Does cannon mesh imposter impact performance?

I am having a hard time deciding between the default basic camera collisions (checkCollisions) and the mesh imposter collisions. Does mesh imposter impact performance a lot if there are a lot of meshes?
Any help will be appreciated.

Adding @Cedric

It all depends on your use case. If you only need collision detection for a camera, then default basic collision might be the most easy and doesn’t need any dependency.
If you also have physics enabled with mesh impostor, then you should use it for camera as well.
I know ammojs (and bullet physics) does a broad phase before testing triangles. The broadphase is basically, collision detection with bouding boxes. And then, at a mesh level is uses a spatial partition algorithm. It’s pretty efficient, more than the basic collision detection I think.
But it also uses more memory. I have no strong opinion here. And it might be a bit different with cannon.


Hi @BlackThunder4007 do you have any more questions on this? :slight_smile: