Yeah I am sort of talking about that screenshot, only the thing is, I opted to just make 4ptSurfaces so I don't have a bump so much as a gab between the extra vertex in a further subdivided region and its less resoled neighbour's edge. Maybe it's the same thing. Since we're on the subject, and I realise this might be elementary, but why would one wnat two triangular nurbs surfaces vs the one with 4 vertices? and in this scenario, would a mesh be more appropriate?
I'm looking to script the solution. I imagine that instead of evaluating the UV of the underlying surface for every recursive step of the procedure that I somehow evaluate these points as divisions along these neighbouring (or least resolved) subsurfaces...I might have just reiterated what you suggested. thanks for the advice On Apr 21, 8:07 pm, taz <[email protected]> wrote: > oompa, > > Are you talking about the gaps visible from the blog screenshot? > > http://culagovski.net/wp-content/uploads/2008/10/screenhunter_06-oct-... > > My first instinct would be that you would need to get the surface > border curves, cull the vertex in question (causing the bump), and > rebuild the border as 3 lines. That would eliminate the gap for > adjacent subdivisions of the same size. > > Then you could test the midpoint/quarter points/eighth points of each > line with <Closest Point> to cinch up the gap of any smaller adjacent > subdivisions within a given tolerance. That would take care of the > gaps between between scalar subdivisions (maybe...) to a specified > subdivision level. Or you could probably script this to make it truly > parametric. > > -taz > > On Apr 21, 6:08 pm, oompa_l <[email protected]> wrote: > > > I was missing an "End Sub", now it works... > > > Anyone have any bright ideas on a good way to force continuity between > > unequally subdivided faces? There are gaps at the junctions because a > > greater subdivided area has evaluated more points...Anyways, it would > > be great if somehow those edge conditions were forced to meet up with > > their neighbours. > > > any ideas are greatly welcomed! > > thanks
