Reflection Pool: A Tangible Interface for Emotional Expression

In a digital world where memories are often flattened behind glass screens, the Reflection Pool restores the physicality and weight of an intention.

• The Artifact: Each coin serves as a physical container for a specific memory.

• The Ritual: The act of "tossing" the coin into the pool acts as a symbolic release, transforming a private thought into a shared environmental expression.

• The Medium: Water acts as a natural amplifier, using ripples and light reflections to make the digital processing of emotion visible and felt.

The user journey is a three-stage process that blends physical interaction with computational feedback:

1. Programming (Speak): When a user picks up a coin, a capacitive sensor triggers the USB Microphone. The user speaks a memory; the system uses a word-bank model to classify the emotion into categories like Joy, Sadness, or Reflection.

2. Release (Toss): As the coin hits the water, the NFC reader identifies the specific coin and signals the Raspberry Pi.

3. Reflection (Feel): The pool responds with a multimodal display:

   • Light: LED rings cast color-coded auras based on the sentiment.

   • Sound: Ambient soundscapes play through integrated speakers.

   • Vibration: Motors create physical ripples, making the digital memory tangible.

Offline Privacy: To encourage vulnerability, we implemented Vosk for local speech-to-text. This ensures that user memories are never uploaded to the cloud, maintaining absolute privacy for the participant.

Tangible "Cost": We found that the physical effort of picking up and tossing a coin makes the interaction more meaningful than a digital click. The "weight" of the interaction matches the emotional weight of the memory.

Robust Scaling: By using a SQLite database on the Pi, we enabled the pool to "remember" coins over time, allowing for a shared, evolving environment of collective memories.

This project was a collaborative effort at the MIT Media Lab involving engineers and designers from Harvard GSD, Harvard GSE, MIT, and MassArt.