I’m curious how much baked lighting efficiency affects the way you all model your scenes. For example, let’s say you’re building an “arena” style FPS shooter level, with a bunch of hallways, small rooms, and different levels. If I were making a level like this, I’d probably start with a tool like ProBuilder, and extrude a cube until I had one giant mesh for the whole level. But trying to do baked lighting on a single large mesh not only takes a very long time, but the results would be rather poor, given that a single lightmap would have to cover the entire large object.
Okay, then I’ll start breaking things into smaller pieces…but…how far do I go with that? Do I break rooms up into individual objects, and separate hallways? Even them I’m likely to be told my mesh exceeds the max lightmap atlas size. So then what? Break every individual surface out into a separate game object? That’s good for lighting, but terrible if I decide to change the layout later. But how far do I go? Let’s say my floor has a tiled texture. Maybe every tile should be its own game object?
I’m curious about people’s actual workflows when it comes to building the geometry of your scenes.
Thanks,
-Dan