Lecture 8: Mesh Representations and Geometry Processing (79)
andrewhuang56
I know that this formula was chosen because the creators simply thought it looked good. However, this feels like it's going to be really slow. How do we decide how many iterations we run this for? Is this trial and error?
Staffjamesfobrien
This was one of the early criticisms of this method when it was first published: too much compute, too much memory. Then computers got a lot faster and memory a lot smaller/cheaper. And there was that paper by Jos Stam that I mentioned that had a very elegant fast-evaluation method that he published with reference code. So for 20 years, hardly anyone cared about CC subdivision and then in about 2 years they became everyone's favorite toy.
I know that this formula was chosen because the creators simply thought it looked good. However, this feels like it's going to be really slow. How do we decide how many iterations we run this for? Is this trial and error?
This was one of the early criticisms of this method when it was first published: too much compute, too much memory. Then computers got a lot faster and memory a lot smaller/cheaper. And there was that paper by Jos Stam that I mentioned that had a very elegant fast-evaluation method that he published with reference code. So for 20 years, hardly anyone cared about CC subdivision and then in about 2 years they became everyone's favorite toy.