Machine Arts Magazine - Online Journal exclusive content.
Peter Beyls and the Art of the Unpredictable Machine
For more than fifty years, Peter Beyls has moved between electronic music, computer drawing, artificial intelligence, instrument building, robotics and painting. The machines have changed radically. The questions have been remarkably persistent.
I sat down to talk with Beyls a week before the opening of his show, Peter Beyls: New Works, at the Bantam Tools Machine Arts Gallery. Beyls creates worlds. His programs aren’t tools for making pictures so much as ways of questioning the systems and principles we live by. In conversation, you might go deep on the materiality of graphite in pencils, or find yourself in philosophical territory: relationships, beginnings, and what it means to be a living creature with the ability to think and imagine.
When I asked about the early days of computer art—his first encounters with computers, plotting and programming—he corrected the premise almost immediately.
“It all started with electronic music, really.”
That turns out to be an important way into Beyls’s work. Long before a computer could sit on an artist’s desk, Beyls was already thinking about machines as collaborators: building electronic instruments, generating musical structures, manipulating signals and creating systems whose behavior could surprise him. Visual art and music did not develop as two separate careers. They grew together.
Beyls grew up in the Belgian countryside with little access to the emerging world of electronic and computer art. He describes discovering it piecemeal: journals, articles, correspondence and, eventually, people. An IBM publication from France led him toward experimental work happening in Paris, and in 1972 he moved there. Around the same period, he began working with computers at Ghent University.
His first computer drawing was produced on an IBM 370 mainframe using punched cards.
The experience bears almost no resemblance to the immediate feedback artists expect from computers today. A program had to be submitted, processed and debugged through an extremely slow cycle. Beyls remembers that producing a single drawing could take months.
Later he gained access to a scientific computing center equipped with PDP-15 and PDP-11 minicomputers and an analog computer. He worked there at night, from roughly nine in the evening until four in the morning. The inconvenience does not seem to have registered as deprivation. He was simply thrilled to have access to the machines.
This is one of the recurring characteristics of Beyls’s story. The computer is never an invisible appliance. He remembers the machines: PDP-15. IBM 370. Commodore PET. Atari. Symbolics Lisp machines. Each created a particular environment for thinking, and when the machines changed, he changed with them.
The Early Plotter Years
Some of Beyls’s earliest plotter drawings emerged from extraordinarily restricted circumstances. At one computer center, the available plotter initially had only blue pens. Rather than treating that as a serious obstacle, Beyls worked within it, developing grid-based drawings in which change could happen inside a constrained visual structure.
The grid became what he describes as a kind of receptive playground. What mattered was not simply the values inside the grid but the relationships between them—the intervals and changes. He makes the analogy to music: an interval can matter more than the individual notes.
The technological limitations were severe, but Beyls says he barely cared. The important thing was access.
Just as important was discovering that he was not alone.
One formative moment came when he encountered Douglas Davis’s Art and the Future. Beyls remembers recognizing himself in the history of electronic art described in the book: “This is me. This is where I belong. This is my family.”
That discovery became literal. He began traveling to meet other people working at the boundaries of computation, electronic music and art. London led to connections at the Slade School of Art. Stockholm offered access to an important computer-music community. In Toronto, his friend Bill Buxton invited him to the Computer Media Lab. Other relationships carried him through Brussels, Utrecht, California, Chicago and elsewhere.
In the 1970s, if you wanted to use a particular computer, you often had to physically travel to it.
Beyls did.
By 1981, he was not merely participating in this international community but helping create it. He organized a large festival of electronic music, video and computer art across Brussels, inviting artists, composers and performers from around the world.
Looking back, one gets the sense that the network of people mattered nearly as much as the network of machines.
Before the Drawings, There Was Music
Beyls’s first musical experience was playing in a local brass band, but his interest quickly shifted to building custom musical electronics. The instruments and the process were as engaging as the music itself.
As a teenager he became fascinated with ham radio. He remembers visiting a nearby radio enthusiast whose room was filled with massive American transmitters and shortwave equipment. Soon Beyls was experimenting himself, including—with some amusement in the retelling—a little unlicensed transmitting.
Then came homemade electronic instruments built into cigar boxes, first with transistors and later with early operational amplifiers.
Between 1973 and 1975, in close partnership with Michel Waisvisz at STEIM (Studio for Electro-Instrumental Music, a center for new musical instruments) in Amsterdam, Beyls designed and built about a dozen early prototypes of the Cracklebox, a small noise-making instrument with exposed contact pads. The instrument let the performer touch the electronics directly, using the electrical capacitance of the human body to bridge and alter the sound-generating circuit. None of the prototypes survive.
Beyls pushed the idea further in his own instruments. He built larger crackleboxes with multiple audio outputs and constructed a floor platform he could perform on barefoot, turning bodily movement and contact directly into electronic sound.
The distinction between instrument, performer and system was already becoming blurry.
He also modified televisions, disconnecting their normal deflection systems and driving them with external signals so that they behaved almost like oscilloscopes. Later he extended those experiments with an analog computer.
These experiments contain an idea that recurs throughout Beyls’s career: rather than fully determine an outcome, build a system with a character of its own and then interact with it. That exploratory mindset has led him through nearly every generation of computing technology, and into friendships with many of its innovators.
In the mid-1970s, music and plotting literally became the same project.
Beyls developed a computer-generated composition for soprano voice, live electronics and generated imagery, after realizing that the same kinds of algorithms could shape both visual and sonic structures. The score itself was plotted—the plotter became the device that materialized the computer’s composition. The work premiered at ARC 2, a contemporary art museum in Paris, in 1976.
Because bringing a computer onto a stage was essentially impractical, Beyls built hardware to do what he wanted computationally. His HSP (Hybrid Signal Processing) machine combined analog and digital electronics, filters, TTL logic chips and a patchable structure. It analyzed a singer’s voice and fed characteristics of that signal into a network that could respond in real time.
Beyls describes the underlying method simply: he would create a machine and then push it to discover its “behavioral scope.”
At first those machines were physical. Eventually they became software.
That transition may be one of the simplest ways of understanding the continuity of his work.
An Unpredictable but Coherent Machine
By the 1980s, Beyls had moved deeply into artificial intelligence and computational systems.
In 1984 he developed an early expert system for drawing: instead of programming a finished composition, he built collections of simple rules that could interact with each other. A single rule might do almost nothing. But many local interactions could produce global complexity that had never been explicitly designed.
For Beyls, the interesting thing was precisely that the result could not be completely anticipated. He later described this approach as speculative computing: creating a computational environment capable of accommodating ambiguity, uncertainty and impulsive intervention from the artist. Programming, in this view, becomes “an act of aesthetic introspection.”
Randomness plays a role, but not in the familiar sense of making things simply look random. Beyls compares randomness to fuel. “I use randomness to activate my machines,” he explains, “to give them momentum and discover what they are capable of doing.”
“I’m more interested in behavior than structure,” he says. The complex drawing is not the goal; it is evidence. The structures left on the page are “residue or trajectories” left behind by a behaving system. And Beyls draws another distinction that may be even more important: “I want to create wonder rather than impression.”
“I want to create wonder rather than impression.”
Music as a Parallel Laboratory
Music never disappears from this story.
Beyls used L-systems—rule-based growth models originally developed to simulate plant development—to generate melodies, and later applied genetic algorithms (techniques that evolve and mutate rules over time, similar to natural selection) to manipulate those rewriting rules. He applied the same algorithms to drawings, collapsing the distinction between a system for visual composition and one for musical composition.
At the VUB Artificial Intelligence Laboratory in Brussels, where Beyls worked as a researcher, he also helped develop new electronic instruments. One was effectively a computerized violin equipped with infrared sensing and touch-sensitive controls, sending signals through an Atari personal computer to a Yamaha TX816 synthesizer system. He performed extensively with it.
For a 1980 live computer-music performance of The Hollow Man in Brussels, Beyls wrote software in 6502 assembly for a Commodore PET, controlling a synthesizer through custom-built electronic hardware. He describes it as one of the very early examples of live computer music.
Later, for his PhD in computer science at the University of Plymouth (2010), the problem became even more explicitly philosophical. He imagined a human performer improvising with a machine performer possessing two competing motivations: the desire to express itself and the desire to integrate with the human musician.
The machine listened. It adapted. It developed musical responses. Its motivations shifted in real time. The ideal, Beyls says, was a system that was unpredictable but coherent—“like a human being.” That phrase could describe a surprising amount of his work.
Programming Across Fifty Years
Talking with Beyls about his work inevitably becomes a lesson in the history of programming environments. He has rarely been loyal to a particular language or machine. Instead, the tools change around the idea.
When he joined the Artificial Intelligence Laboratory in Brussels in 1985, he encountered Symbolics Lisp machines—computers built specifically to run Lisp, a pioneering language for artificial intelligence and symbolic reasoning. He remembers the initial difficulty and then the revelation once the logic of the language clicked. Later came Macintosh Common Lisp, which brought that environment to personal computers and which he remembers with particular fondness. When older software environments disappeared or operating systems moved on, he reimplemented his systems in new tools such as SuperCollider, a language for real-time audio synthesis and algorithmic composition. Current projects often involve Python.
Old ideas also return when computers finally become capable of doing what Beyls originally imagined.
His archive is therefore less a sequence of obsolete projects than a reservoir of ideas waiting for technology to catch up with them. Algorithms, systems and experiments can lie dormant for decades and become active again when a new technological environment makes them possible.
That archive—spanning dozens of academic papers, books and musical recordings—is now entering a new chapter. With support from CVK, funded by the Flemish government, an effort is underway to consolidate, catalog and document it.
Beyls describes the accumulation of these systems as a kind of personal belief system: algorithms and computational paradigms developed over decades that connect to one another, change, collide with history and gradually form something larger than any individual program.
Increasingly, that belief system has moved from bits into atoms.
From Systems to Material
That transition is especially visible in the work Beyls is showing now.
His show at the Bantam Tools Machine Arts Gallery, which opens October 2, includes a project that begins with the image stream of The New York Times. Beyls has collected news images over time and applied face-recognition systems to them, building a database of more than 10,000 detected faces.
Importantly, he is not trying to perfect the recognition. The mistakes interest him.
Faces detected at very low resolution, incomplete faces, unusual expressions and outright failures of the algorithm become raw material. Other programs extract contours and regions and translate them into drawings. Beyls searches the resulting collection for anomalies—things that “stick out” or do not quite belong—and then looks for families among them.
The new work is also where code meets physical matter. After decades of building purely electronic systems, Beyls is exploring what happens when those instructions meet paint, pigment, brushes and the hand.
Some works receive additional color by hand. Different brushes and paints behave differently. A computational instruction becomes only one participant in a physical event.
His workshop reflects the same mix: historic electronic instruments sit beside robotic arms and desktop drawing machines.
Another body of work in the exhibition starts with essentially nothing: an empty field. An algorithm establishes a zone, fills it with hatching, then attempts to place another adjacent zone whose hatching must relate to the first. The process continues until the available space has been occupied.
One series begins with an enormous found system—the constantly changing world of news imagery. The other begins with the blank page.
Both ultimately ask similar questions.
What can a system become?
What happens when it encounters resistance?
How much control should an artist surrender?
And at what point does information become experience?
For Beyls, the answer is never contained entirely in the machine, the algorithm or the resulting object. Meaning happens somewhere in the exchange between them—and between the work and the person standing in front of it. Art, in his view, is never a solitary endeavor but a dialogue sustained by community.
“It all started with electronic music,” he told me. Fifty years later, it hasn’t ended. Every machine he builds is still an instrument, waiting to be played and to play back. Peter Beyls is a world builder of interconnected patterns and systems. Looking back at his prolific career, its history sparkles with innovation, but because his motor of innovation and engagement is still running so strong, I find myself wondering with excitement, “What’s next?” ■
