Google Sean Olive at Harman and check out his blog for more scientific information regarding listener preference. The gist is that the preferences of beginning listeners are all over the map, but as listeners become more experienced, their preferences tend to converge on a very specific set of criteria that confirms the science of making greta sounding speakers and refutes the suggestion that all listeners prefer something different and that speakers are like musical instruments.

Making music is art. Reproducing it is science.

However...some recordings suck and the quality of the listening experience can be dramatically enhanced by some significant coloration in the playback system (this is my postulate, but I'm not a scientist).