This is more of a general question than anything concerning the two specific libraries, however I am extremely confused on this topic. I began setting up a Kinematic Character Controller in Bullet for my game yesterday, and I am using Ogre for my Rendering Engine.

My character mesh (in Ogre) is animated, but I am now confused on how exactly I can get an accurate physics representation of it into Bullet. For a simple game, I understand you can just use a primitive shape to encompass the general size of the character, but for something like an FPS game, having an accurate collision mesh is necessary.

So my confusion is how exactly would I use a collision mesh with an animated mesh? Just taking a shot in the dark, I was thinking perhaps you give the collision mesh bones and then merge them with the relative bones on the visible mesh to make sure it is accurate.

Anyways, I am just extremely confused as to how a physics collision model is kept accurate during gameplay when the character is animating.