Resynth

Concept

Resynth is an audiovisual installation that transforms African political speech into moving light and synthesized sound. Each word spoken is analyzed by artificial intelligence, translated into emotional signals, and expressed through shifting shape, color, and tone.

It's not just a tool. It's an interpreter, one that reveals what's often hidden beneath rhetoric: emotion.

Why We Built This

Political speeches are performances. They persuade, provoke, and promise. But behind every word lies a feeling, fear, joy, anger, or hope. In a world saturated with noise, Resynth invites you to listen differently. Not for what's said, but how it's felt.

We wanted to build something that makes those emotions visible, audible, and alive.

How It Works

Step 2. Emotion detection with AI

The text is analyzed line by line using a machine learning model, DistilRoBERTa, classifying each line into one of seven emotions.

JoyAngerSadnessFearDisgustSurpriseNeutral

Step 3. Emotional data becomes light and sound

Each emotion is tied to a unique color, a distinct shape behavior, twitching, drooping, spiraling, and a synthesized sound built in Tone.js, softer tones for sadness, sharp spikes for anger. The system reacts in real time, blending states into transitions, the way a person might feel more than one emotion at once.

The Interface

The UI stays out of the way on purpose. A dark stage, one shape, a line of text at a time, and a small panel for volume and tone. Nothing competes with the speech or the light responding to it, the interface is there to disappear.

Why African Voices

We center African political speech to make space for emotional expression often overlooked in Western media and data analysis. These speeches are historically layered, shaped by independence, protest, identity, and power.

With Resynth, we amplify those emotions, not to judge them, but to make them felt.

Seeing It React

Live screenshots from the deployed experience, same speech, four different emotional moments. Try it yourself