Methods for Approximating Loop Subdivision Using Tessellation Enabled GPUs (original) (raw)

Subdivision surfaces provide a powerful alternative to polygonal rendering. The availability of tessellation supported hardware presents an opportunity to develop algorithms that can render subdivision surfaces in realtime. We discuss the performance of approximating Loop Subdivision surfaces using tessellation-enabled GPUs in terms of speed and quality of rendering for these methods as well as the implementation strategy. We also propose a novel one pass unified rendering setup for all three methods. Subdivision using the Loop method supports arbitrary triangle meshes and provides for easy transition from polygonal rendering of triangles to the parametric domain. Majority of graphics software applications, especially game engines, render polygons as triangles. The objectives of this paper are to evaluate the performance of smooth rendering algorithms developed to take advantage of tessellator enabled GPUs, provide an easy transition from polygonal to parametric rendering and propose an optimal way to achieve multi-level rendering dependent on performance and visual needs of the application.