John Chowning: Audio Hacker and FM Synthesist

In 1967 FM audio synthesis was discovered by John Chowning during his experiments at Stanford University. It uses frequency modulation on audio waveforms in a similar manner to the way frequency modulation is done on radio waves. FM radio had come along before FM synthesis, and it was Chowning who first came along and did the research necessary to apply the pertinent equations to audio.
​   
John Chowning was born in Salem, New Jersey in 1934. As a young man he had joined the service and studied music at the Navy School of Music. After leaving the navy he went to Wittenberg, Ohio where he got a bachelor degree in music in 1959. Diploma in hand he hopped across the pond to Paris to study with Nadia Boulanger, who introduced him to the music of Pierre Boulez and Karlheinz Stockhausen. He had been bitten by the bug of electroacoustic and music and became fascinated with the idea of using loudspeakers in composition.

AUDIO HACKING

After three years of studies in Paris he went to Stanford in 1962. In 1963 Max Mathews wrote his famous paper on the Music IV program he had made using the computers at Bell Labs. It was an entirely new way of making music. In January of 1964 a friend of passed along a copy of Mathews article to him. At the time he hadn’t even yet seen a computer, yet one of the statements made by Mathews rang inside his head: “There are no theoretical limitations to the performance of the computer as a source of musical sounds, in contrast to the performance of ordinary instruments.” Excited by the ideas in the article Chowning took a computer programming course and it convinced he could learn.

He also got in contact with Max Mathews and made a visit to Bell Labs the following summer. Mathews gave him the punch cards that made up the Music IV program from the Bell Labs compiler. He took this gift back with him to Stanford.

The world of computer music was small, as it was totally new field. His school however was equipped with state of the art computers at the Stanford Artificial Intelligence Lab (SAIL), where a spirit of interdisciplinary investigation reigned. SAIL had been established in 1962 by Professor John McCarthy, a computer and cognitive scientist who founded the field of artificial intelligence, having co-written the paper where that term was first used. John McCarthy pioneered computer timesharing among different users, a solution to the long periods of time when the machines worked out the intricacies of a program to be output.
With the help of the friendly hacker and undergraduate David Poole he got Music IV program up and running at SAIL. They used an IBM 1301 disc, which served as the common storage unit between an IBM 7090 and DEC PDP-1 computer. The music generated by the program and saved to the drive. Poole helped Chowning obtain the audio by writing a double buffer program that eventually allowed for two analog signals to be recorded to stereo tape.

Fellow lab rats like Poole provided a hospitable and encouraging environment for learning and he taught himself the other skills he needed to know as one of the first computer composers and musicians: programming, signal theory and acoustical physics, all fields of study outside of his initial academic wheelhouse.

These new skills opened up further possibilities for Chowning to conduct deep research into the nature and properties of sound and music, and enabled him to translate the algorithms for RF frequency modulation into something that would work for audio frequency modulation.

​“Music is a symbolic art” Chowning has said. The Western classical tradition is accustomed to the role of the composer as someone who often puts music into notation before it is ever heard played by a full ensemble. The music came first from the imagination, then composed on paper, and only later played by musicians. The computer and its programming languages gave composers a different tool for musical realization, and in its capacity as a sonic instrument, gives access to a gamut of timbre that had before only been available in the platonic realm of ideas. With the tool itself realized, further realizations followed, and the ideas were able to be brought down from the platonic realm into listenable form.

SPATIALIZATION AND DOPPLER SHIFTS AND VIBRATO

As with many other composers of his generation who’d been stimulated by the work of Stockhausen, Chowning became very interested in the spatialization of sound. The experiments he conducted using a quadraphonic speaker set up around a listener in the shape of a square led him to his discovery of frequency modulation within the range of audio spectra and the subsequent creation of FM synthesis.

One of Chowning’s experiments was to divide the levels of intensity between the pairs of left and right speakers. The differences of intensity created sound illusions of distance or closeness. Next he worked with Doppler shifts and reverb effects to create the experience of sound moving within what he termed the “listener sound space,” an arrangement of speakers with listeners seated within.

Reverb was a key ingredient in his acoustic work and he discovered that if the reverb is applied equally to all channels it negates any spatialization or perceived distance effects of the audio. From this he learned there are two types of reverb: global and local. The global is applied to all sounds equally in a mix. The local is applied only to certain signals emerging from specific loudspeakers. Chowning then came up with an equation that showed how reverberation within a small space remains basically constant even as signal distance is increased. In a large space the equation can determine the distance of a sound based on the ratio of reverbant and non-reverberant signal.

With equations in hand, he programmed a spatialization routine with the Max program in 1972. It had a graphical aspect that allowed the composer to draw the trajectory of sound movement from one speaker to another. This program had two different aspects of velocity that could be used: angular and radial. The angular velocity is the rate of change of the sound intensity. The radial velocity worked with the rate of frequency shift in sound, i.e. the Doppler effect.

The Doppler effect is most often heard in everyday life as the sounds of objects moving closer to a listener, and then farther away. Striking examples are from traffic of all sorts such as trains, airplanes, and the whizz of automobiles and motorcycles, the blaring sirens of an emergency vehicle. The Doppler effect can also be experienced when there is a loud stationary source of sound, but the listener is moving around it, such as the bump of bass emanating from a house party on a Friday night while a couple walk their dog around the block where the party is taking place.  

Doppler shift was first discovered in the light spectrum by Austrian physicist Christian Doppler who wrote of the phenomenon in his 1842 paper, On the coloured light of the binary stars and some other stars of the heavens. Three years later Buys Ballot ran tests to see if Doppler shift was also present in the audio spectrum, and he showed that it was. He was able to show that the pitch of a sound is higher than the emitted frequency as sound source approaches, and becomes lower than the emitted frequency when it recedes. Further, the French physicist Hippolyte Fizeau independently discovered the property within electromagnetic waves in 1848. Since that time a number of equations came into use to mathematically model the phenomenon. The Doppler effect has gone on to be used in a number of settings such as radar, satellite navigation and communication, medicine, astronomy and the ubiquitous use of sirens, among others.

In working with sound intensity, Doppler effects, and reverberation Chowning realized there was much more going on in the perception of the loudness of sounds in space than just the distance and decay rate of audio as it travels. Vibrato was another factor in acoustics which could change the way a sound was perceived. Vibrato provided the next key he needed to unlock audio FM synthesis.

The computer generated waveforms Chowning created were not natural. In nature sounds are quasi-periodic, yet a computer is capable of making a perfect periodic sound. Some critics of computer music have pointed out the unnatural sound generated by these electrons. To make the timbres sound more natural variations have to be created in the waveform to make them quasi-periodic. Chowning did this by micro-modulating the frequency with vibrato.
​
This led to two discoveries. For one, when a sound is made of multiple partials, he realized that adding small but equal amounts of vibrato to each partial creates perceptual fusion. This fusion creates the illusion in the listener that the sound is one single tone. Perceptual fusion is also at work in film. The eye thinks all the motion is one continuous whole when it is in reality a sequence and series of projected frames. His second discovery was source aggregation. This can be created when small non-equal amounts of vibrato are applied to groups of partials. The listener perceives these as separate tones and sounds. He made extensive use of this latter effect in his 1981 composition Phone.

THE BIRTH OF FM SYNTHESIS

The same principle is at work in radio FM, where a carrier signal is modulated by the input signal, is used in FM synthesis. Audio FM synthesis is achieved by using one signal, called the modulator, to change the pitch of another signal, called the carrier, within a similar audio range. This modulation adds new information to the carrier signal and changes its timbre. The use of multiple modulators on one carrier gives the synthesist further variables for shaping the final sound signal. 

The stage had been set for this discovery as Chowning continued to explore the effects of vibrato. He noticed that when the rate of vibrato entered the audio range at 20 Hz partials started to form within the spectrum. He also noted how the relationship between the modulator and the carrier determined whether a sound was harmonic or inharmonic, as well producing changes in the timbre.

As he continued to explore he learned that if the modulator frequency is a whole number multiple of the carrier frequency, than the partials will be harmonic. Next he discovered the modulation index. This is the ratio between the depth of modulation and modulation frequency. He learned this could be used to change a signals bandwidth over time when the amplitude envelope of an entire signal is added to the value of the modulation index. This creates extra audible partials to change the sound.

​Similar effects had been achieved with additive synthesis, but those often require up to sixteen or more oscillators, whereas FM synthesis could achieve great results with two oscillators, the modulator, and the carrier, though more can also be used. 

FM MUTATIONS

In the summer of 1967 Chowning had visited Jean-Claude Risset and Max Mathews at the Bell Laboratories. A few months later in the fall he had made his discovery. In December he visited Bell Labs again. Risset took notes about what John Chowning had discovered with FM synthesis.

From his notes Risset did his own work and ended up creating the first composition using FM Synthesis, Mutations, in 1969.  Mutations was commissioned by GRM and was composed on computer and two-track tape. Made at Bell Labs it explores the idea of composing at the very level of sound itself, programming it and creating it all on the computer. Gradual changes or mutations occur over the course of the piece, “including the shift from a range of discontinuous heights to continuous frequency variations.”

​The piece used the endless glissando, or barber pole of sound, Risset had devised for a previous piece Little Boy in 1968. This musical barber pole was similar to the Shepard tones also created at Bell Labs using Max Mathews MUSIC software.
Mutations received its premiere at the Moderna Museet in Stockholm in 1970.

TURENAS

Though Risset gets to claim composing the first work to make use of FM tones, Chowning wasn’t far behind with his work Turenas in 1972. It makes use of FM synthesis, his surround sound set up and programming for Doppler shifts in Music IV. The word itself is an anagram of natures and Chowning strove to create realistic timbral sounds with artificial means.

The first of the pieces three movements makes use of the mathematical formula for a Lissajous pattern, also called a Bowditch curve. This is a pattern produced when two sinusoidal curves intersect, their axes at right angles to each other. It was first studied in 1815 by American mathematician Nathaniel Bowditch, while the curves were later studied by the Jules-Antoine Lissajous, a mathematician from France, who used a compound pendulum that poured out narrow streams of sand to study the pattern.

The curve is well known in the world of electronics where it can be made visible using an oscilloscope. With the oscilloscope the shape of the curve shows characteristics of electronic signals. The curves are used to study the properties of any pair of simple harmonic motions at right angles to each other.  

The Lissajous patterns came to be used in determining the frequencies of sounds or radio signals. A known signal frequency is put onto the horizontal axis of an oscilloscope, and the signal that needs to be measured is put on the vertical axis. The pattern that results is a function of the ratio of the two frequencies.

When Chowning had originally made a sketch of the proposed movement of sound in space for Turenas at Stanford, an engineer commented that it looked like the Lissajous pattern. Chowning decided to go ahead and use the Lissajous pattern proper. One of the properties of a Lissajous path is that its rate of change slows down as it reaches its peak, like a car set on cruise control at seventy miles per hour.
​
Chowning used a double Lissajous to surround the listener in these mathematical patterns. The second movement is a tour de force of everything Chowning had learned. He uses reverberation, vibrato, modulation, and many timbral transformations to showcase the veracity of FM synthesis. 

STRIA AND THE GOLDEN MEAN

The great astronomer and explore of the harmony of the spheres, Johannes Kepler said, “Geometry has two great treasures: one is the theorem of Pythagoras, the other the division of a line into mean and extreme ratios, that is Phi, the Golden Mean. The first way may be compared to a measure of gold, the second to a precious jewel.”

Chowning’s most famous work, Stria, from 1977 adheres with rigor to the use of the Golden Mean in the composition of all parameters and aspects of the work.  It also makes strict use of FM synthesis. Goethe said, “Geometry is frozen music.” Chowning took the sacred proportions of the Golden Mean and unfroze them so that they could be heard.

​The Golden Mean is often also called the Golden Ratio, or Golden Section and has been studied since at least the time of Euclid. It is commonly symbolized by the Greek letter Phi, giving it another moniker, the Phi Ratio. The Golden Mean can be found when a line is divided into two, so that, the whole length divided by the long part is also equal to the long part divided by the short part. In math two numbers are in the Golden Mean if the ratio of the sum of the numbers, x + y, divided by the larger number, x, is equal to the ratio of the larger number divided by the smaller number, x/y. Phi is an irrational number equal to 1.618, and then continues on, forever. 

The Golden Mean can be found in the sacred art and architecture of many traditional civilizations, from Egypt to Islam, from China to the great cathedrals of the Gothic Middle Ages, and many points in between. It can be found in many natural forms, such as certain leaves and the shell of the Nautilus pompilius. Wherever it is found there exists a manifestation of this natural harmony.

In his FM research Chowning discovered that when he composed using powers of the Golden Mean, applying them to the carrier-to-modulator frequency, low order side band components were obtained that were also powers of the Golden Mean.  
The macrostructure of Stria is related to the Golden Mean, and the microstructure of Stria relates to the Golden Mean. It all resolves around 1.618. The first frequency heard in the piece is 1618 Hz. All the durations in the piece relate to the Golden Mean.

​Stria was written using MUSIC 10 at SAIL, and travels from highs to lows as it traverses the mathematics of the Golden Mean in different ways. The precise use of computer controlled timbre and vibrato throughout give Stria a sound that is artificial, yet it is also natural sounding because of the use of the Phi Ratio as the structural component. Listening to it is like receiving a geometric download from the platonic realm.

CCRMA AND IRCAM 
                                              
Chowning founded the Center for Computer Research and Musical Acoustics (CCRMA) officially in 1974, though the basis for it had already begun inside of SAIL. The other founding members were Leland Smith, John Grey, Andy Moore, and Loren Rush. The first course in computer composition had already been given at Stanford in 1969, taught by Chowning, Max Mathews, Leland Smith and George Gucker. Having shared the space and valuable computer time with other researchers at SAIL it was soon time for those interested in the specifics of composing with computers to have their own department at Stanford.

One of the technologies developed by CCRMA was called the Samson Box, or the Systems Concepts Digital Synthesizer, the brainchild of MIT graduate Peter Samson. This system was used until Apple came out with their Unix based system. Michael McNabb composed his piece Invisible Cities (based off the novel by Italo Calvino) on the Samson machine.

Just like at SAIL the use of the Samson box had to be timeshared.  “Although the Box was a computer highly optimised for digital signal processing, we didn’t control it in real time because we decided to make it accessible to everyone, and ran a time-sharing environment so that most of the time in composition was spent in preparing the command files for the device. Once those files were written, the music — four channels of audio with integrated reverberation — could be produced in real time and recorded to analogue tape. The Box then became available to the next user in the queue. Running it as an assignable device like a computer printer avoided the problems that would have occurred if we had run it in a studio in which one user could tie it up for hours on end.”

Meanwhile in Paris in 1970 President Georges Pompidou tasked Pierre Boulez with founding an institution for musical research. Boulez then assembled his own team, which included the founders of the CCRMA, to build this sound-house at the Pompidou Center. It became the world famous IRCAM, or the Institute for Research and Coordination in Acoustics/Music. Chowning and his associates set up their French colleagues with the same computer system used at CCRMA. IRCAM was famous for its development of MAX by Miller Pluckette. Other innovations and application of research followed.

Chowning composed his piece Phoné at CCRMA. The piece later had its premiere at IRCAM. In Phoné Chowning expanded upon his previous compositions in FM synthesis to give the work the feeling and texture of the human voice.

For the community in Silicon Valley the CCRMA showcased their works in many outdoor concerts at the Frost Amphitheater, a venue used by the likes of the Grateful Dead, Jefferson Airplane and other stalwarts of hippie culture. The concerts of abstract avant-garde music made with computers became popular with the locals. “All these people who worked in the Valley then heard about these machines on which they're working also being used for a concert. And then we made it into a picnic thing at Frost. People would come early with their family and bring wine and get drunk and sit in the sun with the sunset. It became a happening, sort of. We always did really big sound systems and always quad. It was a big event and lots of fun.”
​
Their work continued on into the 1980s, adapting itself to new iterations of computers and programs, with new compositions by a variety of composers coming out of all the work. Chowning stayed on until 1996 when hearing problems, and the interpersonal fatigue caused by life in academia, caused him to step down into the role of professor emeritus. 

THE DX7

If all of this sounds a touch esoteric, then let it not be forgotten that John Chowning made a lasting imprint on the popular music of the 80s and onwards when Yamaha licensed the technology of FM synthesis to create their DX7 synthesizer. Work on its development began in 1974 but a commercial synth wasn’t available until 1983. With its ability to imitate acoustic sounds of piano, brass, woodwind, and other, as well as create new timbres distinct from earlier analog synths, it quickly became a hit with musicians when it was released. 

The DX7 was hard to program through its complex menus. Many who worked it with it used it’s  out-of-the-box presets, and these sounds became staples in 80s music. Brian Eno got one and it became a crucial part of the setup and workflow in his home studio.

Eno notes, “I use the DX7 because I understand it. I was quite ill for a while, and I filled the time by learning it. I think it’s just as good as anything else. Sticking with this is choosing rapport over options. I know that there are theoretically better synths, but I don’t know how to use them. I know how to use this. I have a relationship with it.”

The DX7 is programmed with 32 sound-generating algorithms, each a different arrangement of its six sine wave operators. These give the DX7 its classic bright and glassy sound. The keyboard itself spans five octaves, and has sixteen-note polyphony. New patches can be created within its deep menu system rather than with cables as had been done with analog synths and these patches could be named and saved inside its memory bank.

After the success of the DX7, Yamaha released a plethora of lower cost FM synthesizers. A cheap version of the DX7 soundchip also went into the Sega Genesis, making it the sound a generation of video game heads grew up jamming their thumbs too.
John Chowning thought the DX7 could also be used to teach about the properties of sound.

“Many basic acoustic phenomena can be demonstrated quite easily using the DX7. It could become an incredibly powerful tool for learning acoustics and psycho-acoustics at a very simple level.”

Since he stepped down from heading CCRMA Chowning has continued to hack audio, compose, and write. He has spent his life investigating the nature of sound and acoustics, he has programmed the music from his head to be output by computers, passing the vibrations from his mind to keyboard and mouse, until the airwaves of this world vibrate with vision. 

​​.:. .:. .:.
​
Read the rest of the Radiophonic Laboratory: Telecommunications, Electronic Music, and the Voice of the Ether.

References:
Frederick E. Terman, Radio Engineering, pp. 483-489 (McGraw-Hill, New York, 1947).

​Lawlor, Robert. 1982. Sacred Geometry: Philosophy & Practice. United Kingdom.: Thames & Hudson.

https://ccrma.stanford.edu/people/john-chowning
https://lifeorange.com/writing/ChowningAnalysis_McGee.pdf
https://ccrma.stanford.edu/sites/default/files/user/jc/fm_synthesis_paper.pdf
https://www.historyofinformation.com/detail.php?id=611
https://ethw.org/John_Chowning
https://cymatics.fm/blogs/production/fm-synthesis
https://en.wikipedia.org/wiki/Doppler_effect
https://www.magazine.columbia.edu/article/edwin-armstrong-pioneer-airwaves
https://www.thespectrummonitor.com/FEB2022TSM.pdf
https://www.britannica.com/biography/Pierre-Boulez#ref917618
https://www.npr.org/templates/story/story.php?storyId=97002999#:~:text=IRCAM%20was%20created%20more%20than,20th%20century's%20pre%2Deminent%20composers.
https://recollectiongrm.bandcamp.com/album/music-from-computer
https://soundart.zone/jean-claude-risset-mutations-1969/
https://www.britannica.com/science/Lissajous-figure
https://lifeorange.com/writing/ChowningAnalysis_McGee.pdf
http://digital.music.cornell.edu/center/wp-content/uploads/2020/01/40072593.pdf
https://en.wikipedia.org/wiki/Doppler_effect
https://www.sciencedirect.com/topics/engineering/frequency-modulation
http://www.personal.psu.edu/meb26/INART55/risset.html
https://cymatics.fm/blogs/production/fm-synthesis
https://electrooptical.net/static/oldsite/OldBooks/Terman-RadioEngineersHandbook_1943.pdf
https://ccrma.stanford.edu/people/john-chowning
https://en.wikipedia.org/wiki/Frequency_modulation_synthesis
https://en.wikipedia.org/wiki/Stanford_University_centers_and_institutes#Center_for_Computer_Research_in_Music_and_Acoustics
https://www.musicradar.com/news/brian-eno-interview
https://ccrma.stanford.edu/~aj/archives/docs/all/646.pdf
https://www.musicradar.com/reviews/tech/yamaha-reface-dx-629156
https://ccrma.stanford.edu/~aj/archives/docs/all/800.pdf
https://ccrma.stanford.edu/~aj/TheSoundOfInnovation.htm\
https://ccrma.stanford.edu/sites/default/files/user/jc/turenas-_the_realization_of_a_dream.pdf