The garden in the machine: disembodied music
Testing a protective shield. Advanced technology has allowed the exploration of Earth, the solar system and beyond. Photographer Cesar Acosta for NASA Ames Thermo-Physics Facilities Branch.
A couple of months ago I set myself the challenge of creating two compositions to explore our changing relationship with technology. In my previous post The garden in the machine I went into detail about this, describing my desire to create one highly embodied composition and another highly disembodied one. I wrote the embodied composition for the handpan and shared it in a previous post. The disembodied composition is complete and can be listened to here.
If possible listen on good speakers/headphones.
Writing music in code
A fragment of the composition visualised as a waveform using the music programming language Supercollider
Using code to write music can feel like departing from the physical world and entering into a realm of ideas, absent of people and spaces, where there is nothing to blow, strike, pluck, or bow, yet there is music. It uproots the ancient collaborative world of sound and replaces it with a more solitary one, familiar to the novelist or painter.
While this might sound like a loss — and were it the only way to work it would be so — a machine can provide a certain kind of freedom. The imagination is liberated, along with the disappearance of constraints such as budgets, physical spaces, and archaic cultural habits. Used in a certain way, machines take on an almost metaphysical nature, summoning intangible worlds of ideas that can influence our reality as powerfully as the ‘real’ world.
The computer artist Vera Molnar even makes the case that certain aspects of our humanity may only be expressible through a computer.
Without the aid of a computer, it would not be possible to materialize quite so faithfully an image that previously existed only in the artist’s mind. This may sound paradoxical, but the machine, which is thought to be cold and inhuman, can help to realize what is most subjective, unattainable, and profound in a human being.
The same can be said for music. Certain compositions are only possible to create with a machine, yet are capable of expressing something deeply human.
A disembodied composition
This composition cannot be notated in a way that makes it readable for musicians, and even if it could, it would be impossible to perform. Each of the 1,000 different instruments plays a unique part in a gigantic polyrhythm, resulting in music that has no measurable time cycle with which to orientate oneself; even though the underlying tempo is as regular as a clock, the multiple layers of time make us lose our place in its flow. A machine is ideally suited to this task, as it is capable of the rapid computation required to accurately coordinate masses of instruments.
For their sounds, the instruments create the simplest possible tone in the universe — an electronic sine wave. Sine waves have a circular wave pattern irreducible to anything simpler. Metaphorically, I like to think of them as atoms, or particles of sound, although this is literally inaccurate since sound is actually as continuous and unbroken as the waves of the ocean.
Sine waves, which sound similar to the concert flute, are regularly used for the sound on assistance buttons on planes. In isolation they are dull, but when combined into a mass they can create rich and evolving sounds with coherence and structure. The simplicity of sine waves allows for thousands of them to be combined harmoniously, without needing to sound muddy or harsh, as they move gracefully like the murmurations of starlings.
A visualisation of a sine wave captured in the music programming language Supercollider
The algorithm that generates the rhythms and melodies in this composition uses a concept called the harmonic series, ostensibly discovered by the Ancient Greek philosopher Pythagoras. It describes how musical tones — like a plucked string — contain a whole series of harmonically related frequencies that vibrate in simple mathematical relation to one another. This means that a single musical tone contains a world of harmony within, which more than two millennia later remains a remarkable and beautiful fact.
As the myth goes, this was discovered by Pythagoras when he was passing by a blacksmith. He realised that the sounds of the metal hammers being struck sounded harmonious when their sizes were in simple whole-number ratios to one another (e.g., hammer a is double the size of hammer b). Inspired by this, he carried out a series of experiments using vibrating strings that confirmed his initial observation, which remains a fact of physics to this day and a powerful influence on music.
This composition borrows ideas from Ancient Greece and modern electronics while leaving out many Western musical ideas developed between these points — such as triadic harmony, the equal tempered scale, time signatures etc. It is as if one dropped a laptop into the hands of musicians in 500 BC!
A visualisation of a complex polyrhythm made in P5JS which influenced this composition. The visual world provided a map for a sound world.
Generative art in contrast to AI
When people talk about AI music they are generally referring to a popular approach that involves using Large Language Models (LLMs). This technique uses LLMs to find patterns in gigantic quantities of existing human-made music and then generates compositions from it. It is a method that leaves one unsure of how the music was created, and with a mountain of ethical questions about the inbuilt bias of the data sets, the exploitation of human creativity, and the environmental cost of using such massive computational power.
This is not to be dismissive of machine learning as a whole, which contains many fascinating developments. One example is the potential to facilitate interspecies communication by decoding animal vocalizations. Another is a promising open-source program called Tone Transfer which allows you to sing a melody into your computer and have your voice converted into the sound of an instrument.
By contrast, generative music is built on a fundamentally different foundation. Ant hills, the growth of neighbourhoods in metropolises, and even the development of the universe become objects of study and inspiration for generative art, exemplifying key concepts in the fields of emergence and complexity theory. With this philosophy, simple rules give rise to complex and often elegant systems, that are more resemblant to human intelligence than giant data sets. This topic was eloquently described by the American linguist and cognitive scientist Noam Chomsky in his critique of LLMs.
The human mind is not, like ChatGPT and its ilk, a lumbering statistical engine for pattern matching, gorging on hundreds of terabytes of data and extrapolating the most likely conversational response or most probable answer to a scientific question. On the contrary, the human mind is a surprisingly efficient and even elegant system that operates with small amounts of information; it seeks not to infer brute correlations among data points but to create explanations.
Unlike the ‘terabytes of data’ that Chomsky mentions, the program I built consisted of 9 lines of code, stored in 4KB of memory, running on a 3-year-old laptop. With a generative approach, a miniature program gives rise to a composition that is much larger than it.
Seeking the garden in the machine
The Tetons and the Snake River, black-and-white photograph by Ansel Adams
Often when an artist mentions having used a computer to help create art, the immediate assumption is that the computer did the work for them. This is not the case with generative music, which requires an artist to have an understanding of composition, a sense of aesthetics, and the ability to write code.
The computer program is in one sense irrelevant, or at least distracting. It's just a tool to compute a musical idea. In this case, the idea uses simple maths to animate thousands of notes to dance in harmonic and rhythmic relation to one another, giving rise to a vast space — or sound world — which can then be explored.
Since it is impossible to predict all the possibilities that will arise in the sound world, it is necessary to enter into it, listen deeply, and develop an intuition for its patterns and the emotional spectrum of possibilities. It is somewhat similar to how a photographer may walk through a landscape before taking a photo, inhabiting it in a way that provides a visceral understanding of the land, the interaction of light at different times of the day, and the changes in the seasons.
The same is fundamentally true in a digital space. The composer’s experience of the sound world forms the basis from which they can capture its essence, reducing its vastness into something finite and communicable. It is composition by reduction as much as production. When the American landscape photographer Ansel Adams gave up a promising career as a concert pianist to pursue photography his mother and aunt pleaded with him.
‘Do not give up the piano! The camera cannot express the human soul!’ To which he replied ‘The camera cannot, but the photographer can.’
The camera neither replicated nor replaced earlier forms of visual arts, such as painting and sculpture, but created an entirely new art form, with a unique set of techniques, aesthetics, and ideas. It provoked traditional visual art forms to change and differentiate themselves, thereby influencing all the fine arts, not just photography.
Something similar might be said today of code. Music composed in code will not replace embodied music. There will still be musicians performing on real instruments to live audiences hundreds of years from now. However, code does offer the potential for new techniques, aesthetics, and ideas to broaden our experience and understanding of reality, and will undoubtedly influence how embodied music is performed and composed.
Inspiration
Captured in infrared light by NASA’s new James Webb Space Telescope, this image reveals the “mountains” and “valleys” of a nearby, young, star-forming region in the Carina Nebula.
Captured in infrared light by NASA’s new James Webb Space Telescope, this image reveals the “mountains” and “valleys” of a nearby, young, star-forming region in the Carina Nebula.
I found the launch of the James Webb telescope in December 2022 deeply inspiring, not least of all when it returned an image of the Carina Nebula that showed the true magnificence of The Milky Way. While the Carina Nebula can be seen by the naked eye on a dark night, its full beauty would have remained hidden without the power of the telescope. It is a case of technology illuminating a corner of the universe.
The word nebula is apposite to this music, referring as it does to a cloud of gas and dust in outer space, which is somewhat akin to the mass or cloud of sine waves in the composition. Similarly, the stars in the image of the Carina Nebula connect with the starlike patterns that emerged from the image of the complex polyrhythms. The image of the Carina Nebula and a composition of 1000 sinewaves share something in common, they represent worlds outside of our everyday experience, that can only be brought to life with the help of technology.
I recently read the novel Dune (with space clearly on my mind) and came across the following quote.
‘What sense do we lack that we cannot see or hear another world all around us?’