Catmull-Clark subdivision

[nobond]

another questoin about cc sudivison, is it my understanding right this is could only handle quaderlaterial polygon rather than triangle mesh .or. more edges in one face(i mean directly?_

 

[rpg.314]

I think for triangle based subdivision surfaces, you need to use loop subdivision surfaces.

 

[Simon F]

another questoin about cc sudivison, is it my understanding right this is could only handle quaderlaterial polygon rather than triangle mesh .or. more edges in one face(i mean directly?_

You can use triangles since after one* subdivision you'd have a quad mesh. The only question is how many "extraordinary" points you have (i.e. those with other than 4 connecting edges) and whether that is a problem for you. It shouldn't be because you can always directly compute where an original point would end up after an infinite number of subdivisions.


(**or is it two - I haven't got time to check)