Archived

This topic is now archived and is closed to further replies.

Splitting and Tesselating NURBS Surfaces

This topic is 5370 days old which is more than the 365 day threshold we allow for new replies. Please post a new topic.

Recommended Posts

Even though my maths skills aren''t really up to it (just started with differential calculus at school), I''m currently experimenting with using NURBS Surfaces in my renderer. I have a few problems tho. My main problem is how to split a large set of control-points (say 66x66 points) into a number of smaller patches (4x4 points). So far I just took 4x4 controlpoints from my original set, so that all patches share controlpoints with the neighbouring patches on the edges. But how would I set up my knot-vectors ? The large surface currently has non-periodic and uniform knot-vectors (0,0,0,0,1/x,2/x,3/x,...,1,1,1,1), which isn''t too great, but if it makes things easier I can keep it like that. So, well, I actually have no clue if my approach even could work, but if it could work, I still have no real clue on how to set up my knot-vectors for the small patches. The second problem, where any help or links would be great aswell, is how to do non-linear tesselation of my patches. The general idea is to do something like what kronq described in this thread some months/almost a year ago.
quote:
Original post by kronq heres a pic of a fractionally tesselated Bezier patch with different tesselation at each edge. Each separate quad strip is in a shade of red, each triangle fan is in green. I''ll forego explaining it in detail because I think looking at the picture will give you a fair chance of deducing how its done, but here are the main points to keep in mind: - each edge has an odd number of vertices -- minimum 3 - each edge always has a point at 0.5, and all the other points on the edge are symmetrical on either side of the 0.5 point - as LOD increases, new points are added at 0.5 and move outward *----*----*----*----* *----*---***---*----* *---*---*-*-*---*---* - lets call the LODs for each edge u0, u1, v0, v1 -- the LOD for the center portion is (u0+u1)/2 x (v0+v1)/2 - if an edge LOD is different from the center LOD a triangle strip is used to make up the difference
This gives my a rough idea, but as the pic is not online anymore, I have problems figuring out how the tesselated surface should look in the end. If anyone has experience with this or has a link to some paper explaining this or a similar method in greater detail, some advice could really save me some trouble and time.

1. 1
Rutin
37
2. 2
3. 3
4. 4
5. 5

• 12
• 10
• 13
• 104
• 11
• Forum Statistics

• Total Topics
632982
• Total Posts
3009689
• Who's Online (See full list)

There are no registered users currently online

×