Horde3D

i am studing horde3D and Ogre.

In ogre, frame(animationtrack) record relative translate of a bone(joint)
first: reset bone to origin postion
then: blend and interpolate the frame
then: bone new Local postion = bone origin local postion * frame postion
then: compute bone world postion, soft or gpu skining

In Horde 3D，i see
first: bone new Local postion = frame postion
then: compute bone world postion, soft or gpu skining

See, both two method can drive skeleton move right. wo can get "horde frame postion = bone origin local postion * Ogre frame postion", as i found in "conventer project"

So why design this not that? for performance, or any other detail?

thank very much!

Hi noetic, not too sure about that but I assume it is done like that to keep the horde scene graph very generic, so joints, lights, emitters, anything with a scene graph/node representation would be dealt with in the same manner, joints wouldn't be a special case here where you need to calculate the bind pose and then apply the delta and then the skinning on top, plus it would be cheaper CPU-side wouldn't it? I'm not familiar with Ogre so I can't comment on this further.

_________________
-Alex
Website