If distortion is already low enough to be audible? Lower distortion but not audible? Maybe what produces those distortions are more audible than the measurements indicate?
That's the question, isn't it. There is always going to be
some distortion, so you may be right, at least up to a point. At what point does reducing distortion even matter? I myself don't think we are there yet.
Have you heard about 'resistor drive' where you add an 8R resistor (makes the amp 1/4 powerful) and that would make it
2:1 - and they say that they can hear the difference. Haven't tried to put that resistor in parallel. They should have tried it.
But maybe you are suggesting they are imagining that they are hearing a difference? And that the measured improvement is largely irrelevant because the distortion is low enough as it is?