I agree with Rushton. One way to put it is that true transparancy, and thusly true detail, has nothing to do with exagerated sense of detail via brightness or edginess, which are distortions. Noise that rides the signal, as opposed to noise floor, can result in a brighter, edgier, superficially more detailed sound but is actually masking true detail. Listen for a lack of harsh edge coupled with a sense of more information and naturalness, being sure to use reference recordings that you can trust to be naturally and well recorded. If you build your system around this you will have a magically revealing system that is very satisfying.
Bear in mind that bad recordings will sound bad, but you won't be adding insult to injury. You can, if you want, prioritize components that are smoothed over as opposed to simply lacking in harshness, and end up with a system that is forgiving of bad recordings, at the expense of low level detail that fleshes out the palpable you-are-there/they-are-here effect.
Assessing each component can be tricky though, since a better down stream component can reveal upstream coodies. It can be done though.
Bear in mind that bad recordings will sound bad, but you won't be adding insult to injury. You can, if you want, prioritize components that are smoothed over as opposed to simply lacking in harshness, and end up with a system that is forgiving of bad recordings, at the expense of low level detail that fleshes out the palpable you-are-there/they-are-here effect.
Assessing each component can be tricky though, since a better down stream component can reveal upstream coodies. It can be done though.