I wanna know if there''s another way to decide which Object (of billions of objects ) is seen on screen.
First way I mention is to have them in a list an compare the objects coordinates to the actual cam lookat. This may be very slow in a bigger world.
hope you can tell me a faster one.
cya && thx
You'll have to use a combination of algorithms to perform fast visibility determination. Visibility isn't an easy problem on 3D-graphics, so you'll have to study it to find the best solution for your app. But here's a few things that might help:
You can use space-subdivision. Octrees, quadtrees and bsp-trees are commonly used for this.
After that, do your regular frustrum-culling for the remaining 3d-objects.
At this point you can do occlusion culling and remove objects that are left totally behind other objects. There are a few algorithms to do this.
If there still are visible objects, do contribution culling for them. This means that you'll check how many pixels the 3d-object will be on the screen. If it only takes 1 or 2 pixels it can be removed since it contributes very little to the final image.