PxPixel
A Talk With John Chowning - Mixonline

A Talk With John Chowning

EXTREME VIBRATO AND OTHER ACCIDENTAL FLASHES OF GENIUS
Author:
Publish date:
JOHN-CHOWNING.gif

John Chowning, circa 1986

John Chowning pretty much sleeps when he wants and works when he wants. That is why when I'm talking to him at 10 a.m. East Coast time — and he's on the West Coast — he's been up and composing for about four hours already. “Now that I don't have institutional obligations, I find it's really great,” he says. “I remember hearing Buckminster Fuller give a talk about his lifestyle, and he said he'd work all the time, and when he was tired, he'd just take a nap. So I was inspired by that. Of course, Fuller says it's really hard on the rest of the family.”

Chowning, for those of you who just got up, was the inventor of FM synthesis, the computational technique that ushered in the era of digital synths, MIDI, desktop music production and much of what we've all been doing for the past 20 years. At the age of 70, he's now a professor emeritus at Stanford, where he was on the faculty for more than 25 years, which means he doesn't have to show up for classes anymore. So what's he doing? He's devoted himself full-time to what a great many of us would like to be doing: composing with all the neat new tools he and those who learned from him helped develop.

Chowning was the founding director of the Center for Computer Research in Music and Acoustics (CCRMA, pronounced “karma”) at Stanford, one of the most successful think tanks for music technology in the world. Some of the most important research in music synthesis and digital signal processing that we use today emerged from there, and among the many major figures who worked there were Andy Moorer, developer of the legendary SoundDroid for Lucasfilm and founder of Sonic Solutions; David Ziccarelli, writer of Opcode's original DX7 patch editor and now head of the wildly innovative software company Cycling '74; and Julius O. Smith, creator of what was to become known as physical modeling synthesis.

I ran into Chowning at the recent AES in San Francisco, where he was on a terrific standing room — only panel about the early days of electronic music in the Bay Area. A question came up from the audience (okay, it was me) about the future of electronic musical instruments, and his answer was short but highly thought-provoking. So I went up to him afterward and asked him if he would be willing to elaborate on it. A few weeks later, we had a fascinating 90-minute phone conversation covering that and many other subjects. So many, that this column is going to be in two parts.

Chowning was always a musician, never a scientist. He grew up listening to the big band music of the World War II era and started violin lessons in public school at the age of 7. A few years later, his junior high school band needed a cymbal player who could read music, so he became a percussionist. He served in the military and went to the U.S. Navy's music school where he learned jazz. “It was an amazing place during the Korean War,” he recalls. “The Adderley Brothers were there and a future member of the vocal group The Hi-Lo's. There was a very high level of playing.”

He then went to college on the GI Bill and studied composition, which he followed up with three years in Paris studying under the legendary Nadia Boulanger, teacher of generations of composers from Aaron Copland to Quincy Jones. In Paris, he heard, and was seduced by, electronic music for the first time, thanks to a concert series produced by Pierre Boulez. “It wasn't Boulanger's music,” he recalls. “She was more fond of Stravinsky and the romantic composers, but she had a fascination with it — Boulez in particular — and she encouraged us to go.” There he heard many of the great pioneers of the early electronic era like Luciano Berio, Henri Pousseur and Karlheinz Stockhausen.

After Paris, he went to Stanford for graduate study, but there was no electronic music there then. That would soon change: “My second year there, someone gave me an article from Science by Max Mathews who was at Bell Labs. I tried to understand it. It made this fantastic claim that any sound that could be perceived could be produced by a computer. So I went down to the computer science department and took a course in ALGOL [one of the first computer languages]. I contacted Max, who was at Bell Labs, and visited him. He gave me a stack of punch cards, which was the BEFAP compiler [Bell Labs' custom FORTRAN language] for the IBM 7094, which you needed to use Music 4, the music composition software that was available. I didn't understand much of what he told me, but then I read an article by James Tenney in Yale's Journal of Music Theory, and after that, I understood everything in Max's article.

“So I had this stack of cards, and I was wondering how I was going to get this to happen,” he continues. “One day, I was standing outside the computer center and this student walked up whom I knew, since he was the tuba player in the orchestra where I played timpani, and he asked me if he could help. That was David Poole, who was a sophomore math major, and he was hanging around what was going to someday be the A.I. lab. He taught me everything I needed to know. Among other things, he figured out a way to transfer the computer sample data in a dual-buffer arrangement so it could be output as a continuous stream. Up to that point, it was a two-step process: At Bell Labs, they had to write the output to a computer tape and then send it to a separate D-to-A converter. So this was probably the first online [real-time] computer music system.”

This process of going around to various sources until he could get his hands around a concept came to define Chowning's development. “I thought maybe I should go back and take some math,” he says. “So I enrolled in Algebra A; I think they called it Bonehead Algebra. I was already 30, 31 years old, and my last math class had been in high school. I struggled through it. I understood everything, I just didn't have the capacity to get through these tests in the few minutes we had. I had to beg the teaching assistant for a passing grade. I said to him, ‘Imagine you were taking a music course and I asked you to play an augmented sixth chord in the key of A-flat major. Musicians can do it right away. You could figure it out, but it would take some time.’ He said, ‘Yeah, yeah, I understand.’ So he didn't give me an A, but he did pass me. But I decided this was no way to learn what I felt I should know. Finding out answers to the immediate questions at hand was more important, but I needed to find the right person to explain it to me.

“At the computer center, the environment increased the number of timeshare users from eight or nine to 20 or more, and now there were all these people there I could talk to: engineers, psychologists, philosophers, linguists. So I built up an incidental education. If there was something I wanted to know, I would ask the same question of all these people, until I could finally get an answer in a way I could understand it. That's how I learned physics and acoustics.”

Chowning's first work at the computer center at Stanford in the early '60s was with reverberators and other spatial illusions in a 4-channel surround environment. “When I started out,” he says, “someone told me I would need vector algebra, and I said, ‘Yeah, right,’ but instead I thought, ‘How else can I capture this information about distance and Doppler Shift?’ Well, I just did it graphically. The lab had an arm with potentiometers in two angles, like a drafting machine, that plotted points on a CRT — sort of a precursor of a mouse. It plotted the points at a constant rate, so if you moved more slowly, the points were closer together, and if you moved faster, they were further apart. So I would just measure the points and that would give me the velocities. And it worked. Some of the mathematicians there laughed at me, but I got this 4-channel system to make these sound paths.

“That was the single most important point of learning for me: the importance of programming. I couldn't solder — I still can't — but I could make all these things with just a modicum of programming skill. I could bypass all the [engineering] detail and go directly from brain to output with just programming. I would write a subroutine to do a spatial path and another to do a circular path, and just use them whenever I needed to. The essential notion of efficiency came to me like a knock on the head.”

In Chowning's view, musicians and computers are not at all an unlikely combination. “Music is a symbolic art,” he says. “A painter gets the sensory feedback immediately, but musicians are used to writing things on paper and hearing them later. So they have to deal with symbols, things that are some distance away from where they are at the sensory level. It might be why music was the first of the arts to make so much artistic use of the computer. I know that other artists were working with computers at the time, but there wasn't this rush of activity — ‘I've got to get back to the computer center to work on my piece’ — that musicians had. And this wasn't the electronic music I heard in France. There was now this whole other dimension besides just producing electronic sounds.”

The idea of a musically oriented research environment with a variety of brains to pick resulted in the founding of CCRMA. Chowning usually gets the credit, but, he says, “I didn't create it — it just sort of happened. Andy Moorer, John Grey and Loren Rush were grad students there, and we were doing projects that came out of a collegial need. We'd ask each other, ‘What are you doing?’ ‘Can you modify that?’ Lots of applications would develop from that. Because what we were doing was interdisciplinary, it didn't fit in the music department, which was dominated by musicologists. So we decided we should form some sort of center that would allow us to apply for funding. I was the one on the faculty, and so I became the director. I chose good people — the idea was to make an open, accessible system and then leave people alone. The downside was that I became the administrator. There were fights to keep it intact and funded.”

Initial money came from the National Science Foundation and the National Endowment for the Arts, but a big break came when the Systems Development Corporation came across with $2.7 million. “Systems Development Corporation was a Defense Department contractor, and they had made an enormous amount of money,” says Chowning, “which they had to dispose of. We were one of four centers to get grants for computer music.” The grant came in large measure because of the efforts of John R. Pierce, another Bell Labs scientist (among whose myriad major accomplishments was coining the word “transistor”), who was so enamored of the center that he worked at CCRMA for more than 12 years as a “visiting” professor without ever asking for a salary.

It was in 1967 when Chowning first discovered the idea behind FM synthesis. “I was experimenting with extreme vibrato,” he recalls, “and I heard these inharmonic sidebands. I did a bunch of experiments, and I brought in an engineer to see whether what I thought I was doing was what the science would say that I was hearing. He looked at the equations and said, ‘Yeah, that's right.’ It was all very counterintuitive: Not a theoretical discovery, it was an ear discovery.

“But I was deep into the quadraphonic stuff so I put it on the back burner,” he continues. “In 1971, I was thinking about work that Jean-Claude Risset had done in additive synthesis [among Risset's contributions was showing that the harmonic spectrum of natural sounds changes with overall amplitude] and that Max Mathews had done in analysis synthesis, and I realized I could do the same sort of thing by coupling an amplitude envelope to a modulation index.” In other words, by varying the amount of frequency modulation over time, he could control the spectrum of a sound by using just two oscillators. “I realized it was all predictable, and within a few tens of minutes, I had some pretty passable brass tones. So then I wrote an article for the AES Journal, which was published in September 1973.”

Like The Beatles being turned down by the first few record companies their manager went to, Chowning's ideas on FM synthesis were rejected by several companies that Stanford's Office of Technology Licensing tried to get interested. Among them were Hammond and Wurlitzer. Chowning says of these companies, “Frankly, I don't think their engineers understood it — they were into analog technology and had no idea what I was talking about.

“But then the office put a Business School graduate student on the project, and he found out that the world's largest manufacturer of musical instruments, even though they didn't have much of a presence in the U.S. at the time, was Yamaha. One of their engineers was visiting their American office, so he came up to Stanford for the day. I guess they had already been working in the digital domain, because in 10 minutes, he understood exactly what I was talking about.” The rest, as they say, is history.

Though not exactly linear history. “Of course, the Yamaha patents made a huge difference,” says Chowning, “but they didn't begin to pay off for a number of years.” And in the meantime, Chowning had lost his job. “Like many universities, at Stanford you teach for seven years and then they either give you tenure or you're out,” he says. “No one understood what was going on in computer music so they didn't promote me, and in 1973, I had to leave.”

Next month: The answer to my question on the future of instruments, the meaning of the DX7, what musical instrument designers should be working on today and how the professor got his spot back.

Paul D. Lehrman is working on music for a film about the Governator and has a nontenured faculty position at Tufts University.