At the Hirshhorn during Sound Scene X.

At the Hirshhorn during Sound Scene X.

At the Hirshhorn during Sound Scene X.

At the Hirshhorn during Sound Scene X.

What is it?

This interactive “radio” has eight stations of computer-generated speech, presented in a faux old-time cathedral radio cabinet with a tuning knob. Each station is a Markov chain generated from texts including Alice in Wonderland, the U.S. Constitution, and the tweet history of Donald Trump. There are 24 hours of looping audio for each station, pegged to the current time.

Have a listen:

Where has it been shown?

The Markov Radio was exhibited at the Hirshhorn Museum as part of Sound Scene X, a one-day show on July 8, 2017, curated by the D.C. Listening Lounge. About 10,000 people visited the museum that day.

DCist, 7/7/2017

Newsroom of the Smithsonian, 6/13/17

What problem does it solve?

The Markov Radio is a deep dive into cognitive dissonance. It has a human-seeming voice reading sensical-sounding text… But if you really listen, it’s all just mash.

What was my role in it?

I designed and created the Markov Radio.

How does it work?

It is driven by a Linux computer (Next Thing Co. C.H.I.P.) and microcontroller (AVR: Pro Mini) running custom software in a bespoke laser-cut wooden enclosure.

Backstory

The Markov Radio is still being constructed; this is a working prototype. It is driven by a Linux computer (Next Thing Co. C.H.I.P.) and microcontroller (AVR: Pro Mini) running custom software in a bespoke laser-cut wooden enclosure.

The radio is about 8x6x4”. It can be positioned on a tabletop and can be heard from a distance. It does not have an external speaker hookup so it would be hard to hear outdoors or in a loud/large space.

The generated text from Donald Trump’s tweet history may be offensive to listeners.

It is audible by many but adjustable by one.

How was it made?

Communication

Fabrication

Hardware

Roles

Software