On Wed, Mar 24, 2010 at 11:56 PM, Adam Mechtley <[email protected]>wrote:

> I'm looking for a good algorithm to compute weighted, multi-way (i.e., more
> than 2) quaternion interpolation that won't introduce artifacting (so
> something more robust than averaging components and renormalizing). I'm
> essentially recreating the functionality of an orient constraint for use
> elsewhere.
>
> I've seen a few options out there, but many of them seem a little
> expensive. I figure there are enough math-smart/well-read people here that
> someone may at least have some suggestions of worthwhile papers.
>

We exponentiate a weighted sum of the logarithms of quaternions (after
making sure all the quaternions that we're going to blend are in the same
hemisphere as the unit quaternion --- which sometimes requires defining a
"local" space in which to do the blend). This reduces to Shomake's "Slerp"
blend in the 2 quaternion case.

Approach described in detail in Michael Patrick Johnson's thesis here:

http://alumni.media.mit.edu/~aries/

Marc



thanks!
>
> --
> http://groups.google.com/group/python_inside_maya
>
> To unsubscribe from this group, send email to python_inside_maya+
> unsubscribegooglegroups.com or reply to this email with the words "REMOVE
> ME" as the subject.
>

-- 
http://groups.google.com/group/python_inside_maya

To unsubscribe from this group, send email to 
python_inside_maya+unsubscribegooglegroups.com or reply to this email with the 
words "REMOVE ME" as the subject.

Reply via email to