So this is a bit of headscratcher for me…
What you are seeing here is the result of using the same sphere mask, based on material world position node excluding material offset, feeding into a alpha mask and an offset.
The offset pulls the pixels which would be clipped outside of the sphere mask.
Now I would expect them to just clip out as they would without the offset, the values they use are correct, the sphere mask is not influenced by the vertex offset.
I’m feeding sphere mask result into emissive
What I am observing however is some kind of difference between “wether the pixel should be rendered or not” and “wether the pixel should write into the depth buffer”
Depth buffer visualisation
As you can see, the pixels which are offset outside of the spheremask range get to write their depth into the buffer, while some of the pixels being pulled into it cannot.
As I say, this is a bit of a headscratcher to me.
I fail to see how writing a white pixel in the back buffer in the right place would be any different from writing depth in the depth buffer in the correct place.
Any help would be appreciated as I seem to be completely stuck