OK, here's another related question. Would it be feasible, or even make sense, to have a python script attempt to transfer the effect of a linear, joint-rotation-driven morph (JCM) into an existing weight map? At least in the sense that each JCM delta defines a vector, and a vector can be derived from the sum of other vectors, one of which could be a vector in the direction of a particular weight map's effect for the vertex that delta applies to. If the combination of relevant weight map vectors could completely account for the effect of that JCM's deltas, then the JCM becomes redundant.

The difficulties I envision would include changing the conversion process for JCMs that apply after joint bending, as opposed to before. And whether the JCM was designed to apply at a specific joint angle (or range of angles), or applied only when a specific set of figure shaping morphs were engaged.

I have seen, from my own attempts to adjust weight mapping on significantly morphed figures, that weight maps relevant to the base shape can become, not so much unstable, as inappropriate if the figure shape moves vertices significantly away from or towards the joint's origin at extreme bends. At which point, one might, say, want to spawn off a different figure and customise its weight maps specifically for that set of morphs or even bake those morphs into a new mesh.

This is in no way intended as a criticism of figure creators painting weight maps, whom I remain in awe of. I'm merely interested in whether use can be made of tools not already in the figure creators' tool boxes. Perhaps what I'm suggesting would make more sense integrated into the Morph Tool - a convert JCM to weight map function.

I can even imagine a mechanism to drive weight maps non-linearly, since Poser has the luxury of supporting multiple transform channels of a given type, where aside from the main rotation channel driving the existing weight maps, two additional rotation channels of the same type could exist. One of which derives its value from non-linear input valueOperations and drives additional weight maps, while the second additional rotation drives no weight maps, but exactly cancels out the rotation of the first additional rotation. I think that could completely eliminate any need for ghost bones at all.

I know that multiple rotations can work in this way (not that I've used them with weight maps, as such) since I added them to the thigh actors of a loose, conformed shirt, so they wouldn't follow the base figure's thighs unless desired.

---

My ShareCG Stuff

Verbosity: Profusely promulgating Graham's number epics of complete and utter verbiage by the metric monkey barrel.