I’ve always wanted to see stats, no matter how rough, comparing the usage of a convex hull to primitives (primarily boxes) for collision.
I’ve almost always only used primitives, where I usually don’t come across a case where more than a couple of primitives are needed, but I’ve always been curious whether there have been some cases where convex hulls would have been more efficient.
I understand it’s completely per use-case but a rough idea of comparative overhead from testing a few different cases of a convex hull compared to multiple primitive boxes (laid out to get a similar resulting shape) would be nice to see.
Having the number of vertices and FPS difference would provide a rough idea for gauging whether a convex hull would be worth considering over using primitives.
Am I missing a reason as to why I cannot find anyone else curious about this?