Re the comments about fast vs slow switching in A/B testing. When I was at university I was involved in many A/B/X listening tests. The point was not to say which sounded better, but rather to say which of A or B was the same as X, which was randomly chosen. If a listener cannot reliably identify X as either A or B, then they can't tell the difference. While subjects were free to switch as frequently or infrequently as they liked, they invariably started to switch back and forth more quickly as the test proceeded.
Our auditory memory for fine acoustic details is relatively short, and people seem to recognize and compensate for this.