I am constantly comparing live sound to reproduced sound - lifelong habit. I often close my eyes and listen to voices around me just to take stock of the character of real voices, and note how they differ from reproduced voices. I usually close my eyes in the presence of live instruments to do the same (e.g. just yesterday upon encountering a group on the street corner: drums, stand up bass, sax). Same when I play my acoustic guitar, or piano, when my sons practice sax, trombone...
When I do I perceive certain tonal qualities often more like tonal color. When I close my eyes in the presence of most sound systems, they just sound wrong. Voices, instruments may be dynamic, detailed, whatever, but the tonal "color" timbral match just isn't there. Tends to sound artificial, steely or plastic, or just like someone has dialed the color dial a bit off on a TV. (I also do this at audio shows - when a nicely recorded vocal is playing it can sound impressive in of itself, but if I close my eyes to directly compare it to the real voices inevitably talking nearby, the artificial nature of the reproduced voice sticks out compared to the fleshy, organic nature of real voices).
Harbeth were one of the very few speakers that survived the "closed eyes" test. That is: eyes closed, they seemed to produce the right tonal/timbral qualities I get in the presence of real voices and instruments. And they survive much better than many others the direct comparison to real voices. They get something fundamentally right about voices that most speakers don't get. IMO. (Some Spendor speakers also do this).
I think this is how a lot of Harbeth owners feel. That to their ears there is a particular "rightness" - where acoustic guitars, trumpets, sax, all sorts of instruments seem not just vivid, but timbrally accurate, right and natural.