Noise Lab.

What the colors actually are

White	Equal energy at every frequency. Bright and hissy — classic static. Good for masking speech.
Pink	Falls 3dB per octave, so each octave carries equal energy — the way ears divide sound. Reads as balanced. The usual pick for sleep.
Brown	Falls 6dB per octave. Deep and rumbling, like a distant waterfall. Best against low rumble, and kinder if white noise feels harsh.
Blue	Rises 3dB per octave. Thin and airy, weighted to the treble.
Violet	Rises 6dB per octave. Very bright and sibilant. Turns up in tinnitus therapy.
Grey	Shaped against the ear's own sensitivity curve, so every frequency sounds equally loud rather than measuring equal.
Green	Concentrated near 500Hz, the middle of the natural-ambience band. Softer than pink, less rumble than brown.

Reading the display. The bars behind are what you're hearing right now. The colored line drawn over each one is that layer's own shape — brown slopes down to the left, violet climbs to the right, and a filter bends its curve as you move it. Shape by eye or by ear, whichever you trust.

Advanced mode adds a filter and a modulator to every layer, plus master tone and timing. Lowpass removes what's above the cutoff, highpass what's below, bandpass keeps a window, notch cuts one out — handy when a single frequency in the room bothers you. Resonance sharpens the effect at the cutoff. Movement slowly sweeps volume or cutoff so a layer never sits perfectly still.

Tones, and what the claims are worth

A tone layer is a sine wave at whatever frequency you set. The picker holds the numbers people look for — the nine solfeggio frequencies, 432 and 440 for tuning, and 40Hz — plus a fine slider for anything between 20Hz and 1200Hz.

Straight answer on the history. The solfeggio set isn't ancient. It was worked out by Joseph Puleo in the 1970s using Pythagorean digit-reduction on verse numbers in the Book of Numbers, and the Gregorian chant link was added afterward — medieval solfège syllables were relative pitch names, and hertz as a unit didn't exist until the 1800s. The 528Hz "DNA repair" claim traces back to one unpublished experiment that was never reproduced. 432Hz has a firmer thread: Verdi did argue for it and Italy did legislate it, though blind listening tests don't show people reliably telling 432 from 440. Nothing here treats or cures anything. The tones are provided because they're pleasant to layer under noise and people want them.

Binaural sends a slightly different frequency to each ear — 220Hz left, 226Hz right gives a 6Hz beat your brain constructs from the difference. It only works on headphones; on speakers the two signals mix in the air and the effect is gone. Research on it is mixed but real, mostly around anxiety and attention. Pulse gates a single tone on and off instead, so it survives on speakers. The beat rate readout names the EEG band it falls in.

40Hz is the one entry with a live research trail — MIT's gamma entrainment work on amyloid clearance is peer-reviewed and still in human trials. Early, audio-only is weaker than audiovisual, and it isn't a treatment. Set it as a pulse rate rather than a carrier: 40Hz as a tone is mostly inaudible on phone speakers.

Tones sit far quieter than noise layers by default, on purpose. A bare sine is fatiguing and reads as tinnitus to some people — two oscillators run a few cents apart to take the edge off, and a little brown or pink underneath makes almost any tone easier to sit with for an hour.

space play or pause · A add a layer · 1–9 toggle that layer