Given the vast array of choices, from tube or solid state to sources, analog or digital (and within either one, the additional choices of drive system, tonearm, cartridge or transport, format and DAC), speaker-amp interaction, subwoofers and overall set up within a given room, you begin to realize how much affect an overall system’s "voicing" plays a role in the results, leaving apart the quality of the recording, which can vary considerably.
To me, the hardest question is one which I cannot answer- that is, a grand or unified theory of system design that addresses all of these variables. The experienced listener knows it when they hear it, but beyond that, how predictive can you be without listening to a given combination of components in a specific room to know what sounds more like real instruments (that also raises the subjective aspects of listener preference/bias). This also takes time and evaluation on a range of material- something that really makes a difference in a positive way can have a negative aspect-- e.g. increased clarity which brings greater stridency on some material.
I came up during the beginning of TAS and the notion of evaluating how closely audio gear could replicate the sound of real instruments. I don’t listen to that much symphonic music anymore, but still listen to a lot of small combo jazz, recorded using acoustic instruments for the most part, without a lot of post production. I also know what different pianos sound like, having owned and played a variety of larger instruments. Those are my measuring sticks--for heavy rock or other genres, it is often a question of scale (how "big" the reproduction is before it sounds "forced"), dynamics (similar-do you hit a "wall") and congestion (does the system lose its coherence when things get cluttered in the recording).
One of many "tells" to me is the acoustic "envelope" of attack and harmonic decay which may be most evident on a piano though that instrument may be one of the most difficult to faithfully record. Seat time and exposure are invaluable.