Quote Originally Posted by RobotCzar
While audio is a dynamic medium, the use of an ABX devices allows the audio equivalent of rapidly flipping from one picture to the other, Even tho music constantly changes, you can play a short segment over and over, switching at practically every point over the duration of the segment. You also have the option of playing a sound that does not change like a pure sine wave or white noise. It is a well-know fact that people can detect lower distortion levels with pure tones. If you cannot hear a difference with some set of pure tones, you will not hear any difference with music.

As for subtle differences, there are long-established statistical methods for finding small differences. One can find as small a difference as one likes by merely taking more samples (e.g., trials). By adopting the proper statsitical design, one would be able to test for very very small differences.

Listening to comparisons longer is just going to make hearing potential differences harder, not easier; but there is nothing stoping a person from taking as long as they like, especially using an ABX.

When somebody says "I can certainly hear differences in cables you tin-eard midfi oaf." one expects them to take a simple blind, level matched test and show that they can. One does not expect them to say, "oh, I can only hear the difference after listening for three months. The true believers will SAY anything to preserve their illusions, I and others would prefer they actually DO something to show that they can hear differences.
It does seem reasonable that a listener who is certain he can hear a difference beween two components in sighted listening should be able to do so in blinded listening. But it also may be reasonable to question the validity of blinded testing used for such a purpose. It would seem fair to ask if the testing does what it is supposed to do. But how do you test the test?

A starting place in taking a critical look at blinded testing might be to look at positive results. What kinds of subtle audible differences in components have been detected through blinded testing? Are any studies available? I'm not entirely sure what I mean by "subtle," but if the difference is very very obvious, the testing isn't being tested.