In article , RPS wrote:
If you were not present at the original recording session, with good
memory, how can you judge the accuracy of the reproduction?

A good question which I have often asked myself too. I don't have a
complete answer above the trite "I know it when I hear it" but I do have
some indicators which (I think) work for me. It's entirely possible
this is just my own personal definition with which no-one else will agree.

First of all transparency or neutrality to me means the absence of
obvious defects. For example:

- tonal imbalance: too much ("impressive?") bass; too much treble
smoothness or contrariwise too much "spit" in the treble

- boxiness: a sort of grey or tonally coloured resonance which
can also leave sound behind, blurring the boundaries between the
musical notes where there should be more silence.

It's a question of trying well-known recordings, especially if you have
heard them on more than one system - even if it's only having listened
on your own system, in the car, on headphones or (better) on someone
else's system. I have a few which I take with me to auditions.

As others have said voices are something everyone knows well from
long real-life listening experience. Voices sounding just right is
a good indication of accuracy. There are also defects to avoid here,
for example:

- excessive sibilance: can arise from a specific form of tonal
imbalance with a mid-upper frequency overemphasis

From attending concerts you will have an idea of the natural range of
sounds especially from acoustic instruments. I find small percussion
instruments if they sound natural to give a good indication of dynamic
accuracy.

Well, that's some of it. However it's your ears which matter and your
own likes and dislikes which are much more relevant than mine.

--
John Phillips