[ODE] visibility detemination?

[skjold@cistron.nl]

Oops...

> You could even consider using a large sphere geometry
> that fits exactly around your intended frustrum, since
> it's meant only for fast culling of definitively non-
> visible objects, and sphere geometries are very effecient
> to check collisions against.

... Disregard the above idea: It would only apply if you actually called dCollide on your "frustrum" geom, and I think it's not a good idea at all to take it that far (overkill). In fact, the only shape that's going to matter is the AABB that will be generated for whatever geom you will use for this.

Another idea that might work well as part of a collision-based approach to quick and rough culling, is to use a plane geom (or perhaps even more than one) to quickly eliminate objects that are behind the viewer, or at the far clipping plane of the frustrum for example. If your environment is mostly on surface levels or in buildings (as opposed to deep space), you'd probably even benefit the most from this if you disregard the pitch angle (up/down) of the camera, when determining the plane's orientation.

Again, I'm just thinking out loud. Please feel free to stop me if it gets annoying.
Greets,
Mark