Everything you've ever heard is air, trembling at a certain speed. This instrument plays that ingredient raw · and shows you what your body does with it.
The signal chain is a single OscillatorNode set to type: 'sine', running into a GainNode → AnalyserNode → output. The AudioContext is requested at 48 kHz; Web Audio processes internally in 32-bit float, and the sine is synthesized on the fly. There's no source file, no codec, no compression. The only place fidelity can be lost is your DAC and transducers · that's the part headphones actually change.
The slider is logarithmic across 20 Hz to 20 kHz, so an octave covers the same distance everywhere. Drag freely · sound starts the moment you move it. The label updates as you cross each band: sub-bass, bass, low-mid, midrange, presence, treble, air, ultrasonic. The interface tints itself to match where you are in the spectrum · ember red at the bottom, violet at the top · a nod to the synesthetes who see it that way. On a keyboard: space plays and pauses, ← → nudge the active tone (hold shift for bigger jumps), and tapping the waveform itself toggles sound.
As you sweep, the italic line under the readout tells you what lives at that frequency: the lowest key on a piano, the hum of the power grid, a cat's purr, the whine of an old CRT television, the "mosquito tone" only young ears can hear. The spectrum isn't an abstract axis · it's a landscape, and these are its monuments. Try walking the slider slowly from end to end and visiting each one.
The six highlighted frequencies cluster around tones traditionally associated with specific states. 40 Hz has the most research behind it · gamma-band entrainment and cognitive focus. The other five (174 / 432 / 528 / 741 / 963 Hz) come from the Solfeggio set: folk tradition rather than peer-reviewed science, but evocative to listen through in order. Treat them as a tasting menu rather than medicine.
The Feelings tab approaches the spectrum from the other direction: instead of picking a number, pick a sensation your body already knows · a cat's purr against your ribs, the shiver of very high "air," the queasy pull of the tritone, the half-remembered whine of a CRT television. Each one loads the frequency (or combination) most associated with that feeling. Two of them are honest folklore: the "brown note" is a myth that has survived every laboratory attempt to confirm it, and 432 Hz calm is tradition rather than science. They're included because testing folklore on your own nervous system is half the fun.
Stack multiple tones with Add layer to build chords or explore interference. New layers default to a perfect fifth (1.5×) above the active one. The slider, presets and waveform all follow whichever layer pill is highlighted · click another pill to edit it. The small ◦ / L / R mark on each pill cycles its stereo position: pan two slightly different tones hard left and hard right and you've hand-built your own binaural beat. Per-layer gain auto-normalizes by 1/N so the master never clips, which keeps every tone arithmetically pure even when stacked.
Below the waveform display, four shape buttons set the active layer's harmonic content. Sine is a single frequency, pure. Triangle adds soft odd harmonics · mellow and clarinet-like. Square is all odd harmonics at full strength · hollow, reedy. Sawtooth contains every harmonic · bright and brassy. Same fundamental, completely different texture. Each layer can have its own shape, so you can mix a pure sine bass with a sawtooth lead. To see the difference rather than just hear it, hit the harmonics button in the display: the view switches from the waveform to its spectrum, and a sawtooth resolves into the comb of overtones it really is. A sine, in that view, is a single lonely spike · the only sound in the world with nothing hidden inside it.
The Combinations tab loads multi-layer presets. Delta / theta / alpha / gamma beats are true binaural beats: a tone hard-panned left and a slightly different tone hard-panned right. The brain hears the difference (2.5 / 6 / 10 / 40 Hz) as a phantom pulse · these only work through headphones, never speakers. The chord presets use just intonation (small-integer frequency ratios) rather than equal temperament, so they're physically more consonant than a piano playing the same chord. Listen to the major and minor triads back to back · one note moves a few hertz and the entire emotional weather changes. Then try the tritone, the interval medieval theorists called diabolus in musica and avoided in church music; its unease is mathematical, not superstitious.
The Share button copies a URL that encodes the full layer set (frequencies, shapes, pan). Opening that URL reconstructs the exact combination on someone else's device. Great for sharing a chord or beat on X with one link.
Most adults stop hearing somewhere between 14 and 17 kHz, and the limit drops with age and noise exposure. The Find your hearing ceiling button runs a slow sweep from 4 kHz upward · press the button the instant the tone disappears and you'll get your ceiling and a rough "hearing age." It's a parlor estimate, not an audiogram: the result depends on your device volume and on whether your headphones can actually reproduce the extremes. Compare with someone a decade older or younger · the gap is real and a little haunting. At the other end of the dial, anything below ~60 Hz is more felt than heard.