Re: Euphoria physics engine GC 2006 demo
Posted: Fri Mar 07, 2008 3:17 am
That is a very good finding there. And in this case I think it is only a good simile. I believe Octrees are used to keep track of 3D world data in rendering engines. For those that know, is this how octrees are used in games?:
Using and algorithm the engine chops up the level when compiled into octrees which it will reference for rendering distance when the game is being rendered. Thus a game world is a bunch of octress holding octrees that all hold 3D objects. Now all of these octrees rely on camera position and view distance of the camera to decide what to render within the octree (and its sub octrees if nessicary). Thus the whole level is not rendered at once as octrees are rendered based on camera view and then objects are excluded from render with a seprarte algorthim that looks for intersections of objects in the field of view. Thus the engine only renders the objects that are visible on screen. Even more advanced engines use an algorithm that looks at what polygons of objects are visible and only renders the polygons that are visible within the octrees that it can "see". That advanced algorithm is called occlusion.
Did I get that information correct? or did I confuse it with another aspect?
Using and algorithm the engine chops up the level when compiled into octrees which it will reference for rendering distance when the game is being rendered. Thus a game world is a bunch of octress holding octrees that all hold 3D objects. Now all of these octrees rely on camera position and view distance of the camera to decide what to render within the octree (and its sub octrees if nessicary). Thus the whole level is not rendered at once as octrees are rendered based on camera view and then objects are excluded from render with a seprarte algorthim that looks for intersections of objects in the field of view. Thus the engine only renders the objects that are visible on screen. Even more advanced engines use an algorithm that looks at what polygons of objects are visible and only renders the polygons that are visible within the octrees that it can "see". That advanced algorithm is called occlusion.
Did I get that information correct? or did I confuse it with another aspect?
I think I got it a little mixed up when I saw the reference image posted there on the wikipedia. But it would be a similiar concept to a degree. But on a model and polygon level, and instead of using it for camera rendering, use it with the physics to calculate which group of polygons are effected.