its a common woo belief that you can "train" the ear to hear these differences and people who work in audio like to think they've done so. the current understanding of the ability to hear the differences, however, is not based on "training" but physiological abilities of the inner ear. double blind tests of discrimination including self-classified "golden ears" doesn't show them to substantively better than a random population sample.