So some engines do better with a 2k or 4k map with a bunch of models crammed in on the same material (or texture), or in other words, a large texture atlas.
However, some engines/games do better with it being broken up into smaller pieces.
From what I have read and understand, UE3 works better with more individual textures instead of texture atlases that refer to several meshes in a scene.
I guess this could be due to the speed of the game, as you're often moving into a new scene, occlusion areas, etc. so there would probably be more draw calls necessary. Also, perhaps it has something to do with the structure of Unreal Engine 3 itself.
So is there any technical information that can further nail down which one is better? Right now I'm of the conclusion that smaller non-atlased textures are better.
However, some engines/games do better with it being broken up into smaller pieces.
From what I have read and understand, UE3 works better with more individual textures instead of texture atlases that refer to several meshes in a scene.
I guess this could be due to the speed of the game, as you're often moving into a new scene, occlusion areas, etc. so there would probably be more draw calls necessary. Also, perhaps it has something to do with the structure of Unreal Engine 3 itself.
So is there any technical information that can further nail down which one is better? Right now I'm of the conclusion that smaller non-atlased textures are better.