Thanks, to clarify, what I meant by consistent is that the results didn’t match when you substitute zoom for mask size. I.e. I assumed a 4 pixel mask should look the same if mask size = 2 or zoom = 4
From your explanation, it still seems this is how it’s supposed to work, but I just tried this out with mask 10 while having transform to shadow mask parameter set to 1.0, the result is different between size and zoom variant. Am I missing something?