first, it is done with "bones" -- just a system of setting up individual matrices for specific parts of the body(model).
second, you move/tweak the matrices to there next saved position(most commonly called keyframes).
imagine a person walking, every second you "snap shot" his postion(remeber every postion of every major part of his body), until he has taken a full two steps, then save the data.
now pretend you can move/tween his parts between every snap shot, in sequence.
yea, thats basically the idea, it of course is a lot more complicated to acculy do it in code.
hope that helps.
--Philip
|