I just launched a new project at radiomusic1956.com that allows me to realize, and visualize, John Cage’s Radio Music (1956) with software-defined radio. Below is some background on how this came to be.

Art built on indeterminacy and chance
I’ve always been drawn to art that is both systematic and indeterminate, where the artist creates a framework or ruleset that defines a range of possible outcomes. This aligns with my designer brain, which is drawn to patterns and repeatable structures more than single, perfected creations. It also aligns with my technologist brain, because even with something as analog as a Sol LeWitt drawing, there is the concept of a “program” that can be run: simple written instructions that result in something unexpected and beautiful. So of course, I’ve always been a fan of John Cage.

John Cage was an experimental composer, visual artist, and mycologist. If people only vaguely know his work, it’s usually for 4′33″, his piece that premiered with a performer sitting in silence at an unplayed piano for 4 minutes and 33 seconds — letting the environmental sounds in the concert hall become the music. In the hands of a less dedicated artist, this might have been just a stunt, but it fits into his body of avant-garde compositions that treat all sound as equally worthy of our attention. He subverted traditional instruments to use them in new ways, got people to really listen to their soundscape, and used chance operations to ensure no two performances would be identical.
One of Cage’s compositions that has always held a particular appeal for me is Radio Music — “to be performed as a solo or ensemble for 1 to 8 performers, each at one radio.” The radios, and the stations broadcasting at that moment, are the instruments. Like much of Cage’s work, the outcome is purposefully indeterminate. The score for Radio Music is a list of numbers indicating dial positions on an AM radio, ranging from 550 to 1560 kHz, along with dashes that indicate “silence.” Even those were left to chance, derived using Cage’s preferred method of consulting the I Ching. The resulting sound is a collage of airwaves at that specific time and place, spanning and scanning the spectrum across 56 different frequencies, with nearly 500 notated changes if performed with all 8 radios.
Software-defined radio
A little over two years ago, I first learned about software-defined radio (SDR), and it planted an idea in my head about realizing Radio Music in a new way.
You can think of SDR as “radio on your computer” — not streaming Internet radio, but real radio, the kind that is in the air all around you. We’re already used to the fact that many dedicated electronic devices are now built into our smartphones: cameras, GPS devices, calculators, voice recorders, flashlights, music players, alarm clocks. But aside from a handful of phones with a built-in FM chip, radio is generally not available on our computing devices.
While traditionally the domain of military and industrial applications, SDR was made accessible to hobbyists in 2012 with the availability of a device called RTL-SDR. This was the first low-cost ($30) SDR device for the consumer market. It’s the size of a large USB thumb drive, or a Roku TV stick, with a USB plug on one end and a coax connector for an antenna on the other. All interaction with it happens on the computer.

An SDR is not just for listening to AM or FM radio. Even my cheap RTL-SDR Blog V3 can access a frequency range from 500 kHz to 1.766 GHz, encompassing the MF, HF, VHF, and UHF bands. That means it can receive signals from airplanes, ships, weather satellites, emergency dispatchers, ham radio, shortwave, and more. When paired with the right antenna, it turns a computer into a sophisticated radio receiver.
All of this is very much still the domain of hardcore nerds. The software for SDR is clearly created by hobbyists, the terminology is arcane, and the learning curve beyond basic listening is pretty intimidating. But — how cool! — real radio on my computer!
So I got an SDR, with the goal of figuring out some way to realize John Cage’s Radio Music with it. Initially, I was stymied. I got it working just enough to do some parlor tricks, like reading ADS-B signals directly from airplanes flying overhead. But in terms of Radio Music, it didn’t seem to offer that much more than an old analog radio. In fact, it would be much more complicated and expensive to get eight SDRs hooked up to eight different computers, so I shelved the idea.
Last month, I dug back into it and discovered an SDR capability I had previously overlooked: the ability to capture radio spectrum as raw I/Q data. This was the unlock! It acts sort of like a RAW file does for photos, where you can process the image data later to create new photos. When you capture the full spectrum, you can demodulate (or “tune”) the radio after the fact. So you’re not recording a particular station; you’re recreating the airwaves as they were at that time and place.
Realizing Radio Music
Which leads us to today — I’m excited to share radiomusic1956.com, which realizes Radio Music with software-defined radio as an ensemble of 8 digital performers. The audio comes from a capture of the full AM radio spectrum, which is then cloned into 8 virtual radios that can perform Cage’s score as if they were sitting under the same sky.
There are two views you can watch while you listen to a performance. The stage view shows each individual radio being tuned, as it would appear on a stage, which also aligns with the left/right spatial mix of the audio (best heard through headphones). Or you can choose the spectrum view, which shows all 8 radios as they move along a single radio spectrum.


The site closely follows Cage’s original score, which leaves the duration of each tuning to the player. My project makes those decisions for each performance through chance operations, modeled on the same I Ching process that Cage used. As the piece plays, you can follow along on a visualization of the score, which shows the playhead moving between all of the frequencies Cage specified and across color fields that indicate the duration of sound, silence, and the gaps between sections.

I will be adding new performances over time, using my newly acquired MLA-30+ loop antenna to capture spectrum in Chicago, and in other locations if I bring it along on trips. Because the performances are built from full-spectrum recordings, I can also use captures made by other people, sometimes from many years ago. The site already includes some of these, and I’ll be adding more as I find historical recordings that meet the necessary technical and licensing requirements.
I’ve done my best to make sure that the site works well on both mobile phones and desktop computers, but the larger-screen experience is better — particularly for the score visualization. I’ve written more about the project on the About page of the site, which includes additional details and explanatory visuals. If you want to get updates about the project, there is an RSS feed to follow and a low-frequency email list.
I hope you check it out, and I’d encourage you to both click around and listen to a full 6-minute performance from the beginning. It’s not “music” in the sense you might traditionally think of it, but there is a quality of immersion to it, and the airwaves’ content acts as an interesting marker of a particular place and time. You will hear fragments of news, talk, music, and (of course) static. The simple rules that program each realization create something unexpected and unique.
With 4′33″, Cage got people to notice the sounds already present in their environment. With Radio Music, he helps us to think about — and briefly hear — the invisible spectrum of sounds that are always there in the air around us.













































