A while back I had a control sceneController that went through a game level and added physics node using simple shapes based on the name of each geom in the model; after a while I when added I used MeshCollisionShape for any object that had a more complex shape or geom wasn’t named. The result was 20 or more physics objects that seemed to work fine. I wanted to ask on a larger level would this control still be favorable or is it better I just use MeshCollisionShape on a whole level section. By level section i mean a section of a larger level broken up into squares and placed next to each other to hopefully allow the engine to ignore certain sections.
It does make sense to have a large scene separated into multiple physics objects and collision shapes to offload some computations to the broadphase, yes. If thats what you mean.