Ah yes, audio mumbo-jumbo. Before everyone jumps on my case, I do a fair amount of audio design and I DO know what i am talking about.
Most people forget there is a way to get very close to 'what it's supposed to sound', even without double-blind tests.
First of all, one can expect that the artist producing the music has given their blessing to the final product, say a CD, and that is a tacit admission that what is on that CD, assuming it was pressed correctly and has no uncorrectable errors, is IT, the reference material.
Next, get a pair of good headphones (it will probably be the last pair you get, anyway) and build a reasonably good amplifier. If you headphones are relatively high impedance dynamic, say 300-600 ohms, you should have no problem getting distortion to extremely low levels, and since it is a well behaved impedance, you will be eliminating 99% of all things that make a particular combo of amplifier and 'speaker' sound different. Then, providing you can separate tonality and spatiality, listen to the material, and you will be pretty darn close to what it's supposed to sound like. Phones have at least an order of magnitude less distortion than any speaker, perhaps excluding directly driven (no transformers!) electrostatics, even at unhealthily high volume.
If a recording is 'bad', then that's exactly how it's supposed to sound from a good system - there should be no 'warmth' where it wasn't in the source, and vice versa. I often find that tube afficionados tend to expect warmth from anything. Such people only declare the liking to a particular sort of distorsion, not to high fidelity. Saying that tubes sound better when clipping than transistors is really the same thing - no clipping is supposed to occur in the first place...
Speakers are the first point of suspicion. And if you go back into the amplifier, having seen many, even so called 'audiophile' ones, a lot needs to be done on education about proper high current wiring, power supply bypassing and filtering, magnetic shielding, and I won't even start on grounding. Some people need to be told the difference between inductors with a core, and ones wound on 'air', as well as between electrolytic caps and all others. Others need more subtle education, such as learning that relay contacts are not linear resistances (and for a VERY good reason) and that they are prone to microphoic effects. Etc, etc.
Saying that cables, tubes, caps, whatever, sound different means nothing without at least theorising why. It was mentioned that the human ear is a curios thing and that it can resolve what measuring equipment cannot. This is nonsense. What measuring equipment? A scope? It operates in the time and amplitude domains and is linear, very different from the ear. There is no ONE parameter that can describe how something will sound, therefore forget about distorsion meters and RLC bridges. There are methods and instruments that can show VERY clearly what the differences between parts are regarding sound, it's just that they are not common, and they are not cheap. A real-time frequency analyzer with spectrogram or waterfall plot capability can tell you a lot of things, but not everyone has one.
Someone mentioned that one listens with the brain. This is very important. I have participated in a study that tested 124 individuals for perception of reverbration. It revealed some interesting facts. One is that although the human ear may have a range which falls off severely over 16kHz, the human brain is capable of operating with phase shifts, that translated into a frequency, extend far above this, perhaps as much as an octave or more. It also suggested that the perception of amplitude is more precise than was originally thought of, and in fact, that it can be trained - a number of people in the test group were 'golden ears', (about 30%) their capability of better percieving minute changes in reverbration time tracked to 94%. Bottom line: things are measurable, what needs to be questioned are traditional methods of measurement.