White, pink, brown, and blue noise. Live in your browser.

White noise is uniform across the audible frequency spectrum. Pink noise drops 3 dB per octave — emphasising the lower frequencies, which makes it sound deeper and matches the muffled in-utero soundscape that newborns are calibrated to. Brown noise drops 6 dB per octave — even deeper, often described as a soft waterfall or the rumble of a faraway road. Blue noise rises in the high frequencies — rarely used for sleep, occasionally helpful for masking specific high-pitched annoyances.

The audio is generated live in the browser via the Web Audio API. White noise is uniform random samples. Pink uses Paul Kellet's refined seven-stage filter cascade. Brown is integrated white with a leak bias to prevent DC drift. Blue is a high-pass differentiator. Generating live means no MP3 streaming, no licensing concerns, and zero file size — the noise is mathematically pure.

The volume cap recommendations come from the use-case selector. For infant use, the cap maps to roughly 50 dB SPL at 1 meter from the speaker — the AAP guidance for safe infant white noise (Hugh-Hazelden et al. 2014 review). For office use the cap is closer to 60 dB, the rough threshold below which sustained eight-hour exposure does not fatigue the auditory system. For adult sleep the cap is 70 dB, well below sustained-exposure risk thresholds for adults.

Critical infant safety notes. The noise machine should be placed at least 7 feet (2 meters) from the crib — never inside it. Volume at the crib should sit in the green zone of most baby monitors, around 50 dB. Run for the duration of the nap or sleep period and turn off — continuous all-night use is not the recommendation. The 8-hour timer option is intended for adult use; the form will warn you if you select it with the baby use case.

This tool ships the four noise colors live; we have not yet added the named ambient sounds (rain, ocean, fan, etc.) that some users prefer. Those are a future addition that requires hosted audio files and licensing review.

The colours are just different frequency profiles. White noise spreads energy evenly across all frequencies and sounds hissy; pink noise weights toward the lows and sounds like steady rain; brown noise weights even further down and sounds like a distant waterfall. None is objectively best, so pick the one you find least intrusive.

Whichever you choose, the job is masking, not sedation. The point is to raise the background floor so a sudden noise does not stand out and jolt you awake, which is why volume and placement matter more than colour: keep it quiet and across the room.

The clearest, best-supported benefit is masking. In genuinely noisy environments, a steady sound floor reduces the number of disruptive noises that reach the sleeping brain. Beyond that, the picture gets murkier: a systematic review by Riedy and colleagues found the overall evidence for noise as a sleep aid limited and inconsistent, with some people actually sleeping worse under constant sound.

There is intriguing specialist research too. Papalambros and colleagues showed that pink noise, precisely timed to the brain's slow waves in a lab, boosted deep sleep and next-day memory in older adults. That is a promising finding, but it involved phase-locked laboratory stimulation, not a bedside speaker looping all night, so it is a reason to be curious rather than a promise for a sound machine.

Keep the volume low, put the source across the room, and use a timer if you do not actually need it all night. If it clearly helps you fall asleep and stay asleep, keep it; if you find yourself sleeping worse or waking when it changes, it is not for you, and that is a legitimate result.

For babies, treat volume and distance as a safety issue, not just a preference. Keep any machine well away from the crib and at a low level, since infant hearing is more vulnerable, and avoid leaning on it as the only thing that gets a baby to sleep.