Hey guys, I'm a little brain dead, and I'm trying to figure out how to correctly create the bind pose for a model.
I've tried a couple of different permutations of multiplying matrices and I'm looking for some support.
So my general understanding is that I read in the nodes with the "JOINT" type. Then I store their matrix data as their local transform. Then I convert the joint's transform to world space by concatenating the matrices starting from the parent node and walking down the joint list.
Is this the bind pose?
If I multiply this matrix with the joint's InvBindMatrix will that work? Or do I need to walk the joint tree again and concatenate the InvBindMatrices as well?
Where should I apply the bind matrix shape?
All I want to do right now is draw the skeleton for a model. I'm not looking to apply skinning yet.
I know this is kind of vague, but I'm just looking for some feedback.