In the vein where the conversation wandered,
I've often thought it might be interesting to get together 4 players (as comparable as possible) a Fisher, Scott, Sherwood, Eico...
7591 based , 30wpc +/-, either all active or passive tone controls, set them up with same speakers (i had A25s in my head, no reason), play them back to back to back (double blind) and get impressions from a dozen or so peers.
Totally unscientific, just seat of the pants impressions.
I've also given some thought to the fact that if you ran same tests with same listeners when listeners were 25 Y.O. and today preferences might be different (compensation for loss of hearing)
Sorry for the side track