Does anyone have a code source example of using the Xith provided Octree with the scenegraph? I want to render very large terrains based on geographic data. Previously I was loading all of this as a single large map. Is it possible or would it be wise to render logically segmented terrain (split on latitude/longitude for example) as separate terrain nodes and then place bounding boxes on each segment so that terrain out of sight would not be rendered? If I use the Xith Octree, how does it work with the scenegraph? The documentation didn’t tell me much.
This might be a bit off topic (for Planner1’s requirements) but has Xith3d people checked Openmind’s (jsr version, CVS head) Terrain demo? It uses vertex buffers and handles big terrains fast. This code does not include any octree methods or splitting huge terrain into several nodes, but the ideas for rendering the current terrain area seem pretty nice.
I’m still struggling with this and going through the source code and javadoc to see what Xith offers. At the design level, I’m failing to see how the spatial data structures relate to the scenegraph in determining visibility and I’m being thrown off course by the OcTree and SphereTree. The SpatialObjectInterface and SpatialContainerInterface have no implementing classes. Are these supposed to be filled in by the user or have they been left for some future development by the developers? I am interested in porting my project over to Xith to get away from some of the low-level tasks of 3d development but it seems Xith hides too much. When creating a scenegraph of anything (terrain or otherwise) what can I expect Xith to do for me and what do I need to implement myself with regards to spatial partitioning?