Propeller Head Plaza Technical and scientific discussion of amps, cables and other topics.

---

[ Follow Ups ] Thread: [ Display  All  Email ] [ Propeller Head Plaza ]

---

[ Alert Moderator ]

---

They aren't very sensitive (long)

Posted by JCA on November 4, 2003 at 15:33:02

In Reply to: Any interest?? posted by ophth on November 4, 2003 at 13:36:38:

ABX tests work really well when you're testing simple phenomena on a single subject. For example, somewhere on the Web (anyone remember where this is?) there's a site where you can download a software ABX tester and test and exercise your ability to hear fairly subtle differences (assuming you have good headphones and a high-quality sound card). You can run these tests repeatedly, get a statistically significant sample on a single individual (you).

But for whatever reason--and there's probably more than one--ABX tests don't seem to be very sensitive, especially when they're conducted on a population. More on this in a moment. Stereophile did a test a few years ago at one of their shows, where the question was whether it was possible to tell the difference between amplifiers. It was conclusive: those listeners could definitely tell the difference between two quite different amplifiers. The results were very clear, but it wasn't an especially challenging test.

This kind of test--where there's a whole population involved--is a very different case from the one where one person can do the same test over and over again. In the former case you're averaging across populations.

It's like clinical trials in medicine--you're trying to determine if there's a net positive effect across a population, which has very little, if anything, to do with the effect on one individual. We're all different. A treatment that works on 80% of the population may not work on me. And just because the FDA determines that, statistically, the effectiveness of a particular medication isn't adequate to allow approval, that doesn't mean that it didn't work for some, and it doesn't mean that for those it did improve, that the effect was psychological. It might have been a very real, physiological effect. This sort of statistical methodology has nothing to say about the effect on a single individual.

Similarly, the failure of populations to (say) reliably here the difference between speaker cables says nothing about the ability of one individual to hear the difference. For that you would need an ABX test conducted on a single invidual over several, or many, iterations. This is a legitimate test; you're testing whether THAT INDIVIDUAL can reliably here the difference. But under such conditions, one kind of distinctions can be heard? Simple ones, like certain kinds of distortion or obvious coloration, are easy to perceive, because there are certain specific earmarks (no pun intended) that you can listen for. That faint buzzing is there in one tone, but not in the other. Here's a guideline: if there's a word for it in the audiophile vocabulary, it's probably not all that hard to hear, unless the effect is very small.

Music is a complex aural experience, and for whatever reason it isn't always easy to get a handle on subtle effects.

But there's more than one way to "know" something. I'm a scientist, by training, but I don't assume that just because you don't have a conclusive ABX test, or some other conclusive scientific validation, that you don't know anything. There are certain situations where you want a definitive test for specific reasons--like new medications--but the absence of such a test doesn't mean the drug doesn't work, and it certainly doesn't mean that the drug doesn't work FOR YOU.

Some things I just know, even if I can't prove it. I'm not talking about belief here, or faith (I'm not religious), just simple, subtle things like whether I'm strong enough to make it up that next hill when I'm bicycling. For others it might be hitting a high note. Whatever. There's something subtle in your makeup, in your chemistry. It's intuition. Whatever it is, you know it, without proof, and sometimes it's quite subtle.

Music is subtle, too. It has time-domain aspects, pitch aspects, harmonics, spatial aspects. Hook your speakers up out of phase, or blow a tweeter, and most people with any experience can pass THAT ABX test with no sweat. But differences in the noise level (unless it's egregious, noise is usually perceived by its absence), or spatial effects...these are things you don't necessarily perceive as particular, simple experiences. You can know something is different, even better (for you) without being able to say precisely what it is.

I have a lousy ear. It's not that I can't hear differences between components, or cables. It's that I hear TOO MANY differences, or think I can, and I can't sort them out. My ear is untrained, undisciplined. My listening methodology is flawed. I can't tell the difference between real effects and psychological effects. I need the support of an ABX test to be sure of what I'm hearing; it also follows that I'd do poorly on many ABX tests. There's an interesting fact: those who need them the most to be sure are the least likely to do well at them. Those who would do well, don't need them. I'd be willing to bet that certain experienced ears would do just fine on a well conceived ABX tests, though it would be very involved and time consuming to prepare a test that exposed some of these subtle effects without being too obvious. And I'd also bet that there are plenty of experts out there who, despite their noted golden ear, would fail miserably.

Science is wonderful, but there are some things we just have to accept without proof. Or not. Your choice.

JCA

This post is made possible by the generous support of people like you and our sponsors:

Parts Connexion

---

[ Top of Page ] [ Contact Us ] [ Support/Wish List ] [ Copyright © 2025, Audio Asylum, all rights reserved ]

---

[ General ] [ Speakers ] [ Tubes ] [ Vinyl ] [ Digital ] [ Hi-Rez ] [ Video Asylum ] [ Cables ] [ Tweaks/DIY ] [ Music ] [ Films ]

---