Reverse LOD's

I thought I could put in a larger scale variant of a model for LOD1, and smaller for LOD0, for reverse LODing. This is done in Civilization and other strategy games. Zooming away from the map, the units, (and sometimes other important map features), get BIGGER, so the player can still see them well.

It does not appear to be working, likely because the LOD system depends on percentage of screen space to transfer to lower detail mesh renderer (up in LOD number). If the percentage is higher, the LOD wants to go back to the higher detail LOD (lower in LOD number).

So I think using Unity's LOD system is out of the question. I was thinking of scaling the skeleton based on relative camera position, But... that opens a whole different set of problems.

Any ideas on this? I seem to remember reading somewhere there is a different way to handle LOD, that doesn't use percentage of screen space, but just plain camera position.

 

You could build your own. Just have a component that references a list of renderers and their respective distances they swap from. Then every frame do a distance check to the current camera to see which renderers need to be enabled, and which need to be disabled.