Reference

One of the most firmly-established audio platitudes is the one which says "The specs don't tell the whole story." One reason for this, of course, is the fact that most manufacturers, preferring to sell their products on the basis of emotional appeals in ads rather than on hard, cold performance claims, do not attempt to make their specs tell the whole story. They publish only the most basic, and the most impressive-looking of their specs, although they are no doubt fully aware that what they do publish is only a very small part of what even the most astute buyer would need to know in order to predict what the component will actually sound like.

Things have come a long way, though, since the days when designers would expect two amplifiers that measure the same on IM and response tests to sound alike. Objective tests have become far more sophisticated, and meaningful, than they were even five years ago, and have reached the point where, at least for some components, measurements do tell just about the whole story.

Recent research on pickups, tuners, and tape recorders have turned up many new tests for these which allow the designers to predict with a fair degree of accuracy how they will actually sound. But in the realm of loudspeaker testing, we are still pretty much in the dark ages.

While it is true that a loudspeaker which measures poorly will almost invariably sound poor, it is equally true that speakers that measure exceedingly well often sound almost as lousy, in a purely subjective listening test, as ones that are objectively poor. This is why, of all components, the loudspeaker is the one that must be selected mainly on the basis of how it sounds. And it is why The Stereophile rarely bothers to confound its readers, or itself, with detailed measurements of the loudspeakers it reports on.

One amplifier manufacturer, in his Comment on one of our equipment reports (the Dynaco Stereo 120, in this issue), indicated his doubts about the reproducibility of subjective evaluations. The point is well taken, for a number of uncontrollable and largely unpredictable factors all conspire to affect the way a given loudspeaker will sound to a given listener. Room acoustics, amplifier damping, speaker placement, personal "preferences" and variations in program material will all determine how a listener will react to a loudspeaker, so it is difficult to see how a subjective report, of the kind we favor in The Stereophile, could possibly be of value to the cross-section of highly critical listeners who subscribe to The Stereophile.

To some extent, such reports must be "interpreted" by readers, much as they learn to interpret record reviews—in terms of whether they usually agree or disagree with the reviewer. The fact that a very large percentage of Stereophile readers find that they do agree with most of our equipment evaluations is not however a result of dumb luck, or of gentle persuasion. It is the result of a carefully developed system for subjective testing, that has been evolved over a period of 18 years of highly educational mistakes in the selection and use of all kinds of components.

Minimizing the Variables
We'll examine in detail The Stereophile's techniques for minimizing the variables in loudspeaker performance and providing test setups that actually reflect the loudspeaker's inherent performance capabilities. One of our readers, who asked to remain anonymous, related to us the results of an interesting series of tests he saw conducted on one of our top-rated loudspeaker systems. The response checks showed that the system had virtually no deep bass, a mid-bass peak, a midrange slump and a high-end rise. Further checks showed gross distortion at input levels over about 6 watts, and a definitely limited (although adequate for row-H listening) maximum output-level capability (footnote 1). Said reader then went on to ask how we could possibly consider such a loudspeaker to be one of the best available.

Well, we weren't going to argue with measurements. All we did, in reply, was reiterate what we had found on the basis of our listening tests: that the system in question sounds smoother and cleaner than any other we've heard to date, it reproduces without apparent attenuation (and very cleanly) bass notes that, according to the musical scores, are well below the limit indicated by the reader's response tests, and the system is as natural a reproducer as anything we've come across. And that, in the last analysis, is what counts.

A Matter of Taste
Now, this is the point in our argument where some of our more knowledgeable readers will leap from their seats and say "Aha! That's where they're making their first mistake. It may sound good to them, but different people hear differently, and what's real to The Stereophile may not be to somebody else."

Only partly true, we counter. Different people do perceive sounds differently, for the simple reason that different people listen to different aspects of a total sound.

Suppose we take two listeners, whom we shall call A and B, to a live performance of the Eroica symphony. If they sit side by side, both listeners will be exposed to essentially the same set of sound waves, but listener A may concentrate on the pattern of harmonies and instrumental textures, while listener B may be "tuned in" to the main themes and the rhythmic groupings. Each is hearing the entire complex of sound, in the sense that his eardrums and nervous system are responding to it, but each is perceiving—that is, is consciously aware of—different aspects of the sound.

---

Footnote 1: You may guess what system was under test, but don't bother to write and ask us to confirm your guess. The reader did not wish the loudspeaker identified either.