So, mesh collider calculates it all in one, but box collider calculates faster even with lots of them. So my question is, at what point will a mesh collider calculate faster? I am looking to have a few 10-20k vert mesh colliders in scene, non moving.
When you need precision, or shapes that are hard to render in boxes. Frankly, I rarely use boxes in the environment because I hard get that “perfect” box shape for something that never moves. Crates tends to be blown away.
However, when using Mesh Collider, we ALWAYS make our own highly optimized mesh. Never use the visual mesh for the collision.
Yeah, I planned on using a lower poly, but it cannot be much lower because the terrain itself is fairly low poly, in the sense that it is actually massive to the player. But I’m unsure if having large mesh colliders like that would be a huge performance cost on say a modern or better PC.
Oh, and so my point with the mesh being already optimized, it that, with a reduced mesh for colliding, there might be some very noticable issues with traveling around on it, penetrating through the visible mesh. So not sure if its ok to run it as is, or if breaking up the mesh and having several mesh colliders would help at all, So I can disable some it.
But if a decent PC can run it no problem than, woohoo!
EDIT: sooo, my wording in first post kinda sounds like I wanted to use tons of singular mesh colliders, but I really meant 10-20k VERTS in a few mesh colliders.
Usually, physic engine are good enough to not try to collide with every triangles. If you see performance issue, try splitting your mesh into smaller meshes.
I suppose I’l just wait until then and see how it does. Was hoping to hear about some experiences with large meshes using mesh collider.