From research over the past couple of day s I have identified 2 methods for handling Level of Detail using the Geometry shader provided by Direct3D 10.
The first method is vertex clustering which can be implemented quite efficently and contrary to my previous knowledge on the method is capable of producing meshes of a reasonable quality if run in an incremental fashion. This method would also be applicable to implementation on the CPU.
C DeCoro, N Tatarchuk. Implementing Real-Time Mesh Simplification Using the GPU. ShaderX6.
The second method is to use triangular patchs and control vertex density as required through tessalation, detail can be added with height or displacement maps in the vertex shader.
T Boubekeur, C Schlick. Generic Adaptive Mesh Refinement. GPU Gems 3
GPU Gems 3 http://http.developer.nvidia.com/GPUGems3/gpugems3_ch05.html
I still need to check recent Siggraph proceedings so there may be more to uncover, hopefully I will get time to look into them tonight or tomorrow.
Subscribe to:
Post Comments (Atom)
No comments:
Post a Comment