Skeleton and Xform transforms - How do they compose, if at all?

Community USD
1

We received a bug report where the user had roughly the following hierarchy. As far as usdview is concerned, it is as-if the 2 Xform transforms are ignored. Our importer however had offset the Mesh according to its hierarchy by T2 and T1 etc.

SkelRoot
    - Skeleton Foo

    - Xform with some transform T1
        - Xform with some transform T2
            - Mesh bound to Foo

I may have missed it but is there an indication in the documentation about how this situation should be handled?

2

I think this is implicitly described by Universal Scene Description: UsdSkel Introduction - it’s the geom bind transform which transforms the mesh’s local space points to the bind state in world space, and then the skinning transforms are applied

3

I’ve been informed by our crowds folk that we don’t rely on Xform positioning/animation in our Skel agents in our pipeline, which is not to say that it may not work, but like Cameron said, there’s definitely not a promise that it should.