Nobody recorded a singer. The voice is arithmetic — a buzz shaped by five rooms in the mouth.
Every vowel is the same raw sound — a buzzing tone from the throat — filtered by the shape you make above it. Push your tongue and lips into different rooms and the same buzz turns into ah, ee, oo. Those rooms ring at fixed pitches called formants, and it is only their pattern, not the buzz, that your ear reads as a vowel.
CHOIR carries three of those rooms as filters and slides them as your cursor crosses the nave. On the far left they sit where a mouth sits for ah; drag right and they part and climb through eh and ee, then round and drop for oh and oo. You are not choosing a recording — you are re-shaping a throat in real time.
- What you move
- Left to right morphs the vowel through the five cardinal shapes. Up and down is pitch — low at your feet, two octaves higher at the vault.
- Second reading
- The rose window at the vanishing point is not decoration. It brightens with the loudness of your voice and warms toward gold on the closed vowels — watch it while you hold an ee.
Five shapes, and everything between them. Your cursor lives on the line that joins them.
Each vowel is a fingerprint of three resonances. These are the anchor values CHOIR interpolates between as you slide — the numbers a phonetician would measure from a spectrogram exactly like the one above.
A cathedral answers every voice with itself. CHOIR wraps yours in a synthesised stone reverb — a three-second tail modelled on a long vaulted nave — so a single moved finger becomes a chorus of its own reflections.
- Hold still
- Stop moving and the voiced tone falls away within a breath. What is left is not silence but aspiration — the sound of air in a throat that has stopped singing, still coloured by the last vowel you held.
- Move again
- The instant your cursor stirs, the buzz returns and swells with how fast you move. Fast strokes sing out; a slow drift barely voices. The loudness is your velocity.
- Never recorded
- There is no audio file anywhere on this page. Every sample is computed live in WebAudio the moment it reaches your ears, and it exists only while you are here to make it.