Quad Tree

Ok as nobody has answered as yet, something else I need to ask, as I store my bounding coordinates for each patch/node, do I also need to store a pointer to which index this references within my vertex buffer or do I store vertex information within the actual quad tree node ? At the moment I have a huge vertex buffer with all the terrain vertices in, I can draw from here if I know the index to start to draw from.

Sorry to sound stupid.

Steve

Quote:

Original post by steg ... Is this the right way of going about this ? ...

Yes. That is how the basic quadtree works -- each node is split into 4 equal sections, which are also nodes.

Quote:

Original post by steg ... do I also need to store a pointer to which index this references within my vertex buffer or do I store vertex information within the actual quad tree node ? ...

Without describing what you are trying to do and how your algorithm works, it is impossible for anyone to give a meaningful reply.

Not sure if this is helpful or not but here you go. I store only references to patches in my quadtree.

I use one big vertex buffer (depending on the terrain size, but I am currently only targeting smaller terrains (257x257 - 1025x1025) so the one big vertex buffer is fine.

Each patch stores a start vertex. By doing this, I can re-use the index buffer for every patch.

Since the quadtree only stores references to patches, I traverse the quadtree, determine which patches are visible, and tell those patches to render themselves.

Thanks all,

I am wanting to use the quad tree for terrain rendering, I am going to put in geomipmapping firstly and then quad tree after. I was thinking of using one big vertex buffer and several index buffers ( one for each LOD ), then store the start positions in the trees node.

Regards,
Steve