It's a fascinating subject. For starters let me give my opinion on reproduction: music reproduction is a representation of an original -- NEVER the original. In the same way that a portrait is NOT the original, likewise, blind (i.e. non video) music at home is an interpretation of the original material; in this case the LP or disc -- NOT the live event.
Trying to give a structured answer to Tom's question, I'd say that my present take on the subject is that the "differences" relate to FOUR factors... in no order:
* The kind of music we favour.
Classical is acoustic, rock is electric often with processing thrown in. So, harmonics captured by the classical recording *should* reappear in the reproduction -- or we don't like the result. Classical has a lower bass content of ~40% to rock's +50%. So, the system priorities in either case would be different, in a real world rig (I readily concede that a well balanced 100k 2channel system can probably do both well -- but that's exotic for mere mortals).
* The sonic preference or goal pursued: Linkwitz (of xover & Audio Artistry fame) once remarked that he wanted the reproduced sound to be as close as possible *to the original* i.e., the recorded signal. That's one view. Another is to make a system emminently enjoyable (usually giving a few db boost to the 100-5000Hz region). Again, the appraisal of a system, by people favouring one or the other approaches, will be different.
* What are the sonic details we can live without vs. those we cannot accept compromise for. Viridian was very clear on this.
I, for one, love to get transient attack, as if experiencing a 20db headroom (i.e. 1-100 difference) is the minimum threshold to preserving my psychological well-being. Of course I also need pitch otherwise I get a stomach ache. So, of course, when s/one asks about, say, Spectral or Symphonic Line, I will find both products exquisite. Of course they have the ultimate extension in the highs & the lower low's. Do they have the pronounced & extended midrange of a good tube? No, in both cases, and by comparison the midrange is a bit "recessed".
*FINAL: most of us have scant knowledge of electroacoustics & the maths related to speaker performance. Unfortunately, no one seems ready to give us a few intelligible hints -- appart from this & other communities exchanging info.
So, we readily concede that speaker X performs better than speaker Y with amp X (amp X,Y vs. speakers, etc) when all that's happening is that the loads are iffy, the speaker's crossover components are of the 0,5c variety (and we buy expensive cables to correct the uncorrectable), etc.
Worse, very often even speaker placement is terrible CHEZ the professionals -- hence, a somewhat iffy tonal balance speaker placed facing straight ahead (we're listening well off axis) with an iffy amp/pre combo sounds good, compared to a good speaker playing the same pre/amp combo. Dangerous conclusion: the "better" speaker is worse.
It isn't.
Anyway, pls excuse the rambling