yikes tackling Collada huh? Well that will really open things up per experimentation with skeletons. Thanks for the break down of the abstraction layers, it's helpful. Good luck with the Collada Beast ; )
On May 16, 7:15 pm, Dave <[email protected]> wrote: > Skeleton Animation is definitely very complex. There are many classes > involved that you need to understand to get animation working. > > This is my understanding of the organization of these classes at the moment: > > - Skeleton/SkeletonJoint - inverseBindMatrix for each joint > - SkeletonPose/JointPose - orientation, position of skeleton for a single > 'keyframe' of animation. > - SkeletonAnimationSequence - sequence of SkeletonPose(s) with durations > - SkeletonAnimator (or whichever flavor you choose) - top level animator > class that application creates to drive animation - play() interface here. > - SkeletonAnimationState - mesh-level animation state info > - SkeletonAnimation - geometry-level animation state info > - SkinnedSubGeometry - the SubGeometry subclass that holds jointWeights > and jointIndexes. > > There are few working examples of SkeletonAnimation other than MD5 - and > those have been working sporadically - they may work now, have not checked. > Richard has been working on AWD2, but I do not know if there is any working > data or examples there yet. > > I am working on Collada and banging my head against this right now. > Hopefully I can share working stuff soon on that front. > > -Dave
