Previous | Next --- Slide 34 of 36
Back to Lecture Thumbnails
hlin1

I don't quite understand the purpose of subdivision modeling. Is it just an efficient tool to create 3D models, or does it have any runtime purposes.

pw123

In the industry, what's the most common mesh shape to use? It is triangles or are we just doing triangle meshes for simplicity?

keenan

@hlin1 Both. On the one hand, it's a pretty intuitive paradigm for modeling, where you can put a few control points "just where you need them." On the other hand, subdivision surfaces lead to very high-quality rendering (nice smooth surfaces, good normals, etc.), and are the foundation for things like Pixar's Renderman renderer. Consider if you had to model directly with triangles (for instance). You'd need a ton of triangles to describe a smooth surface, and you'd have to move thousands of vertices around to make changes. Also, when it comes to rendering, you can get as much resolution as you want: no matter how far you zoom in, the surface remains nice and smooth. Not the case with polygon meshes. In another, very different part of the world, subdivision surfaces provide excellent basis functions for numerical simulation, via techniques like isogeometric analysis.