blink just to keep the character alive" which is a complete contradiction of Shawn Kelly's recent post. (I side with Kelly)

2 comments:

smearframe said...

A lot of Pixar animators seem to have a similar workflow or at least the ones I have seen animate. Its interesting to see almost all of them manipulate the curves by using weighted tangents and keep the keys to a minimal amount as opposed to setting keys to 'shape' the curve and the movement. I am wondering if their proprietary software is more geared towards such an approach/workflow. Just thinking out aloud here :)
I got the 3d world magazines that carry the Kyle Balda tutes. Its the first time someone has put up an article that describes the layering method in such detail. I am a slave to the pose to pose method but I think I am gonna try the animate from the root out method just to shake things up.
Once again thanks for the informative, no-nonsense blog.
- Himanshu

Alonso said...

I think that they can't break their tangents, so they can only scale them wide or small.

But I think that they don't use tangents to do overshoot, they always have a key pinning the curve at highest and lowest points, then they use tangents to get in and out of there. That way if they make changes they don't get surprises.

But I'm just guessing :P

As they say, whichever way works best for you, that's the best way to do it