This is always a tough question because it’s a balance between a couple things. Modular is definitely better for bug fixing and reuse which is awesome, but suppose you make one floor tile that is 2m in length and you want to make a floor that is 20m. You would have to place down 10 tiles and that would cost you 10 draw calls at the mesh level. So the key is to block out your environment with small modules then analyze the scene to see how you can make bigger modules out of your groups, so if you see you use 20m floors a lot then go to your 3d program and export out a floor tile version that has 10 of those tiles in a row. The other side of that evil sword is that if you make your new big modules too big than they will not cull out since most of the mesh might still be in the view frustum. There is no absolute answer for this one, you just have to be mindful and approach each area with these fundamentals in mind.
1 Like