Re: [Algorithms] normal calculation in soft skinning

[Andy R. <an...@ne...>]

Chris Hecker wrote:
> Jon Watte wrote:
> >The fully correct solution is to multiply by the transpose inverse of the
> >skinning matrix.
> 
> Actually, this is not correct.

Actually, it's sort of correct.  What he said is true for the normal
of a PLANE.  The problem, of course, is that the "normal of a vertex"
is a poorly defined thing that is arrived at via some kind of
heuristic averaging of the normals of the associated triangles.  This
averaging operation isn't a matrix, and doesn't commute with the
inverse transpose operation.

I'm sure you know that.  I just thought it would be useful to point
out why the answer was wrong in a way that gelled with the "normals
transform by the inverse transpose" rule (which is a really important
thing to grok, IMHO).

Andy

-- 
Andrew J. Ross                NextBus Information Systems
Senior Software Engineer      Emeryville, CA
an...@ne...              http://www.nextbus.com
(510)420-3126                 Why Wait?