Discovering a New Sound World
The brief: Create a device that makes use of a micro-controller platform, sensing, actuation, and digital fabrication. Deliverables: a physical prototype, a demo video (shown below), a poster, and a live demo presentation.
This mini-project is an exploratory design for the MSc that features extensive use of physical prototyping.
I took the opportunity to indulge in my love for noise music, selling it to a layman audience using the metaphor of discovering a parallel dimension through short wave radio (made famous in the Netflix show Stranger Things and in the Conet Project).
The idea was framed thus: Shortwave radio hobbyists enjoy discovering new and weird shortwave stations from across the globe - an activity known an DX-ing. BoxDX aims to bring that same sense of discovery to the user's surroundings - a noisy cafe, a bustling street, or your neighbourhood.
Note: live audio distortion is not a novel idea. However, audio distortion that can be transported everywhere, and processed on the spot, could have interesting applications in entertainment or relaxation. This meant that BoxDX's portability became extremely important, so that it would passively distort all environmental sound surrounding the user when it is switched on - to provide a surreal, alien experience of the world.
Demo video. Showing the construction, assembly and the engineering behind BoxDX.
Sketching the Design
Design started with sketching how I would envision the product to work and look like. With my limited knowledge at the time, I only knew that I would need an audio signal input, signal distortion with the micro-controller, and an audio signal output.
I knew the final product would need to be portable, to collect sound from a wide variety of environments, and to process it live to create an 'alien but familiar' representation of reality.
Eventually, this simple picture became more complicated, as I experienced more technical challenges and learnt how to overcome them.
Basic sketch. Simple schematic of how BoxDX would work.
Form factors. Multiple sketches of possible form factors as a way of creating a design space.
Logbook (excerpt). Personal tracking of needs and issues over the project.
Hardware
1) Sound Input
The challenge here was to find a microphone that would be able to gather sound from an entire room. Most breakout mics are either only good to detect sound, or will only pick up sound that is close to them. I settled on the Adafruit MAX 4466 as it had a powerful onboard amplifier that picked up sound from a long distance away.
2) Variable Distortion
To provide the user an additional way to vary the sound, I included a variable low-pass filter to create distortion. I chose the low-pass filter as its construction relies on very little space (pot + capacitor + wires), and has the potential to completely change the sound.
3) Sound Processing Code
Converting the analog signal (input) to digital was simple as the Arduino Uno has an onboard ADC. Outputting the post-processed digital signal (see below section on Digital Signal Processing for details) was a challenge as the Uno only outputs in digital. Eventually, I decided to buy and assemble an Adafruit Wave Shield as it was the most time and cost efficient option.
BoxDX schematic. Created with Fritzing. Red lines refer to live; blue lines to ground; and green lines for audio signal.
Digital Signal Processing
Before even looking into coding for Arduino, I experimented with Audacity to distort various sound recordings I had made in London.
Through much experimentation, I settled on 2 possibilities to achieve the "alien but familiar" sound I was looking for: 1) Reverb and 2) Downwards pitch shifting. Subsequently, I needed to decide on just one effect as the Uno has the power to handle only one sound effect at a time.
Further qualitative testing amongst coursemates helped me to decide on the downwards pitch shift effect; the reverb effect was felt to be too harsh and jarring for the ear.
Field recordings. Audio experiments using Audacity.
Enclosure
With the prototype working, I worked on deciding the appropriate enclosure for the device. The enclosure needed to be portable and to provide additional distortion. Since I had decided on not including a hole for the mic (for maximum distortion), the size and material of the enclosure became very important.
I tested a wide range of materials and linings with course mates, including acrylic boxes, wooden make-up drawers, plastic cups, ceramic bowls, cardboard boxes, foam and bubble wrap. Key criteria was durability, portability and most importantly, acoustics.
Eventually, I settled on using wood, with some extra internal space to allow for a small amount of reverb (16.6cm X 11.6cm X 7.2cm). This had the most pleasant acoustics, was durable, and was small enough to carry around. I created the box using a laser cutter.
Choosing materials. Different boxes and materials tested for the enclosure.
Putting it together; Presentation
All components were soldered together and glued inside the plywood box to be run on a 9V battery. Further testing was done in different environments - the street, a cafe, a workshop, and a garden, with manipulation of the low pass filter - to stress the box and to explore the different sound outputs.
Finally, BoxDX was presented successfully in a live demo for the class.
Internal view. A shot of the internal wiring of BoxDX.
Live Demo. Me posing with BoxDX. In the background, the poster used to accompany the showcase.
Future Steps
Ideally, I would like to further refine the prototype to a smaller form factor. Considering BoxDX is meant to be used outdoors as the user walks around, improved ergonomics has to become the main priority.
This could be done by experimenting with printing the circuits so that less space is used. Additionally, sound processing could either be moved to code or a separate module, creating possibilities for more complex forms of distortion.
Future sketches. Potential directions for BoxDX.