The Music of Sound
Guest editorial in The Audio Amateur, issue 5, 1982.
Reprinted in Hi-Fi News & Record Review (England), Sept. 1985.
Copyright © 1982, 1985, 1997 James Boyk

 
Over the last few years, I have listened to a lot of reproduced sound with musicians, audiophiles, recording engineers, and students in "Projects in Music & Science," my course at California Institute of Technology.

I find myself thinking more and more that there are two modes of responding to reproduced sound. I'll refer to them as though they were people—which they're not—and call them the Musician and the Technician. The Musician knows and loves the sound of live music, and judges reproduced sound by how well it preserves the beauty and emotional impact of the original. The Technician does not know live musical sound, and judges according to some mental checklist of technical categories.

The two types were neatly separated for me recently by a remark made by Caltech professor F. Brock Fuller, a knowledgeable audiophile (meaning lover of sound, not lover of audio). In connection with our digital listening test, we were comparing the sound at a recorder's output with its input, a microphone feed of live music. Dr. Fuller said that if you ran down a recording engineer's diagnostic list of technical maladies while listening to the output, you would find nothing amiss, and would conclude that the machine was perfect or nearly so. A Technician, that is, would give it a clean bill of health. However, he continued, the fact that the doctor finds nothing wrong doesn't mean that the patient is necessarily all right; and he felt that this patient was indeed ill. For when we switched from output to input, we could hear that the unit had taken away much of the beauty of the sound, leaving it uninteresting and without musical impact.

I am not saying that the Musician listens to the music, and the Technician to the sound. I'm trying to describe two different ways of listening to the sound. The Technician listens in categories defined by the technology. The trouble is, there are not enough categories and never could be, because what's important in musical sound changes with the meaning of the music. What does not change—and this is what the Musician realizes—is that the sound itself is an organic part of the meaning.

For example, consider Chopin's Étude in E Major, Op. 10, No. 3 for piano—you would recognize it in a moment if I could play it for you. One phrase of this piece ends on E just above middle C. The next phrase begins on the E an octave higher. You know that every musical tone is actually a series of harmonics, and that the upper E is the second harmonic of the lower E and is therefore implicit in the end of the first phrase before it explicitly begins the second. Now on a Steinway—the instrument I play—the various harmonics of a single note develop over time. Which ones are prominent may change from moment to moment. When I perform this étude, I want a particular emotional relation between the two phrases, one which requires the second to appear in the most gentle possible way out of the first. I listen carefully to the development of the upper E implicit in the end of the first phrase, and I join the actual played note to it at the moment when they will meld most smoothly.

Few recordings convey the harmonic development of piano tone; most would damage this tiny detail of interpretation. Hearing such a recording, the Musician might be frustrated by inability to follow the meaning of the phrase join; but the Technician would not notice this lack, as the problem does not come under standard 'diagnostic' categories.

For another example, consider the clever and beautiful thing Brahms does in one of his trios when he leads a violin melody downward and has the 'cello take over at just the point where the sounds of the two instruments allow the most seamless join. (We tried crossing the melody at other points in my "Alive with Music!" class at Caltech, and verified that Brahms really did get it right!) If the reproducing chain treats the formants of the two instruments differently—formants are resonances of constant frequency characteristic of an instrument—the join will not be smooth, and the way Brahms makes you hear pure music instead of two instruments will fail. The Musician may be puzzled here, or may doubt the performers' sensitivity. The Technician will not notice anything amiss.

Such intimate relations between sound and meaning are the rule, not the exception, in music; and they show that listening categories must be flexible and music-oriented. Beyond this, the various aspects of the sound must be recorded so as to keep their organic relationships intact, because music is an organic expression of human communication, not a product with 'features' which can be added or subtracted at will.

What do I mean by this? Consider the performance situation. The tempo chosen by the artist depends not only on mood but also on the acoustics of the room. One plays slower for clarity in a reverberant hall; but in a 'dry' room such as the typical recording studio, a faster tempo helps to keep the music alive. If reverberation is added to a studio recording, the disagreement between tempo and acoustics makes nonsense; and since tempo is crucial to emotion, the feeling may become nonsense, too. The error here by the Technician producer is to separate the elements of musical sound.

The performance situation is a feedback loop of artist, instrument, room, sound, and audience. The artist is trained to maximize communication within that loop. Changing one element outside the loop—by adding reverberation, equalization or anything else—upsets the organic sense and throws away the artist's training and work.

Sometimes we must listen as Technicians, I suppose; but we should regard what we learn this way as no more necessarily relevant to musical fidelity than what we learn at the test bench. We do it simply to check things out. When testing our work against the goal of high fidelity, however, we must welcome into our perception all aspects of musical meaning.

We may never succeed in reproducing live music's combination of power, delicacy and beauty, nor its ability to involve us emotionally; but in the attempt to do so, we will learn much, not only about audio but about our perceptions and ourselves.