Put on headphones, dim the lights, and let it play. It runs about seven minutes; the score, the sounds and the world are all made live as you watch.
On a phone, turn it sideways for the widest view.
3D graphics could not start in this browser, so the vision will play as sound, narration and scene descriptions.
Scene descriptions add a second voice that describes what is on screen, in the pauses between verses. With the narrator silent, your screen reader reads the verses instead.
Contains flashes of light in its storm passage.
Every image here is drawn in real time with WebGL (three.js), and every sound is synthesized in your browser: the hammered dulcimer is plucked-string physics, the ancestral voices are formant choirs, the rivers and the fountain are sculpted noise. The narration is a Kokoro neural voice, generated for this piece. Poem and preface are in the public domain.