That shows how important it is to control for all variables. I understand it is difficult to match volume/loudness when AB testing speakers, amps or codecs and that might subtly influence someone’s judgement.