Horde3D

Hi,

I'm new to this forum. Horde 3D has caught my attention, it looks really nice. However, I'm curious about how scene partitioning works in Horde3D. It doesn't seem to have any mesh type specialized for rendering large static geometry (BSP, Octree, kdtree, etc.) like most other graphics engine do. Is the user supposed to provide it himself ? The features page states that Horde3D has a "Unified scene system where world, models and skeletons are just scene graph branches and no special objects". Does this mean that scene partitioning in Horde3D is supposed to be done by separating the static world geometry in small scene nodes and building a tree of scene nodes with them ?

Thank you very much for your answers.

If you needed super-optimal rendering of some species of structure then you could write up an extension for it.

Most other graphics engines and projects are using the cookie-cutter pieces approach, often with a modernized form of BSP for the "frame." Cookie cutter works well because its batch friendly and elements like columns, pipes, and grates tend to repeat in an environmet. The recent enhancements to the 1.0 beta such as occlusion culling should really help out rendering performance.

If you want to do BSP etc, you'll have to do it inside your application by toggling the visibility of horde nodes manually, or, by extending the horde code internally (by writing an extension) so that this can be done automatically (probably by creating a new horde3d BSP resource type).