Dataset Viewer
Auto-converted to Parquet Duplicate
Search is not available for this dataset
audio
audioduration (s)
96
135
End of preview. Expand in Data Studio

World Sound — the hisia.live library

Every track here is AI-generated: a mirror of one city's weekly chart, made with Stable Audio Open Small from a prompt derived from that chart week's measured tempo, key, instrumentation, mood and genre label. No recording is ever an input to a track, and no track is a cover, a soundalike or a real song. The artist and title on every track are invented.

  • audio/<id>.mp3 — the published tracks, as played on https://hisia.live
  • library.json — per track: city, invented artist and title, chart week, recipe, version, loudness; per city: readout and the generation prompt
  • prompts.json — the prompts, one per city

Licence: audio and metadata CC BY 4.0 (credit "World Sound / hisia.live"). How it is made: https://decimalpoints.substack.com/p/world-sound-how-its-made Updated at every publish by radio/refresh/push_library.py.

Downloads last month
63