The Horse's Mouth (Stereo Review's 1983 cable tests) Page 2

"The striking outcome is that the panel accurately heard and named speaker cables in 5 out of 6 comparisons. The listeners felt pleased after this listening test battery: they had heard real differences! After the 50 hours of testing, scoring, comparing, and just plain listening, they were exhausted but felt accomplished. Both Monster Cable and 24-gauge wire could be heard reliably under double-blind conditions." (footnote 6)

Big difference, wouldn't you say? Yet Greenhill's original actually allows for two totally different interpretations of the data. One measure of the reliability of a series of identifications is known as chi square, or binomial, statistical analysis. Using this method, a prediction is made as to the probability of the identifications having occurred by chance. Another measure of reliability, one commonly used in psychoacoustic experiments, is the 75% rule: if a subject's identification (of the speaker cable, in this instance) is correct more than 75% of the time, the difference identified will almost certainly be audible—presumably to most people, most of the time.

When Stereo Review's editors wrote their conclusions, they essentially threw out all the tests where a subject was unable to identify the cables more than 75% of the time; Greenhill, on the other hand, was impressed that his group of 11 listeners could pick, with better than 999 to 1 odds, between 16-gauge zipcord, 24-gauge zipcord, and 11.5-gauge Monster Cable (using the chi square method of statistical analysis). Greenhill's original report, which is three times the length of the published one, makes it possible to evaluate the two methods fairly. It's certainly possible to distinguish between 24-gauge and either Monster or 16-gauge; one can reliably distinguish between 16-gauge and Monster (using the best test signal, pink noise), but the differences are right at the edge of audibility so that many listeners won't notice them, and the most acute listeners will.

I think it's unfortunate that more of the original report didn't make it into print, instead being sacrificed to Stereo Review's editorial viewpoints. Greenhill and his friends from the Audiophile Society (who provided eight of the listeners for the cable test) have tried several double-blind listening tests, and none of them have come up with positive results (ie, reliable identification of the components in question). Here they finally achieve success in the form of interesting results, but those results are obscured by the time the report makes it into print.

The first interesting result is that the listening panel's preconceptions of cable performance had a large effect on perceived differences between cables when they knew which cables were in use.

Second, differences were still perceived in double-blind testing, but to a much lesser degree.

Third, panel members were surprised that the differences between cables were so subtle and difficult to distinguish.

Fourth, the performance of different panel members varied widely: there was one truly amazing "ear" amongst them, and four very good ones.

Fifth, differences between very similar cables (none of them using exotic materials or cable geometry in their construction) could still be reliably picked out, even when (in one trial) the resistances of the different wires were artificially matched with a potentiometer.

Sixth, pink noise is a better test signal for discrimination than the choral music selection used (not necessarily all music). With a list of positive results such as this, it really makes you wonder why Stereo Review chose to emphasize only the negative.

Much can be learned from the coverage afforded the cable article. Stereo Review has used Larry Greenhill's article by distorting it to represent their well-established editorial positions. International Audio Review has used it to draw attention to its role as savior of the consumer and of the high-end industry. The Absolute Sound has capitalized on the humor possibilities, and taken the opportunity to again attack Julian Hirsch—at whose feet I feel this matter is unjustly laid. Hans Fantel, who must have read only the conclusions and not the data, used the Stereo Review article to justify his hitherto-announced scorn for esoterica. The Wall Street Journal seemingly has no axe to grind, but Gregory Sandow has used their pages to not unfairly promote The Absolute Sound, for whom he also writes.

Significant harm has been done, however. Truly esoteric speaker cables (of which New Monster Cable is definitely not one) have been maligned in the eyes of that portion of Stereo Review's 550,000 readers who are not intellectually discriminating—and without even being tested. The cause of "scientific" testing has not been helped; here was a successful test which earned its author and participants not fame but infamy. The test really pointed the way to many new and interesting experiments: does the ABX switcher (reviewed in these pages, in Vol.5 No.5) help or hinder such experiments; would esoteric cables show more marked differences in a similar test; what other components could be evaluated under double-blind conditions?

Larry Greenhill's position in the high-end community has been changed substantially, at least in the short run. Several manufacturers now refuse to talk to him, and a magazine he works for (High Performance Review) has even been denied the opportunity to review a certain manufacturer's products. Attacked by all manner of underground magazine—whose ranks he once felt a part of—Larry has hardly known which end was up these last few months.

On the other hand, he has only himself to thank. He did sign off on the article as printed, with adequate time for review, although there was a lot of pressure to accept the changes suggested by Stereo Review's editors. His personal opinion is that it would be wise for him (and for others who write for those magazines whose primary inte rest is attracting advertisements) to be most careful about what they approve for publication.

In keeping with Larry Greenhill's desire that the whole issue simply go away, we close our review of the cable article situation with a plea for a more objective stance from Stereo Review, who had the chance to expose as a success this attempt to differentiate between slightly varying components, and a plea for many more tests such as the one carried out by Greenhill (footnote 7).



Footnote 6: "Listening Tests On Speaker Cables," the original submission, pp.23 and 29, Larry Greenhill.

Footnote 7: Subsequent to the publication of this essay, Larry Greenhill joined Stereophile reviewing team, where he can still be found in 2007. See, for example, his May 2007 review of the Roku SoundBridge 1001 network music player. The debate on audible differences between cables continues to rage a quarter-century later.—Ed.

Article Contents
Share | |
Site Map / Direct Links