Sat Nov 1, 2025
An art project commisioned in 2025 by Fondazione Il Lazzaretto in Milan, Italy for Il Festival della Peste.
Does our voice change depending on the topic of conversation? If so, by how much? Could we see the difference between various topics with our own eyes if we converted our speech into abstract drawings? Not because of the words we use, but because of the emotions connected to different topics?
I had these questions in mind while developing this interactive installation, which invites visitors to contribute voice recordings related to the festival topics “Power” and “Pleasure.” The installation uses a pen plotter to transform these recordings into abstract designs on paper.
The installation runs on three computers. Two of them feature a microphone and a small display that invites participants to talk. I wanted the interaction to be limited to the sound of the human voice, without using a mouse, a keyboard, or any speech recognition software. I came up with the idea of having participants produce a long enough “shhhhhhh” sound to move forward. My thinking was that most people would find it easy and that asking them to start with a simple sound might help “break the ice” (talking to a computer can be awkward).
A computer waiting for a participant.
An amusing side effect occurred: people would start excited conversations around the pen plotter, and suddenly someone in the back would say “shhhhhh!”. Everyone would then quiet down for a moment, wait, and start laughing after realizing it wasn’t meant for them.
I analyzed the voice recordings left by participants using Praat, a software widely used for speech analysis in various fields of linguistics. I observed five changing sound properties: the intensity, the first formant frequency and its intensity, and the first pitch candidate and its intensity. The average loudness of the recording was also measured.
A third computer, connected to the pen plotter, received the sound analysis values and used them to create a unique, looping, self-intersecting contour. Then, I applied a chordal axis algorithm to generate the complex shape that gets drawn on paper.
Participants observing the pen plotter working. Photo by Silvia Gottardi, edited by me.
During certain periods, festival attendees were busy with workshops and did not make any new recordings. I programmed the computer to indicate these “quiet times” by drawing a small bubble every ten seconds. It was as if the computer were bored or lonely. At other times, a “drawing queue” formed because people recorded voices faster than the machine could draw them. In this situation, a large display provided instant feedback, showing a more literal visualization of each recording in two colors: green for recordings related to power and red for those related to pleasure.
This project consists of approximately 4,000 lines of code (written by me) and uses various free, open-source tools, including OPENRNDR, the Processing Geometry Suite, Minim, SoX, Denoiseit, Praat, rsync, bash, and Manjaro Linux.
Each shape features a time-stamp and a unique ID.
Over the course of five days, nearly two hundred recordings were collected and transformed into a collective, abstract design spanning approximately five meters.
I was pleasantly surprised by people’s curiosity and the abundant feedback. The pen plotter fascinated them with its mesmerizing movements and sounds, while they eagerly awaited seeing their own contributions appear on paper.
Someone told me that this work had helped them realize the importance of not only what we say, but also how we say it. I’m taking that comment home with me.
A large display showing voice recordings. Red shapes are about pleasure, green ones about power.