Robert Bateman
07-16-2006, 12:49 AM
I'm not sure that would even work. Going from bones to blend shapes is obviously straightforward, but going the other way seems a nightmare. I think really you should answer this question before starting,
how do you intend to map morph target data (essentially random offsets from the base vertex positions), into rotational and translational data for joints, the sum effect of which will be uniformly applied via it's skin weights?
for example, consider the perfect case, all vertex offsets are in one direction (i.e. 10x10 poly plane, the target only moves the verts up in y). In that case, you could simply assign skin weights to a single bone (where each vertices weight was the length of it's target offset vector). However, move a single vertex in another axis, and then it cannot be mapped to that bone, and instead would require a 2nd bone to deform the surface. I could imagine it would easily replace each vertex offset with a single skinned bone, and thats just for one target!
I imagine, that the sum cost of trying to force a blend shaped model into a bone rig, would actually be more than simply adding support for blend shapes into the engine, or simply re-doing the facial targets using bones....
CGTalk Moderation
07-16-2006, 12:49 AM
This thread has been automatically closed as it remained inactive for 12 months. If you wish to continue the discussion, please create a new thread in the appropriate forum.
vBulletin v3.0.5, Copyright ©2000-2012, Jelsoft Enterprises Ltd.