Real or Surreal. Do you throw accuracy out the window for "better" sound?

Exactly... Thanks...

@snilf There is no corresponding "objective reality." That’s right. Everything that "is" must be somehow taken by us. No raw given, no way to check. Even the "real, objective" cello on the stage, playing live, is heard by me -- my sitting position, my ears, my distracted mind -- and, most important -- my interpretative taking of that acoustical experience.

If, in my home, I want to experience what I did in the concert hall -- ok, then I try to figure out how to do that. (And, as @mahgister points out: there are a hundred interpretive acts which are between me and that moment: engineers, mastering, etc.) But in this enterprise, let me not fall into the trap that I’m "really" getting back to something "more real." That’s folly and, worse, obfuscation. But it makes for some great chest-beating online.

I find that a speaker’s frequency response should measure flat on axis under anechoic conditions, and follow a downward tilted curve at the listening position something close to the Harman curve to sound best overall to me. That seems to be one measurable metric and established standard I can count on to work for me. Of course it’s not all that matters, but it is important and works for me every time. Indeed, some surreal and interesting effect can happen with boosted treble. I’ve been working on some new speakers recently and have been down that rabbit’s hole again, adjusting levels up and down by ear on the treble, mid, and bass to see what happens. The initial goal is to do tiny adjustments but then I sometimes go hog wild with it and hear some beguiling things. Like the OP said, the boosted treble and bass can be good at low volumes, acting as a loudness curve. Overall I just prefer to let the midrange become naturally more prominent as the volume goes down.

 I estimate myself to be in the more "real" than "surreal" camp. Some recordings sound surreal no matter what you do, and that’s fine. I don’t want to add my own particular flavor of surreal to every recording.

When I'm at a concert I don't hear precise locations of each instrument but can tell their decay easily. In a home system I'm trying to get all the small details as I can that to me makes listening more enjoyable. So I'd take hyper detail but not etched sounding over anything else. Each one of us has our own set of biases and that's why there are thousands of different types of speakers and tubed vs ss gear. We build our system to enjoy it ourselves not for others. 

I see a personal preference. What one likes is what one likes. It’s not going to be the same with every person.  
“Accuracy,” within the context of this, is a dubious definition to begin with.  
Accurate…how?  
Unless we were in the studio at the time something was played, we have no reference in determining how “accurately” our audio playback represents those performances. Even if we were in the studio at the time a performance took place, thusly providing such a reference point, after a few weeks (heck, depending on someone’s capacity for retention and auditory processing, maybe even after only a few hours) we have “forgotten” how it “really sounded” in the studio anyway.

I say, “have fun with your music, however that may be.”

Mahgister and hilde45:

I'm not a "nominalist," either, nor was I trying to suggest that either of you might be. ("Not that there's anything wrong with that.") My philosophical intuitions are mostly Kantian these days, after half a century contributing to the discipline. But this is not a philosophy forum.

Reality "in itself" (in the Kantian sense) is inaccessible and not worth even trying to talk about anyway, since it's inaccessible. Still, for practical purposes, "reality" can be said to be the source of the "source," as it were: the performance which the recording attempts to record and that our audio systems attempt to play back. Now, obviously we were not present at the recording session, nor do we have access to the acoustics of the recording studio in which mastering decisions were made. Etc. 

Nevertheless, a violin has a slightly different timbre than a viola (violists and violinists would challenge that qualification "slightly," but I presume you see what I mean). For that matter, a Guarneri violin has a slightly different timbre than an Amati, or a Stradivari. BUT...a "live" violin has a characteristic timbre, whether played in one's living room or in Carnegie Hall. If you are familiar with the sound of a live violin—especially if you are familiar with that sound in the acoustics where your audio system is set up—then you know, in your own head, whether or not, and to what extent, your system reproduces that timbre (mutatis mutandis: again, the differences between a Cremona instrument and a good modern "copy" are subtleties we can argue about in a different thread, and one can parse those differences in terms besides "timbre"). 

Bottom line (for me): I want to be able to hear the "voice" of the first violinist as against that of the second, and I certainly want to hear the voice of a violin in a string quartet as against that of the viola, or the cello. That's partly a matter of timbre, but also of imaging: I like to be able to "see" where the sound is coming from in the recreated "soundstage" of my listening room. That's essential to really following the music: I need to be able to discern the "voices" that make up the Gestalt. And that point bridges the "audiophile" concern for sound quality with the music lover's love of the music: it's difficult to fully appreciate what's in the music without a high level of fidelity to the subtle features of sound quality.

"Reality," as I was using the word (without all the philosophical asides), simply refers to the signature features of the live instrument as a perceptual baseline, against which the "accuracy" of the reproduction must be judged. 

Is this really problematic? Or unclear"