I’m in the process of recreating a simple arcade style AR game in Unreal and have replaced simple sphere to sphere collisions with actual physics objects. This is all working fine except for one aspect of the original game design. https://www.youtube.com/watch?v=cIJSAHMZFK4
It stuck to the principle that any collision ‘error’ should make the player feel better about their skills. In the original, enemies and the player have very small collision spheres to make it easier for the player to dodge through gaps while the pickups have much larger spheres to make them easier to pick up. I assume this would be easy to achieve if I was to build collisions outside unreal but I note that when I select spherical collision in the mesh tool, I can’t scale or move the sphere to bias the collision size.
Beyond that, in the original game, when the player picks up a powerup and chases the bad guys, the collision spheres expanded to make them easier to hit. As far as I can make out this is also impossible. I’d be happy to do this from C++ but it seems that the collision is baked into the static mesh which in turn, has to be the root object for the physics to work.
I’m thinking the most likely solution would be to have my meshes and invisible collision objects as separate components that get enabled or disabled as the gameplay state changes