Your culling system should not be directly tied to your renderable-object system (no dependency between them). Usually a 3rd system will use the results of the culling system when preparing renderable objects.

Two options would be to link all instances of that tree to a single cullable (either draw the whole forest or don't), or link each tree to it's own cullable, and have the instancing system adjust it's per-instance data buffers appropriately upon recieving the culling results (i.e. only include data for visible trees and adjust the instance-count), but this highly depends on your instancing system and your circumstances.

Thanks.
Well, I was asking, because I'm using hardware occlusion culling with DX 9.0 queries. I needed to implement a big forest, had the frustum culling of the trees tied with terrain quadtree culling, but for the occlusion culling I was facing a dilema :
In order to make my occlsusion culling to work, I needed to render the trees individually and fetch the result to see which tree acts as occluder and which as occludee, issuing a draw call per tree, but in order to make my instancing actually be helpful, I needed to draw many trees with one draw call. As my forest is actually culled with visible terrain chunks, using the results of a quadtree culling, maybe I should treat my terrain chunks as a whole, render instanced trees per chunk and occlude next chunks with trees, mixing the occlusion and instancing that way.


Can you guys point me to a nice article/sample code about software occlusion culling.

Hodgeman - Why would the artefacts be so bad for a one frame delay? Most objects wouldn't move that much in a single frame, and those that do could maybe be removed from the occlusion system.