The More I Know, the Less I Know
Indeed. It has been nearly 15 years since I started accompanying this magazine's equipment reports with measurements. Not only have I measured every loudspeaker Stereophile has reviewed since the fall of 1989 and every digital component since June 1997, I have been responsible for measuring every amplifier since June 2000. Editors as a breed are both pedants—I keep a workbook for every component I measure—and pack rats: against the wall opposite that New Yorker cartoon resides a large filing cabinet that holds not only all those workbooks, but a CD-ROM on which are stored all the measurement data files for 560 loudspeakers, 330 amps and preamps, 160 digital products, and about 100 other products ranging from microphones to turntables and even listening rooms.
That's one heck of a lot of information! Gotta mean something, right? So when you read, in Art Dudley's December "Listening" (p.47), the sentence "In this Promethean conflict [between objectivity and subjectivity], audiophiles who are technically oriented insist that measurements, because they are objective, are the best if not the only reliable way to describe audio gear," you probably thought that I could throw some light on what measurements can tell us about sound quality. Particularly as Art went on to state that he is forever bugging me to share my considerable experience on the subject, adding that "looking for and occasionally finding correlations between science and sound [is] fun."
Fun. Back in August 1989, announcing our intention to include measurement data in Stereophile's reviews (Must We Test? Yes, We Must!,"), I wrote that one of the motives behind this decision was to build up a measurement database that would eventually throw up correlations with what is heard. Yes, we have made some progress in some areas with some kinds of equipment. However, the more I learn about measuring audio equipment, the more there appears for me still to learn.
I was reminded of this Zen paradox—the more I know, the less I know—when I was cruising the Audio Asylum just before this issue's deadline. I visit the Asylum daily because its moderators' light touch maximizes debate by minimizing flames. Even so, the audio Internet is populated by people who seem to have never had a doubt in their lives. About anything. Especially not audio. One of these Internet bullies offered me some advice on the Asylum on how to write equipment reports. "I think you should be more critical about obviously bad engineering," I was told. "A good example is the first EgglestonWorks Andra. Who designed that terrible crossover???"
"Terrible"? Well, it certainly wasn't optimal. As you can read in the measurements accompanying our October 1997 review of the original Andra, the "terrible crossover" did produce an alarming suckout on the tweeter axis. However, this axis was only 33" from the ground. A listener with his ears 40" high or so would experience better integration between the tweeter and the midrange drivers. In addition, the farther back that listener sat from the Andra, the more his perception of its balance would be influenced by the room's reverberant field, which would ameliorate the influence of the speaker's flawed first-arrival sound. As I commented in the review, it was surprising how inaudible the Andra's crossover-region suckout could be much of the time.
But yes, I could hear it, and did comment in the review on the speaker's inability to reproduce timbres correctly, noting a slight degree of hollowness "that made violin and viola, for example, sound a little as if played with mutes." (Which was one reason I scratched my head when I read the letter from Stereophile's founder, J. Gordon Holt, on p.9 of this issue, in which he condemns all audio magazines, saying, "you'll...never find a statement like 'This loudspeaker cannot properly reproduce...timbres' in any review.")
More important, with the Andra, just as with the Horn Shoppe and Moth Audio speakers reviewed by Art Dudley in this issue, what one person dismisses as "bad engineering" is still the balance the designer worked to achieve. One man's "bad engineering" is another's acceptable tradeoff. Yes, the Andra changes instrumental timbres, but it will also never sound aggressive, and its distant soundstage perspective will compensate to some extent for the in-your-face balance of so many modern recordings. The speaker is not "accurate," nor is it a speaker I would choose to use, but does that mean it is "bad"? I don't think so.
I regard the job of an audio critic as being, first, to try to approach a product from the viewpoint of its designer, then to try to determine the "fit" between what that designer was trying to achieve and what the reviewer needs from an audio component. As part of that process, the reviewer must describe what a product does right and what it does wrong. But I do not believe that reviewers should dismiss or trivialize the efforts of those who have put their livelihoods on the line, particularly when they themselves do not have the creativity or design skills to do likewise.