> Just fiddling around with methods to approximate isos using triangle meshs
> based on the algorithm described here
> http://astronomy.swin.edu.au/~pbourke/modelling/polytetra/, 27s to render
9m
> 27s to parse using 40 segments along each axis, and 68124 triangles in
> total. I'm bundling it all up into an include file, I would image given
the
> ability to save the mesh so it dosen't have to be reproduced every time
> would make it handy for test renders of scenes that are heavy with isos.
> Mind you, there is only a noticable speed increase when using complex
> functions,
> I rendered the iso f(x,y,z) = cos(2*pi*x) + cos(2*pi*y) + cos(2*pi*z) and
it
> took the same amount of time as the mesh (27s).

If you decrease the amount of triangles, it would render a great deal
faster, with the obvious loss of detail, right? And if you didn't need the
parsing time either.. What about using vertex normals / smooth triangle ?
Maybe that could remove the need for isosurface all together in some
situations?
This is very interesting stuff, keep us posted!

-Peter

Post a reply to this message