Hi. I want to change the way I use vertex buffers in my engine. Now, I am creating a lot of VB (one for every objet), and when I go render, I have to change a lot of time between obtects per frame. Now I want to change this and do it like this: I will only use 2 VB: - 1 STATIC VB holding all my MAP vertices, and 1 DYNAMIC index buffer to index the geometry. - 1 DYNAMIC VB created for 4000 vertices to hold all the other meshes (every thing that is not the map). Before render each frame, I will lock and fill up the VB with the vertices of the mesh I am going to render. the FVF for each type of VB would be: // for the static VB (only for the map) struct D3DCUSTOMVERTEX_map { float x,y,z; float u,v; float s,t; }; // without normals and with 2 pairs of texture coordinates for the lightmaps. // For the DYNAMIC VB (for the other geometry) struct D3DCUSTOMVERTEX_obj { float x,y,z; float nx,ny,nz; float u,v; }; So, I''ll change only between 2 VB and 2 FVF per frame (and not per object per frame) What do you think about this?? Thanx.

Hey pleeeeeease, gimme a clue!! :_