The girl's dress and the italian cooks pinafore (?) cps are assigned to several vertical bones (all located at pelvis height) which then have dynamic constraints to swing around, pretending to be cloth. That's all.
No ambiance occlusion at all. That would have taken too much time in rendering. Lighting was done with three lights - one main and one rim to give definition. To emulate ambiance occlusion, or global illunination, or whatever you'd like to call it, there's the third light as part of a spinning-light-rig, which travels up a spiral path surrounding the whole scene within one frame (at 36 passes per frame).
Using 36 pass rendering had this other purpose, too:
The fuzzy surface / look of the characters was achieved by using an animated displacement map for their 'skin'. That means that 36 times a frame, another displacement texture is rendered. A:M then blends these passes together, so the fuzzy, yet smooth surface develops.
Elmar Keweloh *** http://www.soulcage-department.de *** email@example.com