I think the uv points <0,0>, <1,0>, <0,1>, <1,1> will correspond to the
texture's "corners" (easier to think of in terms of an image_map) and
each patch should have a subset of that. eg:

say you have 4 patches

A--B--C
|  |  |
D--E--F
|  |  |
G--H--I

where A,B,C, etc represent the corners of the patches
And say you want the image map uniformly mapped onto the four patches,
then for patch GHDE you would need <0,0>,<0.5,0>,<0,0.5>,<0.5,0.5>
similarly HIEF <0.5,0>,<1.0,0>,<0.5,0.5>,<1.0,0.5>
EFBC <0.5,0.5>,<1.0,0.5>,<0.5,1.0>,<1.0,1.0>
DEAB is left as an exercise for the reader :-)
as long as the uv vectors are the same at the boundaries of the patch, it
should be okay (this is all theory, not tested ;)

Do get Mike's patch from the binaries.programming group, it has support
for uv_mapping type 2 bicubic patch, which is easier to use:
type 2 accuracy 0.01
instead of
type 1 flatness 0.01 u_steps 4 v_steps 4
the type 2 patches use less memory and don't suffer from little black
dots on the surface. The patch has a text file explaining the uv_vectors
usage as well.

Post a reply to this message