Aren't you measuring distortion at the output then? Certainly, if the measurements tell the whole story, then the null will be below the threshold of hearing just like the measurements suggest.
Well, I know better than to have these arguments here, but I got sucked in because I felt the new...