CG Modeling 3: Advanced - An Introduction to Computer Graphics for Artists - page 272

Chapter 15
CG Modeling 3: Advanced

15.1 Subdivision Surfaces

Subdivision surfaces are a hybrid form of geometry that blends polygons with
Bezier patches by using a polygonal control mesh to defi ne the shape of a group of
connected tangent patches (Catmull and Clark 1978 ) (Fig. 15.1 ). The problem sub-
division surfaces were meant to solve was the topological limitations imposed by
NURBS surfaces. Those limitations increased design time by creating topological
design challenges for artists. Subdivision surfaces were not embraced by industry
until some modifi cations were made that made them practical to use, such as in the
Pixar short fi lm Geri's Game (DeRose et al. 1998 ).
Until Geri's Game , characters made for CG fi lms were typically built of
NURBS patches. What Geri's Game proved to the animators at Pixar and else-
where was that subdivision surfaces could deliver a solid mesh with variable
control over tangency without any possibility of tangency breaks in models with
arbitrary topology (DeRose et al. 1998 ). The Catmull-Clark type of subdivision
surface used by Pixar continues to undergo refi nement. Several competing sub-
division surface defi nitions now exist (Patney et al. 2009 ; Müller et al. 2010 ),
though the Catmull-Clark model remains the one most likely to be found in CG
applications.

15.2 Working with Subdivision Surfaces

15.2.1 Box-Modeling

Subdivision surface editing is very much like box-modeling with polygons. An artist
begins with a mesh primitive like a cube, then progressively edits it by adding edge
loops, extruding faces, modifying vertex positions, carving holes, etc. Subdivision