how would I do clipping

I am making a game that has a huge game world. first of all I dont want to render the whole world so how should I do that? should I have the world broken off into smaller pieces or something similar? my second question is how can I do clipping, only rendering what the camera sees?

1(simple) You could break down the world into zones, then load everything for that zone when the player gets to the “zone” spot. (Like the FF games)

2(Little harder to code) You could have a LOD system like WOW and GTA that breaks the game down to zones, but with LOD you load things in strips or aka what the user only going to see and unload anything that user not going to see.

All in all your going to need a LOD system with either of em for smaller or more detail objects so not to over load the scene.

As for the camera clipping you can use…


If you mean frustum culling, panda already does that for you.

what do you mean by frustum culling pro-soft?

thank you adr for your post, i was thinking about doing number 2

With frustum culling, only objects that are in the camera’s view will be rendered - that’s what Panda does by default.

Panda supports an even smarter form of occlusion culling: portal culling. That’s if you have a window or door, Panda would not render the objects that are behind the wall, only the objects that might be visible through the door or window.
That’s not enabled by default though - look into the PortalNode in the API reference.

Also see: … ermination

wow, i didnt know that panda did that. thank you for help

That said, it is still probably a good idea to break your world down into sections, particularly if you have anything interacting with it on a regular basis. Just because geometry isn’t being drawn in the render pass doesn’t mean it’s not kicking around in memory or reacting to transformations/collisions/etc.

wow, thanks pro soft. is the portal culling doing a hidden surfarce elimination (determination) or the complete object?

Yeah, portal or frustum culling happen on the rendering side, so if the world is large enough, you still have to deal with loading too much into memory at once, having too many agents or physics objects processing at once, etc. I think most big world games load the scene in chunks, only loading, say, the chunk you are on, and all of the chunks surrounding it. For objects in the distance, I think they have a very low resolution model that they use for the entire scene, and show that for distant scenery.