TARASOUND is a generative music system that listens to the world.

It continuously gathers real-time data from global monitoring networks and translates that data into music as a genuine attempt to make the state of the world audible. The music you hear is not composed or pre-recorded. It is calculated fresh, phrase by phrase, from live information about weather, air quality, conflict, economics, disease, and news sentiment around the world.

TARASOUND draws from four real-time data pipelines:

Weather — Open-Meteo provides current temperature anomaly, wind speed, and storm indicators from meteorological stations worldwide. Extreme heat, cold, and severe weather push scores lower.

Air Quality — OpenAQ aggregates PM2.5 particulate readings from monitoring stations across dozens of countries. Higher pollution loads reduce scores.

Conflict & News Sentiment — GDELT processes news media in real time, scoring the emotional tone of global reporting. Conflict, violence, and negative sentiment reduce scores while positive coverage raises them.

Economic Stress — World Bank indicators including unemployment rates, inflation, and poverty indices contribute a slower-moving but persistent signal reflecting underlying economic conditions.

Each region produces a score from 0 to 100, where higher numbers generally reflect more stable, positive conditions and lower numbers reflect stress, instability, or environmental pressure. The Global Index displayed in the header is the average of all active region scores — a single number representing the overall state of the world as TARASOUND understands it at that moment.

Scores are not static. They update on every data refresh cycle, typically every few minutes, and the direction and magnitude of change is as musically significant as the score itself.

Simulated — All scores are generated by an internal random walk algorithm. Each region drifts slowly toward a randomly chosen target, creating organic-feeling variation without any external data.

Live — All scores come exclusively from real-world data feeds. When a fetch fails, the affected region retains its last known score until new data arrives.

Hybrid — Live data is used wherever available. Regions with failed or stale fetches fall back to simulated values, ensuring continuous musical variation even during data outages.

TARASOUND has five audio engines — Percussion, Rhythm, Bass, Melody, and Harmony. Each engine is assigned to a geographic region on the map and responds to that region's score.

Scores influence tempo, harmonic complexity, rhythmic density, timbral character, and the probability of various musical events. A high score tends toward stability, consonance, and steady rhythm. A low score tends toward slower tempos, dissonance, irregular meters, and sparse textures.

The relationships between engines matter as much as individual scores. When two engines diverge significantly — for example when Harmony and Melody are pulled in opposite directions by their respective regions — the music reflects that tension through increased harmonic dissonance and rhythmic instability. When engines converge, the music becomes more unified and coherent.

TARASOUND uses probabilistic meter selection. Rather than locking into a fixed time signature, it rolls weighted dice at regular phrase intervals based on the current score. High scores favor familiar meters like 4/4 and 3/4. Low scores introduce complex asymmetric meters like 11/8, 13/8, and 15/8, creating the sense of rhythmic unease that mirrors a world under stress.

Under certain conditions — when a data refresh brings a sudden significant change in regional scores, or when the divergence between specific engines exceeds a threshold — TARASOUND briefly enters a glitch state. For sixteen musical phrases, additional note triggers are fired alongside the normal music, overwhelming certain synthesis engines and producing characteristic digital artifacts. This is an intentional compositional choice: the world's sudden changes deserve a sudden sonic response.

TARASOUND does not claim to be a precise scientific instrument. The relationship between raw data and musical output involves many layers of transformation, weighting, and artistic judgment. What it offers instead is a sustained, sincere attempt to make global conditions feelable — to give the abstract numbers of the world's vital signs a texture, a rhythm, and a voice.

To report a bug, suggest a feature, or just say hello —