If I'm building a good sized First Person Shooter level, and want to have physics objects scattered around it like most modern shooters, what would be the most efficient way of doing this?
I could just make a giant triangle mesh for the level, shove in all the physics and collision objects, and scream in horror as my framerate drops to a geographical timescale :), but I'd rather approach this intelligently, so I've got a bunch of questions. If you can answer one or two of them, please do. Perhaps we should add the answers to the Wiki?
- Does JigLibX already do most of the collision optimisation for me, or do I have a lot to do to make a full-sized level like this feasable?
- Should I split the level itself up into smaller meshes or models? Is there any benefit to this, aside from being able to manually turn off individual parts as they dissappear from sight?
- Should I turn physics, or collision, or both, on and off depending on distance from player/whether or not they can see them? If so, whats the best method to stop them from using processing power? Do I delete them and re-create them if the player approaches them again, or is there a better way?
- Is there a 'low quality' physics setting, which I can apply to items that the player can't see or which do not need perfect physics?
- Does JigLibX automatically turn off/scale down physics on objects that are no longer moving? If not, can I do this manually?
And so on…