I've been searching high and low for a solution to blending / interpolating of multiple weighted quaternions.
There is simply no good solution - even the second order quaternion quadratic is s***.
I'd like to talk to someone who has a clue.
I know about Ken Shoemake's take on the Bezier solution, it's also s***.

My proposal is to solve the system of 4D equations in 5D, based on my observation that any system of linear complementary equations is easiest to solve if we move the whole system up by one dimension, solve it, then transform the answer back down by one dimension.
(That is something I learned from coding physics contact solvers with anisotropic friction - we could solve it much more cheaply in 4D Contact Space than in 3D WorldSpace or any other 3D space).


Please, if you think you understand Hamilton and imaginary number systems, come to my rescue!
Posted on 2012-10-09 02:39:22 by Homer
Dual quaternions?
Posted on 2013-03-26 15:12:00 by Ultrano