Reading the official format specification:

Although the skeleton property is not needed for computing skinning transforms, it may be used to provide a specific “pivot point” for the skinned geometry.

That makes no sense. How is the root node not needed to compute the skin matrices? One must know the explicit node hierarchy of each node to make sure one is multiplying the right matrices.

How can you compute the final transformations for each individual vertex if you do not have an explicit skeleton root?

CodePudding user response：

Regardless of whether an explicit skeleton root is provided, each joint node is still a member of the scene node hierarchy. Consider:

• Scene
  • Node A
    • Node B
      • Node C
        • Node D (Joint)

Regardless of whether the skeleton root is node A, B, or C, all three nodes contribute to the final transformation of Joint D. The skeleton root may have some semantic or runtime meaning in particular applications (not all applications have this concept), but that information isn't required to compute a joint's matrix transforms.