A few years ago I built a 3d voxel engine proof of concept akin to minecraft that generated a large mesh (a chunk) by iterating over 3d voxel data (dirt, grass, etc) and creating a vertex/UV list to build the mesh from. This happened on a seperate thread for speed. With cubes, this process is simple, however I want to make a top down game with more complex 3d tiles. I often wondered if I even needed to combine them into a single mesh or if I could leave them as separate meshes under one gameobject with multiple renderers.
Here’s my main question:
Is it necessary to combine the individual tile meshes (a wall tile, a floor tile, etc) into a single mesh (or series of larger mesh “chunks”)? Why or why not?
I don’t want to prematurely optimize and make my level system more complex than need be. Tiles will be getting changed throughout gameplay so the geometry is not static throughout the game. Advice or examples of how this can be done would also help!