There is a small industry now built around sleep sounds — apps, playlists, devices, branded soundscapes. Most of it is fine. Some of it is genuinely useful. A surprising amount of it is doing the opposite of what it claims, particularly the more elaborate generative soundscapes that promise to optimise sleep through complex algorithmic mixing.
I have used a fair amount of this material both for myself and for clients during sessions, and I have come to a small short list of what actually works for rest. The list is shorter than I would have expected. Most of what is sold is not on it.
What works
Rain. Just rain. Falling on leaves, on a roof, on a quiet street. Not a thunderstorm — the thunder activates the startle response — but steady moderate rainfall. This is, by some distance, the most reliable sleep sound for most people. The reason is partly evolutionary (rain meant a body was sheltered, fed, and safe) and partly acoustic (rain produces a broad-spectrum white-noise mask that covers other sounds without itself being patterned enough to attend to).
Distant wind in trees. Same mechanism. Slightly more variable than rain, so slightly less reliable.
Brown noise. Quieter and lower than white noise, with the high frequencies rolled off. Genuinely sleep-supportive in a way that white noise often is not — white noise has too much high-frequency energy and many people find it activating rather than calming over time.
Single instruments at very low volume. A piano played slowly, a cello held on single notes, a hang drum being struck without rhythm. The key word is single. Music with multiple instruments engages the brain's pattern-tracking circuits and prevents sleep onset.
What does not work
Most sleep music. The genre, broadly. Generative ambient mixes designed to optimise sleep are almost always too rich. They have too many instruments, too much harmonic motion, too many subtle changes. The brain stays engaged trying to follow them.
Binaural beats, in my experience and observation. The research is mixed at best. For most clients I have suggested them to, the effect was negligible.
Anything with vocals, even wordless. The brain finds the human voice irresistible. A whispered guided meditation is fine as long as it actually intends you to stay awake and engaged. As a sleep sound, voice is the wrong choice.
At what volume
Lower than you think. The sound should be at the threshold of audibility — present if you listen for it, easily ignored if you do not. Loud sleep sounds become something the brain is hosting rather than something the brain is using as a calming wall against other sounds. The point is mask, not entertainment.
At what time
Start the sound about ten minutes before lights-out. Stop it, or have it stop itself, ninety minutes into sleep. Sleeping through the entire night with sound playing is, for most people, slightly less restorative than sleeping through the second half in silence. The first ninety minutes are when the sound earns its keep — through sleep onset and into the first deep cycle. After that, the brain prefers silence.
Rain at low volume from ten minutes before sleep is the simplest sleep-sound protocol I know. Most of what is sold is doing more.
