%@ include file="../../check-login.jsp" %>
Re: pccm knot insertion questionIn Reply to: pccm knot insertion question posted by soren on December 14, 2004 at 21:41:08: : I am stuck. I am attempting to implement "Patching Catmull Clark Meshes" from siggraph 2000 by Jorg Peters. I dont understand the knot insertion step, which I am told is really easy. : I need to insert three new knots at the endpoints to interpolate my nurbs patches. fine. : I need to insert three new control points to preserve the relationship beween degree,ctrl, and knots. fine. : what I dont understand are the rules for the new control points, : Q0 = (P-1 + 4P0 + P1)/6 : how can I use these in a control array that already contains P1, P2 ??? Isnt that a problem? : I understand that when the spline function is evaluated, these guys will match up with P0, P0,P0, P0, but I feel like I am setting something up wrong, or miusunderstanding something. I haven't read the paper recently. However, the rules that you specify look like the cubic B-spline subdivision limit mask (for the first point) with some end-point rule. Are these rules used on the boundary of the patch, perhaps?
|