[sldev] Body motion and facial expression tracking, Microsoft did it

[Argent Stonecutter]

On 2009-06-05, at 23:35, Tateru Nino wrote:
> Avatars don't occupy a common coordinate-space, however. All we  
> really know about the avatar is that it is drawn based on the  
> position of the dimensionless point representing the agent (and  
> offset up to 10m or so within a volume centered on that point). The  
> avatar itself (AFAIK) exists in its own independent coordinate- 
> space, sans reference to the common coordinate-space in which other  
> agents are positioned, and with even less relation to the coordinate- 
> spaces of other avatars.

But in the viewer the avatar shares a coordinate space with the world.  
Flexible prims attached to the avatar object global forces, not local  
ones. Particles from prims attached to the avatar obey a global  
coordinate space and can target prims attached to other avatars.  
That's the coordinate space this hypothetical "IK tweaking call" would  
operate in.