5 replies to this topic

[makuto]

I've just been playing around with OpenGL and came to the problem of 3D collision detection. I have implemented SAT and AABB collision detection with some 2D games, but 3D seems to be quite a bit different. I suppose I could use SAT with a Z axis, but in 3D that seems to be very taxing even with simplified collision meshes.

If I should implement it in order to learn something, can you point me in the right direction?

If I should ignore it for now and just get a physics engine, which one would you recommend (Bullet, ODE, etc.)?

[ApochPiQ]

I'd strongly recommend starting with a library. I've used Bullet in the past and it isn't too bad; but it's been ages since I touched any of the other alternatives so others would probably have more up-to-date opinions on them.

[Jarwulf]

Most people seem to eventually end up using physics engines to handle collision detection in 3d games. PhysX is a good option but changes quickly. Also beware of small 3rd party wrappers. The port of PhysX I was using for an Ogre project died when the one man team holding it together decided he had enough and quit.

[makuto]

I'm thinking the Open Dynamics Engine or Bullet would be the best for me. What do you guys think?

[Cornstalks]

I'm thinking the Open Dynamics Engine or Bullet would be the best for me. What do you guys think?

I've never used either (so take this with a huge grain of salt), but most of the stuff I've read online has been in favor of Bullet. I haven't heard of anything wrong with ODE, but the majority I've read about seem to prefer Bullet (haven't read why; googling may be revealing). I will say, though, that if you can implement realtime Voronoi fracturing in your game, I will pay good money just for a test drive...

[makuto]

That stuff is amazing!