andrel wrote:

> The way I do it is as follows:
> For both representations I compute the values at the 16
> points (0,0),(1/3,0),(2/3,0),(1,0),(0,1/3),(1/3,1/3),...(1,1)
> as a function of the input parameters. This wil give a 16 by
> 16 matrix.

16 x 16 ?

> Then I compute the inverses of these matrices.
> conversion from bezier to hermite is then by first applying
> the bezier matrix followed by the inverse hermite matrix.
> You can do this in one step of course.

I'll try that.

> Hermite patches use the value at the 4 points, plus the
> two first order derivatives in the u and v direction,
> plus the second order derivative in the uv direction.
> So we have for each corner i: A_i, dA_i/du, dA_i/dv and
> d^2A_i/dudv, if this notation makes any sense to you.

I've read about the twist vectors being the mixed partial derivatives, I just 
can't imagine what that means :-)
In the bezier form I can visualize the surface "bending towards" the 
control-points. In the hermite form the first derivatives are the tangents, but 
is there a geometric interpretation of the twist vectors?

Thanks a lot!

Post a reply to this message