Dutch artist Maarten Heijnens worked, together with sound designer Wouter Snoei on “Parallel Realities”: an installation that gives one person at a time the opportunity to enter worlds through different sound portraits. Worlds in which his or her presence is not so obvious. With a console mounted on a soundproof booth, a soundscape is chosen that can be experienced in surround sound. The installation had its premiere in May 2025 at the “Moet je horen” festival, a festival centered on listening at the Theater aan het Spui in The Hague. Later, the installation will tour along museums and festivals.
My assignment was to design the “doorbells” board that could be used to select the sound portraits. This modular board should be able to communicate with the already realized technology of this art object via a serial interface.
Technically, the installation consists of a soundproof booth for one person. Six speakers and a subwoofer are installed in this booth. The listener is transported by audio to another place with the different sound portraits. Depending on the choice, these sound portraits can be real or fictional but always run synchronously with the current time. In addition, a live stream of very high quality is available.
All audio was recorded with so-called “spatial” Ambisonics microphones. The soundscape of choice is selected on a console on the outside of the listening room. Lights in the various selection buttons of this panel provide feedback on the different streams and make it clear which 'audio world' the listener is in.
When Maarten and Wouter contacted me, the “audio landscapes” and technology were already at an advanced stage. The audio part was already set up on a Mac Mini in the SuperCollider program. The live stream was already running with a Zoom H3VR recorder on location linked to a Raspberry Pi communicating via SonoBus.
Construction of the booth had already started, but there was not yet a, preferably splash-proof, “doorbell” style board for selecting the sound portraits to be added. Only less more than one month of time was left for its development and realization of this because the premiere was already scheduled.
There were several possibilities for the board to communicate with the underlying software, each with its own advantages and disadvantages. In part, these depended on the choice of microcontroller to be used. For example, a Raspberry Pi seemed like overkill for functionality, but the serial communication over the USB bus would be relatively simple. However, when choosing a simpler microcontroller (such as an Arduino) rather than a complete single board computer, this in turn would be more challenging.
Universal MIDI was eventually chosen for serial communication to best match the hardware already used for the sound booth prototype. Although MIDI (v1) is not a tremendously fast bus (31.25 KHz), in practice it turns out to be too fast to reliably emulate in software on a 16 MHz microcontroller. For this reason, the choice fell on an Arduino Mega because of its extensive (hardware-based) bus and interface capabilities. This microcontroller also has sufficient digital inputs for the buttons and pulse-width modulated (PWM) output ports for controlling the individual light intensities of the lighting in the eleven push buttons.
Development time for the “Bell Board” was limited. Fortunately, the necessary parts to build a prototype were available in my electronics workshop and I could begin development immediately. To investigate the need for a hardware MIDI bus, I tested the reliability of a MIDI interface emulated in software. Then I focused on making the MIDI inputs and outputs "electronically robust".
Fortunately, the relatively 'slow' optocoupler (4N25) in stock was able to suffice with an electronic trick during the ordering time of the final component (6N138) in the prototype, so further developments were not delayed.
In its simplest form, the board would only have to report which button was pressed. The board itself would then have to take care of the lighting (and extinguishing again when the button is pressed again) of the corresponding button. It would of course be nicer and more flexible if the lights could be controlled externally. And it would be even nicer and more useful if the brightness could also be controlled completely via the MIDI input.
Despite the approaching deadline, it was decided that we would not compromise on the technical possibilities: the eleven ring LEDs in the buttons of the “Bell Board” should become fully controllable. To investigate the need for eleven PWM outputs, this led to an experiment with a PCA9685 board (16-channel, 12-bit PWM Fm+ I2C bus LED controller). With this module it is possible to provide even a simpler microcontroller (such as, for example, an Arduino Nano) with sufficient controllable (PWM) outputs.
During the experiment with the PCA9685, it did not become sufficiently clear whether the I2C bus necessary for this would be able to process the stream of commands for the eleven lighting LEDs simultaneously. For this reason, but also because the “Bell Board” should eventually be equipped with fast MIDI inputs and outputs as well as having sufficient control outputs for the lighting in the buttons, the choice for an Arduino Mega became increasingly clear.
The next step was to build the final board. Since the necessary additional parts had arrived in the meantime, the MIDI interface could also be built on this directly in its final form. I designed a PCB that was a modular housing for the whole thing on which the various connectors (MIDI, DC 9 volts) and an on/off switch could be housed.
By now, the pushbuttons had also arrived and it became clear that they all still had to be individually provided with a driver stage in any case, since they required more current (>20mA) than can be supplied directly from the microcontroller outputs. Since I wanted to give the LEDs a common ground (just like the pushbuttons at the inputs), I built eleven simple PNP-transistor drivers (with a BC557) for each LED output.
The eleven (in the photo quite dusty, my apologies) LEDs came in handy especially before the final buttons were installed. They make it possible to keep an eye on the outputs from the back of the panel during the experiments. Upon commissioning the panel, this readout could be deactivated by moving the yellow contact jumper.
Meanwhile, for further firmware development, it was thus possible to gradually switch from the circuit built on a breadboard to this, more final, circuit.
After the electronics, microcontroller and programming phase, the actual construction of the board could finally begin. The final wooden housing could be designed and realized. For the text plates I designed and printed card holders that could be mounted on the front of the console. Once the installation of the selector buttons was done, the whole thing could be painted and fitted with aluminum frame profiles so that it would be optimally visually aligned with the “industrial look” of the sound booth that was now taking shape in Maarten's studio.
After the “Bell Board” was united with the technology and booth, it was extensively tested by Wouter. In the meantime, the idea had arisen to use the LEDs as VU meters of the sound portraits that were not active at that time. So the decision to stick to the requirement of being able to control each LEDs individual brightness certainly was a good one.
Fortunately, no necessary adjustments to the software came to light during testing and the button module could smoothly be connected to the cabin by Maarten.
Meanwhile, the premiere of “Parallel Realities” was successful and the installation passed the practical test. During the summer months of 2025, the booth will travel with De Parade and may also come to your neighborhood. Experiencing this special sound booth is definitely worth recommending!