![]() Regardless of that: There are certain errors in implementations of skinning that can be observed occasionally. Regarding the "scaling to 0,0,0" idea, I'd have to check whether this is actually allowed by the spec. (And first having it as a non-normative section would increase the chance that it will undergo a critical review by people who are actually using it as the basis for their implementations). It might be part of the next glTF version, though. Regarding the "non-normative" part: Things that are going into the spec have to undergo a ratification process, so it's not entirely trivial to change the spec "just so". In that case the skeleton property would be omitted. If they have no common ancestor other than the scene, the scene itself is necessarily the common ancestor. If a skeleton property does not have some meaning in the software exporting the model, you might as well omit it.Īlso, what happens if there is no common ancestor and the joints come directly from the scenes array? It indicates a semantic root of the skeleton, which may be helpful to software in which nodes and bones are not mixed in the node hierarchy, to preserve object types. Looking for a transform is probably not a good way to assign the skeleton property. If one uses the skeleton property, it can set it to be the common ancestor with the transform but also the one without the transform. Each joint's world transform - relative to the scene root - is affected by all of the joint's ancestors. The skeleton property (or lack thereof) has no bearing on the transforms of joints. Should we apply that transform to the skeleton? The skeleton has a common ancestor with a transform. How should we interpret the transform for the skeleton itself?. ![]()
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |