I know theres occlusion culling and LOD managers that swap out objects at a distance but I had another idea.
So I have a terrain thats about 1500 polys, and i made 2 more that are around 700 and 150. I was planning on making 3 cameras combined and each one has a range it can see. (See the picture)
So the first camera renders the most polys and the last camera only renders maybe 50 polys at a time.
Would this method work well and any suggestions?
That sounds pretty badass. I’d love to see some screens of it in action once you get it working. 
I fail to see how this is different than UniLOD or our LOD meshfilter script. If you are talking specifically about the terrain; the built-in Unity terrain already reduces the amount of detail in the distance, you can see that by switching on wireframe.
Well for an iphone game theres no built in terrain, I made one (or 3) in blender.
Also for characters i would only have to first camera see the animation and the second camera sees a posed character and the far camera probably wont see anything but the terrain. This will be good for big groups of enemies.
If the terrain mesh is just 1500 polygons, is it worth it? With this approach you’ll push 2350 polygons in three draw calls instead, having to clear the depth buffer twice (different near-far clipping per camera). Cutting up the terrain mesh and using occlusion culling instead might perform better, depending on the game.
Also, you’re going to get seams at the LOD boundaries. Without modifying the meshes on the fly, you can’t get rid of that. Unless perhaps you have a regular grid-based terrain, constrained by the lowest LOD.
Thanks Jasper! that was the kind of help i wanted.
I was trying to think of a way to get good graphics and performance because my game will have a lot of moving character and physics.
