That thread has all the information we have so far - these are NANs (Not A Number) pixels, which means there is an error somewhere with Corona, which we are investigating but haven't pinned down yet. It's an unusual situation as we've not seen NANs that only appear after a certain number of passes before (in either C4D, or Max), that I recall anyway - NANs are usually there from pass 1. So that makes it a complex issue to investigate - we are on the case, though!