Similarly, I use Wilson, Reference, Water Lily Acoustics, and Sheffield Lab classical recordings to assess departures from tonal neutrality, pink noise to look for the presence of resonant problems, pretty much any modern rock album to assess the ability of amplifiers and speakers to present tight, well-defined midbass, good organ recordings and sinewave sweeps to investigate low-frequency extension, drum and percussion recordings to investigate "jump factor" and overall dynamics, in fact any recording where I can break the logical judgmental loop due to the fact that I have independent knowledge of how the recording should sound. As an aside, it is for this reason that I regard it as essential for reviewers to be actively involved in making their own recordings, because this gives them that knowledge absolutely. (Though, as has recently been pointed out with respect to Dick Olsher's use of his recordings of his wife's voice, this removes the capability of readers to repeat his test).

Which of all these parameters is most important I believe to be a personal choice, associated with the listener's musical taste. J. Gordon Holt, for example, would agree with Mr. Bernstein's apparent feeling that it should be tonal neutrality; as would Peter Mitchell, who, when asked by Bud Fried at the 1989 WCES (full report in this issue), replied that "good frequency response is a 'necessary but not sufficient condition' for great sound. Other things are important, but if the frequency response is irregular, the other qualities don't matter because the speaker won't reproduce the real sound of music."

But the most important thing for a reviewer is to use as wide a range of recordings as possible in order to frame value judgments which can be communicated intact to someone else. As, again, Peter Mitchell pointed out in a recent "Industry Update" feature (December 1988, p.47), merely to compare the sound of a component with the reviewer's memories of live sound can lead to error unless he or she uses a sufficiently varied selection of recordings.

So where does my idea, as expressed in November, of the reviewer using his visceral response fit into this neat framework of diagnosis and judgment? And how does live music fit into it?

As follows: The whole business of diagnosis described above seems designed to allow the reviewer more to get a handle on what a component is doing wrong than on what it does right. The latter, however, is more important—the fundamental purpose of a hi-fi system is to enable its owner to enjoy music in the home—and the listener who is sensitive to the music has a shortcut to discovering it. But in order for a listener to become musically sensitive, it is incumbent on him or her to regularly experience the real thing. How can you make yourself receptive to Mahler's music, as opposed to the sound of Mahler's scoring, if you don't expose yourself to the music as the composer originally intended it to be heard? If, then, a listener who regularly attends live music finds that a component somehow destroys his or her ability to enjoy the music—and this is true of all kinds of music—then all else is irrelevant, even tonal accuracy.

As J. Gordon Holt emphasized many years ago, a component must have the ability to allow the music to raise goosebumps on the listener's arm. Mr. Bernstein states that a reviewer must provide his reader with an intelligibly stated standard by which he will judge components coming before him. Ultimately it's goosebumps, Mr. Bernstein. Goosebumps.