What is the new way to determine occlusion in post processing?

I used to use the technique from the Tom Looman outline project that would do a

SceneTexture:CustomDepth -> Mark(R) -> Subtract (1) -> Ceil -> Clamp(0,1)

Then 1 would be an occluded part of the object and 0 would be visible. This only gives everything as 1 now saying it is all occluded. I have tried a bunch of different combinations with the scenetexture node and other things but haven’t had any luck.

Works like a charm. Thank you :slight_smile: