A Conversation with John Chowning
David Burraston
Stria LP
Important Records
A mysterious ratio, known since antiquity and rooted in nature, is used to create music quite “unnatural” that would not be possible without the precision of computer synthesis. Both the structure of the sounds themselves, inharmonic but orderly, and the musical scale through which they are revealed, are determined by the “Golden Section” – John Chowning
John Chowning is well known as the discoverer of FM synthesis in 1967 as well as being significantly active as a composer, researcher and professor at Stanford University. He famously collaborated with Yamaha on the commercialising of FM, resulting in the now legendary DX7 synthesizer, as well as many other instruments.
Each piece on Stria is an exploration of FM synthesis by its original discoverer, encompassing the 1960’s to the early 1980’s. All this was done at Stanford University on large scale computers, at a time before digital synthesizers as we now know them existed. The pioneering work of Chowning and many others helped make possible the existence of digital musical instruments today, whether in hardware or software form.
I discussed the pieces, their cultural context and the creative technological process involved in their realisation with John over email and video call. This article presents a summary of our dialogue, along with further research links/reading for those wishing to dive deeper.
John: “It [Stria] was a completely digital project from the very beginning. Although it was not easy to do back in the time of non-realtime synthesis. We had to make seven tapes I think.. so here’s the story.. after I finished it, it was to be premiered at IRCAM, at the Centre Pompidou in Paris in October 1977. So as we couldn’t put all the data on one tape or one disc file, we had to make seven different tapes, and then when I got to Paris a colleague of ours, James Andrew Moorer, was the systems programmer at IRCAM. So the two of us did the feathering of tapes together so it ended up being a tape of the complete piece. Then it was presented on October, 18th 1977. It was part of a series that Luciano Berio had initiated at the behest of Pierre Boulez called Perspectives of the 20th Century. So lots of people were there, and some of the people who became important in the spectral movement in France were present, and it made a big impression. Then three years later in 1980 Pierre Boulez presented it in a live presentation at the Théâtre d'Orsay, so that was very important because that was his stamp of approval, it gave Stria a lot of life. So that’s the story of how we got it to Paris, from a non-realtime system.”
It is not only a new sound; it’s not just a sonority that entices us, a new sonority or a new pleasure for the ear, but it is also something that forces us to change our thinking on how we put these sounds together and because of that this piece indicates a very clear direction, in that how we arrange these sounds together will for sure change completely and that the idea of composition will need to evolve precisely as a function of the new material that the computer brings us. Pierre Boulez recorded at the Théâtre d'Orsay, Paris. April 22-28, 1980 [translation]
Dave: “Was it rendered using the Samson Box?”
John: “No it was not, it was done on a PDP10 computer. I think it was a Digital Equipment Corporation [DEC] KL10 computer. It had washer machine size discs you could load and unload, so we put the output on the disc. But I couldn’t take the disc to Paris, so we had to put it on analogue tapes. So then we did the mix and presentation on analogue tapes in Paris.
And also interesting for your notes.. Well, when I first started working with the computer on the PDP6, which is the predecessor of the KL series, it was all in machine language but it was highly optimised for the DEC computer systems. Then they developed the language called SAIL, the Stanford Artificial Intelligence Language. The guys who wrote the SAIL language were of course in the lab. So when I began working on Stria I had this vague idea of what I wanted to do, the kind of structural relationships between the powers of the Golden Ratio and the tuning system, as you now understand. And so when I got this SAIL program there was a little note that said the procedures could be recursive. Well I was completely naive, my background was nothing but music. I had no math, I wasn’t a HAM operator, all I knew about was music, and I was well studied in music. I had studied with [Nadia] Boulanger in Paris etc, but I knew nothing.
So I was very naive and I asked one of the authors of the program, I said, ‘What does that mean, recursion?’ So he explained to me what recursion was and then what a recursive procedure was. And so I said ‘hmmm so a recursive procedure then, I can call the procedure within the procedure itself to replicate the procedure in some manner?’ He said ‘Yeah.’ I said ‘You mean like, as an example, I can write a canon, and say having 4 or 5 pitches in the sequence, and I can say one of the pitches will then become the reference pitch of another canon. So I have a canon within a canon?’ He said, ‘Yeah.’ So immediately I saw the application and then I developed these recursive procedures, and thats in one of the images I sent you. All the recursive procedures are highlighted kind of in a brilliant turquoise I think. With those recursive procedures I was able to build up this great kind of acoustic mass at the climax of the piece, which of course is at the Golden Section.
Everything is built on the Golden Section, subdivisions and everything. That was kind of an example of how computer languages themselves, like the natural language and poetry, computer language can give to the composer in this context more than he asked for. Which then generates new questions, which then generates new ideas. So its kind of a spiralling effect, a very powerful idea. I explained this to [Gyorgy] Ligeti in 1972 and he got very interested in that idea of program languages that can interact with the composer in a very rich kind of spiralling context. So at some point in his career he said ‘I write computer music but I don’t use computers.’ He was, I think, one of the greatest of that whole generation.”
Dave: “In your interview with Jim [Gardner] you mentioned Ligeti’s influence on Yamaha regarding microtuning as well. I have the mark one [DX7], I got the Grey Matter upgrade board which had the microtuning and I’ve got the mark two.”
John: “Yes, the mark two had it built in. Ligeti said he wanted microtuning, so thats one of the reasons Yamaha came out with it. There were other reasons, I asked for it. I said to them, look you can get a better piano sound because you can do stretched partials. The mark two was also used by Simha Arom, who did the study of Pygmy music.”
Dave: “Yes I read that in Jim’s interview too. Theres a paper somewhere that he published, you mentioned in that interview..”
John: “Yes, somewhere, you can find that online somewhere. They gave the Pygmies mallets, and they had them beat on bars that touched the keyboard, the pitches, of the DX7 mark two. So they had them tune them according to their ears. Something like that, it was a pretty complex arrangement. They had to carry in batteries and all sorts of stuff. But it was something that got Lygeti’s interest in Pygmy music and their tuning system. Yea, the DX7 mark two has an interesting history.”
Dave: “Just want to touch back on the previous aspect of recursion. Something I found interesting about that was the SAIL language implemented recursion, and you had been working with Fortran before that. Now Fortran is interesting because people were getting into recursion in Fortran but it wasn’t officially supported in Fortran until Fortran 90, so its quite remarkable that the SAIL team saw that was an aspect they wanted to put in there.”
John: “Yeah, well it had some LISP features too, because [John] McCarthy was director of the AI Lab and he was the LISP author.”
Dave: “Ah, yeah, of course. Because you also have Common Lisp Music at Stanford.”
John: “Yeah, yeah”
Dave: ”I played around with that back in the 90’s briefly, and LISP again earlier this year.”
John: “Yeah, CLM, Common Lisp Music, yeah”
Dave: “Regarding some of the other compositions if we go to Sabelithe... Tell me a little bit about that.”
John: “I began in 64 and Sabelithe I realized in 71, and it was the culmination of my work in spatialization using quad surround sound with Doppler shift. But it was really my test for what would become Turenas which I finished a year later. Sabelithe had all.. the spatialization was fully developed, the quad system. FM I had worked hard on and gotten it so that I understood all its relationships between Bessel functions and whatnot. So that was a short piece, I think it's about 8 minutes, and it propelled me into Turenas. I thought, I've got these ideas and I can do better. Sabelithe is an anagram of Elisabeth who was my wife at the time. It was a tribute to the long hours I was gone. Because I could only work at the AI lab in the off hours and on weekends. It was time-shared machines, so the best time to work was two in the morning, when the funded researchers had gone home. Because there was no competition for compute cycles on the time-share system. So that was a tribute to my wife Elisabeth, who we actually see quite often, she lives quite close to us here.
Then Turenas, which is an anagram of natures. That I think was the culmination of my previous 8 years of work. And at that premier Ligeti was there, Martin Bresnick, Ivan Tcherepnin and other notables.”
Dave: “Is that a relation to Serge Tcherepnin?” [Creator of the Serge Modular synthesizer]
John: “Yes it is, his brother. Ivan died early unfortunately. But his family.. it’s a long history. Ivan was a very close colleague. He presented the piece at Harvard soon after he left Stanford. He got a position at Harvard University in the Music Department. And he wrote the program notes that I like to use and incorporate in my notes. He saw the poetry in it.”
Dave: “Is Turenas the first piece where you used the Doppler shift and Lissajous figures?”
John: “Lissajous figures was Turenas. Doppler shift I used also in Sabelithe to some extent, but it was not very apparent. But the Lissajous figures I used extensively became the signature of Turenas. Lissajous figures were very important because they had a kind of naturalness that my ear told me was very close to how nature works. If you have a phase relationship between sine and cosine, as the sine curve reaches.. [it’s peak], its slope changes and as the slope changes the rate it goes is slower. So, the Lissajous figures had this characteristic when they changed direction they de-accelerated and then accelerated, as a bird or a mass in motion would have to do. So that was the attractive feature of the Lissajous figures – this naturalness that corresponded with what happens in the natural world. So nature is the title because it really was a gift. And all the FM is not my invention, it was a discovery.. although a kind of perverse application of a well known formula defined by Armstrong I think in 1928.”
Dave: “Because I worked at telecom I did all the different types of radio frequency modulation in communications theory, so I was... at the time when all of this was coming out in the 1980's I was just finishing my apprenticeship, and fortunately for me I had learned all about FM, AM, Ring Mod and all of that stuff.”
John: “So you understood right away then, right?” Laughs...
Dave: Laughing... “Yeah, it was great.”
John: “Because in the radio engineering text we looked at [Terman] used a sine wave as a way of explaining how FM works. And of course that was exactly the case in which I was applying it in the audio domain, where the carrier wave was in the audio frequency.”
Dave: “And all the Bessel functions were worked out in there so it all neatly falls into place, it’s just great.”
John: “Though they're nowhere present, they're used to explain it. I have to remind students that when they think about this there are no Bessel function numbers anywhere. It's all just an abstraction that explains what's going on when you modulate one wave with another.”
Dave: “The spatialization stuff is really interesting as well. When you consider at the time you were also consulting with Yamaha...”
John: “They first came in 72. We had also invited The Office of Technology Licensing [at Stanford] and invited others like Hammond, of course, Lowrey, Wurlitzer; they all sent engineers but none of them understood digital domain. But the Yamaha engineer they sent understood in 10 minutes what was going on. Because they had already foreseen that the digital domain would be the ultimate future for synthesizers and electronic organs and whatnot. So they had studied that. And this guy he understood right away. He went back and reported, and they quickly took an option on the license. And the patent was issued a little while later. But they were quick to buy into it.”
Dave: “On the archive page there's some great material about your initial conversations with Yamaha about the first tones that you were trying out, and some of the first ratios. I've noted here you've got 1.4 which is obviously the square root of two. We got 3.14 which is obviously pi. This is all mentioned in the 1974 correspondence. I find this stuff fascinating.”
John: “Yes and the golden ratio, I could remember that. Pi was no good because it's too close to 3 to 1. Square root of two was good because it gave distributed partials and the golden ratio did also. It was not close to a ratio of integers.”
Dave: “I've got the book [Chowning & Bristow 86]. This was a lifesaver when it came out, it really was. Even for someone like me with the technical knowledge, having something like that set everything out really well.”
John: “Dave Bristow was amazing.”
Dave: “How did that come about, for example, producing a book with Yamaha. Obviously there was a need for people to understand the technology because it was so radically different.”
John: “Yes, so.. [nods] because it was so radically different.. and no one really understood it.. there were a few people who got into it and understood, but most people just modified existing patches. The key of course was the little cartridge, where people could put their data on the cartridge and send it off anywhere. So, people in Singapore would make a modification and send it to Paris on an overnight flight, and the next day they had that in Paris. That was a very rich way to.. in the context of communicating data, it was exactly reproducible because it was digital. So that, unlike analog synthesizers, which were never stable and, well, you understand. So, that book came about because.. David Bristow, was given a fully equiped MIDI studio [at IRCAM] with TX816's, 2 or 3 of them, the whole thing.. keyboards. So, I then had a year to spend at IRCAM while on sabbatical. Then he and I got together and thought well let’s write a book about how you do this. So I would teach him the theory, and he'd write it all down and together we'd put together these examples. It was a wonderful experience. Then he proposed to Yamaha to publish it and they agreed to. Unfortunately, their products moved on, so pretty soon it was no longer applicable. They were always interested in new products, so they de-emphasized the book. I think they stopped publishing it in 85 or 86, I’m not sure.
Dave: “It's not easy to get a hold of now.”
John: “No, I know. But it's free on the download, right?”
Dave: “Yea, I’ve seen it. I've still got all my original photocopies with all my notes in so I didn't write on the book. From the eighties. I got the book from the library.
So back to the algorithms. My understanding would be that Sabelithe was a 1.4 to 1 pair?”
John: “Yes. I used a lot of 1.4”
Dave: “Turenas would be 1.4 to parallel carriers?”
John: “Yeah, and then Phoné is 3.”
Dave: “Were they [Phoné] fixed frequency carriers?”
John: No, not fixed. The carriers were assigned to the harmonic closest to a desired resonance.
Dave: “Ah, so they would change with pitch?”
John: “With pitch height, yes.
Dave: “Ah, okay, thanks for the clarification because I was wondering about that. I’ve been reading the Voices information, which presumably is the..”
John: “Yeah, the same idea.”
Dave: “It uses the same 1 modulator to 3 carriers.”
John: “Yeah, 3 carriers. And for the deep voices I modulated the modulator, which is like feedback FM. Feedback was a Yamaha development. In fact, the DX7 was the work of about 100 really good engineers at Yamaha. They determined the algorithms, that was not my work. I never saw a DX7, I saw all the breadboard versions in the many trips I made. But I never saw one until my wife and I went for a beer after a concert. He [the keyboard player] said come on over and showed me. What a surprise.”
Dave: “It must have been so surprising. That's a great story. Thanks ever so much for your time.”
----
Based on a video interview that took place on 13th July 2023 and some associated email correspondence. A huge thanks to John Chowning for his co-operation with this interview – DB
All images courtesy John Chowning, except video stills by DB.
Pierre Boulez recorded at the Théâtre d'Orsay, Paris. April 22-28, 1980 [Translation by Françoise Tourniaire November 22, 2016 corrected by Frédéric Dufeu March 8, 2019], courtesy John Chowning.
Scala microtuning files for anyone interested in experimenting with both the 9 and 18 note Phi tunings in Stria, rendered by Dave Burraston in 2023. Thanks to John Chowning for confirming these are the correct tunings:
Further info:
John: “One thing that's really important to me is that somewhere on the document that you write to note that the original 4 channel version is available through jc_at_ccrma_dot_stanford_dot_edu so that people know that they can also hear it [Stria] as it was originally intended.”
John’s page at Stanford University
https://ccrma.stanford.edu/people/john-chowning
John M. Chowning papers, 1974-2014. An extensive archive of John’s work/correspondence on FM at Stanford University
https://archives.stanford.edu/catalog/sc0906
Chowning, J. 1971. The Simulation of Moving Sound Sources. J. Audio Eng. Soc. 19, 2-6, 1971.
Chowning, J. 1973. The Synthesis of Complex Audio Spectra by Means of Frequency Modulation. J. Audio Eng. Soc. 21 (7).
Chowning, J. 1980. Synthesis of the singing voice by frequency modulation. In E. Jansson and Johan Sundberg, Sound generation in winds, strings, and computers, Stockholm, Royal Academy of Music, pp. 4-13.
Chowning, J., and D. Bristow. 1986. The Theory and Practice of FM Synthesis, , Yamaha Foundation, Tokyo.
Chowning, J. 2007. "Stria: Lines to Its Reconstruction." Computer Music Journal 31(3).
Chowning, J. 2008. Fifty Years of Computer Music: Ideas of the Past Speak to the Future. In Computer Music Modeling and Retrieval. Sense of Sounds (pp. 1-10). Springer Berlin Heidelberg.
Chowning, J. 2011. Turenas: the realization of a dream. Proc. of the 17es Journées d’Informatique Musicale, Saint-Etienne, France.
Baudouin, O. 2007. "A Reconstruction of Stria." Computer Music Journal 31(3).
Menneghini, M. 2007. "An Analysis of the Compositional Techniques in John Chowning’s Stria." Computer Music Journal 31(3).
Zattra, L. 2007. "The Assembling of Stria by John Chowning: A Philological Investigation." Computer Music Journal 31(3).
Interview: John Chowning, Jim Gardner
https://www.rnz.co.nz/concert/programmes/hopefulmachines/20130917
Son[i]a #212, John Chowning, Ràdio Web MACBA [interview]s
https://rwm.macba.cat/en/podcasts/sonia-212-john-chowning-2/
MTA SYNTHFEST 2013: Dave Bristow & The History of Yamaha FM Synthesis
https://www.matrixsynth.com/2013/03/mmta-spring-synthfest-2013-dave-bristow.html
Arom, Simha. 1991. A synthesizer in the Central African bush. A Method of Interactive Exploration of Musical Scales. In Für Ligeti. Die Referate des Ligeti-Kongresses Hamburg 1988, Laaber, Laaber-Verlag,, pp. 163–178.
Johnstone, Robert. 1994. The sound of one chip clapping : Yamaha and FM synthesis. MIT Libraries. [pdf]
https://dspace.mit.edu/handle/1721.1/17123?show=full
Terman, Frederick Emmons. Electronic And Radio Engineering, various editions from 1932 onwards, McGraw-Hill. This is the definitive text John mentions where he first studied FM theory.