Transfo Intiem 2026
Bachelor thesis · Devine, Howest

Holding Our Breath

12.12.2026
17:00 – 22:00
Zwevegem, BE

From tech research to Firefly Forest: our first development steps

What could three people realistically build in six weeks? That was our starting point. Before choosing a concept, we began exploring the technology, testing our assumptions and working out what would be feasible within our time and budget.

In short

  • We mapped the technology first: sensors, controllers, software and outputs.
  • We turned that into four tech directions, then combined two of them into Firefly Forest.
  • The biggest change: visitors touch the flowers instead of blowing into a sensor.
  • Next: build and test one flower before scaling up.

What can we build?

On 29 October, we put together an inventory of possible technologies: sensors for detecting people and breath, controllers, software, and outputs such as LEDs and projection. We compared cost, reliability and suitability for five hours of use by families in a cold hall.

A few things became clear early on:

  • No microphones. Detecting someone blowing would require them to be only a few centimetres away, while the hall’s strong echo made audio interaction less promising.
  • One shared foundation. An ESP32 microcontroller to read sensor inputs, TouchDesigner on a laptop to coordinate the experience, and LEDs as an output could support several concepts. This meant we could start learning the tools before settling on a concept.
  • Free licence. The free TouchDesigner licence appeared sufficient for our needs, although we still needed to confirm this with the school.
  • First budget. Around €400 for a minimal setup to €1,300 for a more extensive version. Renting a projector could add another €600 to €1,500.

Checking feasibility

We then assessed whether both assignment options were achievable. Our conclusion was that either could work, provided we kept the interaction simple: one input and one output.

At this stage, the Pompenzaal seemed the safer option. It offered an indoor, dry environment, and the build would not depend on another team delivering materials on time.


Four technical directions

For our pitch to the other groups, we translated the research into four possible directions:

01

Projection mapping
Visitors’ breath becomes moving imagery on the walls.

03

Light and sound
Lights appear to breathe in time with a soundscape.

02

LED light
Each breath lights up LED strips inside pipes and tubes.

04

Interactive projection
A depth sensor detects where visitors move, and the projection responds.

These were starting points rather than fixed proposals: different ways to explore how visitors could influence the space around them.


How the research became Firefly Forest

Once we chose the theme “breath as nature,” we combined elements from several directions rather than selecting just one.

  • From interactive projection we kept the depth sensor and projected imagery, allowing fireflies to gather around visitors as they move through the installation.
  • From the LED direction we kept the ESP32 controllers and lights, placing them inside flowers.

The biggest change was the input. Instead of asking visitors to blow into a sensor, we shifted to touch.

A small vibration switch inside each flower head detects movement when someone touches it and the stem sways.

This removed the need for shared breath sensors, a concern during flu season, and gave us an interaction intended to feel intuitive across age groups. Whether it works without explanation is something we still need to test.


Our current setup

The planned installation brings together physical flowers and a projected forest:

PartWhat it does
20 flowersEach with four warm LEDs and a vibration switch, mounted on a flexible fibreglass stem.
3 ESP32 boardsEach supports up to seven flowers and controls the breathing light and the ripples between neighbouring flowers.
Laptop with TouchDesignerReceives signals from the flowers and controls the forest projection.
7 short throw projectorsProject the forest onto the wall and a scrim along our curtain.
Depth sensorLets the projected fireflies respond to visitors’ movements.
Light haze (optional)Only with approval from Transfo.

Our current estimate for the flower and firefly components is around €160 of our €300 budget. This excludes the projectors, laptop and depth sensor, which we hope to borrow.

The plan has grown beyond our initial “one input, one output” approach. That makes prototyping especially important: we need to establish what works reliably before committing to the full installation.


How we used AI

We used Claude to help assemble the initial technology inventory and feasibility check, including comparisons of sensors, controllers and approximate costs. We treated its output as a starting point for investigation, not as a verified technical plan.

Next steps

Our next priority is to test the core interaction before scaling up:

  1. Build one test flower, connecting its LEDs and vibration switch to an ESP32 and then to TouchDesigner.
  2. Test the depth sensor in TouchDesigner.
  3. Confirm the TouchDesigner licence requirements with the school.
  4. Ask Transfo about power, projector availability and permission to use haze.
  5. Prepare the paper prototype for our first test on 20 November.

The first week gave us a technical starting point. The next step is to find out how it behaves in practice and whether a single flower can deliver the interaction we imagine for the whole forest.