Brown, pink & white noise — mixed live in your browser
Press play and this page becomes a noise machine. There are no recordings and nothing streams: all three colors are synthesized on the spot with the Web Audio API, level-matched so none wins just by being louder. Play one color on its own, blend all three with the sliders, and let the sleep timer fade the sound out instead of cutting it off.
Press play — every sound is synthesized live in your browser. No audio files, nothing streams.
No timer set — playback continues until you pause it. With a timer, the sound fades out gradually at the end instead of cutting off.
Volume first: start low and raise it only as far as you need. Any sound — including gentle noise — can damage hearing if it is played too loud for too long. Noise is a sound-masking aid, not a medical treatment; if you have persistent sleep problems, ringing in the ears or trouble concentrating, talk to a doctor rather than turning the volume up.
The three colors, briefly
White noise
Equal energy at every frequency — a flat spectrum, 0 dB per octave. Bright and hissy, like an untuned radio or a fan on full.
Pink noise
Energy halves with every octave (−3 dB/octave), so the highs recede. Most people hear it as steady rain or wind in trees.
Brown noise
Energy quarters with every octave (−6 dB/octave): a deep, soft rumble, like a distant waterfall or the cabin hum on a long flight.
Not sure which fits your room? The side-by-side comparison lays out the spectra and the situations people reach for each color, and the sound masking explainer covers why a steady sound makes sudden noises feel smaller.
How the sound is made
Each color is generated sample by sample. White noise is a buffer of random values. Pink noise runs that randomness through Paul Kellet's filter — a small cascade of one-pole sections that tilts the spectrum down by about 3 dB per octave. Brown noise feeds it into a leaky integrator, which accumulates the signal while slowly forgetting it, tilting the spectrum by about 6 dB per octave. Before anything plays, every color is normalised to the same RMS level (0.22), so switching colors changes the character of the sound, not its loudness. Buffers are ten-second loops with the tail blended into the head, so you never hear the loop point click. We verified the generators offline at 44.1 kHz: measured octave slopes came out at about 0 dB/octave for white, −3 dB/octave for pink and −6 dB/octave for brown — the numbers the spectrum diagrams on this site show.
The sleep timer fades — it never cuts
Set a timer (15 to 90 minutes, or any length up to 600) and the last stretch is a fade: the gain falls in a straight line from full to zero over the final 10% of the timer, capped at 60 seconds. A 30-minute timer therefore plays 29 minutes at your volume and spends its last minute quietly disappearing. The countdown above the player shows exactly when the fade starts. Many people use steady noise to soften the sounds that would otherwise wake them — a door, a truck outside, a partner coming to bed late — and a fade means the room is truly quiet once you are asleep.
Getting the bedtime itself right
Noise can soften a room, but when you fall asleep matters just as much. If you are pairing a sleep timer with a planned bedtime, the Sleep Cycle Calculator works out bedtimes that land your wake-up between sleep cycles — set the noise timer to cover the falling-asleep window, and let the cycle math handle the alarm.
Export your mix as a WAV
The export button renders a fresh 60-second WAV of whatever the sliders are set to — useful if you want the same mix on another device. It is generated in your browser from the same code as the player (not a recording of the live output), peak-limited so a full three-color stack cannot clip, and capped at 60 seconds. Nothing is uploaded at any point.
Questions
- Is the noise really generated live, or is it a recording?
- Generated live. The page ships no audio files at all: white noise is random samples, pink is a filtered version of that randomness, and brown is its integrated form. Leave the tab open for hours and it never repeats a "track," because there is no track — just a ten-second loop of freshly generated sound, crossfaded so the loop point is inaudible.
- Which color is best for sleep?
- There is no proven best color — it is a preference. Brown is the most popular for sleep because the missing high frequencies feel less "present," while pink suits people who find brown too muffled. What matters is that the sound is steady, quiet and boring. The noise for sleep guide covers volume and timer habits.
- Can this cure insomnia, tinnitus or ADHD?
- No, and we will not claim otherwise. Many people use steady noise to mask environmental sounds, and some find it helps them settle; that is the extent of it. It is not a treatment for any condition. Persistent sleep problems, tinnitus or attention difficulties deserve a clinician, not a louder speaker.
- Why does the timer fade instead of stopping?
- A sudden silence can be as noticeable as a sudden sound. The fade (the last 10% of the timer, up to 60 seconds, in a straight line down to zero) lets the room go quiet gradually — and by the time it does, you are usually asleep.
- Does the player work offline or in a background tab?
- Once the page has loaded, playback needs no connection. Background tabs keep playing in most desktop browsers; mobile browsers may pause audio when the screen locks, which is a system policy we cannot override — the sleep timer is wall-clock based, so if audio is paused by the system the countdown still reflects real time.