Audio & Alternative Medicine Letters
Blithely unaware of the simple truth?
Editor: Even though I am a confirmed objectivist when it comes to high-fidelity systems, I continue to enjoy Stereophile. Your coverage of music and transducers (mostly speakers) is valuable enough for me to continue my subscription. It would be nice if you took a more balanced view and tested more modest components, but I'm sure that I'm in the minority of your readers in this request.
Having said all that, George Reisch's "Undercurrents" in the May issue really shows a complete lack of understanding of testing methods as used in modern science. Reisch is, as usual, trying to demonstrate that subjective testing is something other than nonsense. In the subjectivist tradition, he seems blithely unaware of the simple truth that any proposition can be proven by analogy. Let me discuss and refute each of his major points in turn.
GR begins by trying to draw an analogy between alternative medical techniques vs conventional medicine and subjective vs objective testing of audio components. He totally ignores the absolutely critical point, which is that the testing of these techniques will be done using standard double- or single-blind methods. This is not an endorsement of subjective testing at all, but exactly the reverse. I agree that keeping an open mind about what might be true is important before testing has been carried out. That's completely different from saying that subjective observations constitute scientific demonstration of a fact. Incidentally, that's true even if a US Senator disagrees.
Second, George sets up the straw man that audio objectivists demand not only that the audible differences between cables be proved, but also that they be explained using the laws of physics. There may be those out there who say this, but generally, he's wrong about this. What we actually say is that subjective testing techniques demonstrate nothing scientifically. We also say that there is no explanation in known physics of why, for example, using silver or OFC in a cable would have any audible effect on music. In effect, we are just warning people that not only hasn't the case been demonstrated, but that it doesn't even make sense. We also sometimes point out that the profit margins on very-high-end equipment provide a mighty incentive for salespeople and owners to claim they are "hearing something."
Then GR gets into the whole muddled area of quantum mechanics. Frankly, I understand (as well as most intelligent people) that there are paradoxes in the standard physics model. However, there is no such paradox known to exist in the double-blind testing model. In fact, the EPR paradox is theoretically reproducible, over and over again—it's testable. This, incidentally, is one of the important standards of testing scientifically. We know something is true when it is reproducible. I am suspicious of anyone who doesn't report a result that can be repeated. Yet that's exactly what the subjectivist audio testers strive to do: no standard room, no single change in system, no measurement of distances, no use of sound-level meters, etc. At the risk of being redundant, the whole point of the EPR paradox is that it is repeatable; that's how we know it exists.
Fourth, GR states that individual differences in the ability to perceive what music "really" should sound like are what separate him from me. I can't hear any differences between well-made cables; he can. Okay, If that's the case, it's easy to prove it. Just have someone substitute cables or any other component without his knowledge in 50 trials over a two-month period. If he can identify in a daily diary exactly what's been done, publish it. I can be convinced. I even bought a Japanese car last year after eight BMWs. But he doesn't do this, and the manufacturers of cables somehow never do this, etc., etc.
Fifth, GR carries on about how controlled tests destroy his ability to discriminate. Well, again, it's easy to do the experiment I outlined above and learn that he has no idea at all of what's been changed in his system. At this point, I would have to ask how not knowing in advance makes him unable to hear what have been described as audible differences. I don't see why you cannot do single-blind, single-substitution testing with a number of listeners and a number of trials to determine the facts about subjective listening tests. Why does this work perfectly in all other fields?
As a relevant aside, I was enormously amused to read Sam Tellig's column in the same issue. Tellig was fooled (some 20 years ago?) by a pair of ProAc speakers, thinking them Quads. Did he realize that he actually couldn't tell the difference? Of course not; he simply ignored the results of his own ears. Now he says he knows better. Come on, Sam, why not demonstrate it? If you can't, then why not just say you know this is all just crap?
Finally, Reisch talks about the idea that there may be some small fraction of the population who "somehow" benefit from an alternative treatment while the majority do not. Of course this is possible. However, when scientific testing is done, this possibility can be ruled out by using placebos, multiple trials, and a large sample. Using probability and statistical techniques, we can determine "how" true something is. As anyone can read, it's very common for medical conditions like back pain to simply disappear over time. In a recent Consumer Reports article, several of the "complaints" had 30% "cure" rates regardless of the therapy. You could look it up.
It's funny, because I suspect that all of this subjectivist nonsense really causes everyone who is interested in accurate sound to waste time worrying about what are, at best, nuance issues. The frequency response of my listening room is ±10dB and that of my speakers is ±3dB. They also have a non-uniform polar response. So do yours! If we want sound to be close to what's been encoded on the CD, then we need to figure out how to tailor the frequency response and dispersion in our listening rooms to get the best possible sound.—Ross Salinger, email@example.com