I have objects in my scene, which are combined at runtime. My environments are procedural rooms with four walls. How I combine them is all four walls at once, grouped by material, into a single mesh (I’m not sure this explicitly matters for my case but I’ll bring it up later). There is no directional light.
One of these sections can count as a dozen+ of shadow casters with a point light in the center. I believe this happens due to cubemap stuff, where one point light actually counts as six lights, but there’s very little online on how this works, hidden amidst a sea of similar but irrelevant information. Plus, it’s still just one object and one light, how can it be more than six? (again, no directional light)
Related to this, could I expect better performance by grouping meshes by singular wall (so I have four flat meshes) rather than one larger more complex mesh, since one side of the point light will be hitting one or two sidea which technically contains the data for the other sides?