Interviews

It is a widely held belief that musicians do not assess hi-fi equipment in the same way as "audiophiles." I remember the British conductor Norman Del Mar—an underrated conductor if ever there was one—still being perfectly satisfied in 1981 with his 78 player, never having felt the need to go to LP, let alone to stereo. And some musicians do seem oblivious to the worst that modern technology can do. I was present at the infamous Salzburg CD conference in 1982, for example, where Herbert von Karajan, following one of the most unpleasant sound demonstrations in recorded history, announced that "All else is gaslight!" compared with what we had just heard. J. Gordon Holt proposed a couple of years back ("As We See It," Vol.8 No.1) that sound is not one of the things in reproduced music to which musicians listen. I have also heard it said that even the highest fidelity is so far removed from live music that a musician, immersed in the real thing, regards the difference between the best and the worst reproduced sound as irrelevant to the musical message: both are off the scale of his or her personal quality meter.

Whatever the truth of the matter, there are, as with all widely-held beliefs, as many exceptions as examples, one such being pianist James Boyk, Artist-in-Residence at Caltech in Pasadena since 1974, where he runs an interdisciplinary course called Projects in Music and Science (footnote 1). Jim is also president of, record producer for, and sole featured artist of, the small record company Performance Recordings. (He tells me that he also sweeps floors, and answers the phone.) Despite being a classical pianist of renown, Jim is acutely sensitive to small nuances in reproduced sound, and, perhaps as a result, his recordings are object lessons in how to capture the natural sound of a piano in a real concert hall, as opposed to the usual multimiked, cotton wool-acoustic, musically-sterile, sound featured by such companies as DG.

At Caltech, Jim has the wherewithal to perform one of the really meaningful listening tests: to compare the sound of a live microphone feed with the sound when a hi-fi component—gain and absolute signal polarity adjusted appropriately—is inserted into that feed. As a result of such tests, Jim is committed to tubes for amplification, and his opinion of what digital has to offer lies in stark contrast to, say, J. Gordon Holt's or, indeed, to those of the engineers for the major recording companies. Needless to say, his own recordings are strictly analog, and made with tube amplification, hence Performance Recordings' "All-tube Analog" trademark. He also uses ribbon microphones exclusively rather than the ubiquitous condensers. I visited with Jim in July, and we talked about his listening tests and his aversion both to transistors and to music by numbers. But first, at the beginning of what composer Percy Grainger might have called a "free ramble," we talked about the purpose of high fidelity . . .

James Boyk: You're trying to preserve the sound of the music with high faithfulness, but in aid of what? The purpose of the exercise is to convey the emotion within the music. Music is about the communication of emotion.

The first thing to know about audio is that live musical sound is beautiful. That's the absolute bottom line. By "beautiful" I don't mean namby-pamby pretty, but something very definite: it means that the sound draws you in instead of puts you off. It means you can listen for almost arbitrarily long amounts of time without getting fatigued by the mere act of listening. (You may get fatigued from your emotional involvement with it, but that's a different matter.) When was the last time you heard a hi-fi system at any price where listening for three hours wouldn't fatigue you? You hear live music for three hours and it doesn't fatigue you.

The first requirement for doing a good listening test, therefore, is that the sound must have a reasonable amount of that beauty. It's easier to do that with tubes than with solid-state, in my experience. It's easier to do that with ribbon mikes than with condensers. You're helping yourself a lot, of course, by coming straight from the microphones rather than using stored material—I don't care whether it's LP or CD or tape, straight from the microphone is just radically better—but you must make sure that the sound you hear out of the loudspeakers on your direct feed is beautiful. (That's already very unusual: you go into the typical commercial recording studio here in Los Angeles, and what you hear straight from the microphones is a travesty. I would set up what they felt was their best microphone on piano, somebody would hack on the piano, and I would listen on the direct feed through their board—there wasn't any remote semblance to piano sound!)

So you set up two microphones with musicians playing into them, and you run the signal through your microphone preamps, gain control, power amp, and speakers. You break the line at one point with a switch and insert the device to be tested. Obviously you can't do it with all components. For instance, to take the transducers at the two extremes of the playback chain, microphones and speakers, you have to listen to live acoustic music, then listen to what that transducer does in a system. That's really problematical, because no microphone set-up hears spatially the way your ears do. But some components—tape recorders, for example—are perfect for insertion in a direct feed. All a tape recorder is supposed to do is to act like a black box that does nothing to the input signal except delay it by an amount of time at your command. And repeat it at will. So you drop the tape recorder into a line, adjust for unity gain, and see if it changes what happens. The question then becomes, not "Do I like it or do I not like it?", but "Does it change the sound, is it different? Is it the same as the piece of wire?"

Of course I'm leaving out all kinds of technicalities, but it seems to be a completely valid idea. If you can't tell the difference when you insert the tape recorder, then you've got a tape recorder that nobody's yet made—a very good tape recorder!

Atkinson: I understand that you've been using this listening test to evaluate tube and transistor amplifiers.

Boyk: This is a kind of open-ended project that was started in my Caltech course about three years ago. What we have are two stereo line amplifiers, identical in the following ways: they're both designed to work more or less around the 1V or 750mV level; they're both x10 voltage amplifiers; each has a follower after the voltage amplifier, so that they can drive some kind of output load without too much difficulty; they're both very simple; and they're designed as much as possible to have similar figures for frequency response, distortion, overload level and so on. Then each is padded down with a resistive network so that the net result of each of these things is nothing, no gain. However, one uses tubes and the other transistors.

The idea is compare the sound of one with that of the other. (They both invert phase, so there has to be provision to compensate for that.) We set up our microphone feed with a three-way switch: the middle position, labelled "reference," lets you listen straight to the microphone and mike preamp. You listen for a while till you feel familiar with what you're hearing. Another position drops the tube line amp into the circuit, and the third position does the same for the transistor line amp. Then the question that is asked is "Which one comes closer to having neither of them in?"

When the student doing the project feels that there's a definite difference between the two, I give him a blind test, and if we get to the point where the student can tell them apart, I say, "Fine, now you're allowed to have an opinion about which is better. Which one is better?" He says "That one!"—whichever it is—and pulls off the black velvet cloth. Lo and behold, the tube line amp is better.

---

Footnote 1: James Boyk retired from Caltech in 2004 and is devoting himself to writing and performing. You can find his website here, his bio here, and his bibliography here.