White Noise and Sleep: Newer Meta-Analysis, Pink Noise, and the Limits of Sound Masking

Bottom line: White noise has a stronger evidence case than it did several years ago, but it is not a universal sleep treatment. A 2025 meta-analysis found meaningful subjective sleep-quality improvements in several adult patient groups, while older reviews found highly inconsistent and low-certainty evidence. Pink noise should not be assumed to be superior: a 2026 polysomnography study found that continuous pink noise could reduce REM sleep even while partly masking intermittent environmental noise.[1-5]

Sound masking is different from sleep enhancement

The most defensible reason to use white noise is acoustic masking.

A steady background can reduce the contrast between baseline sound and unpredictable events such as:

• traffic;
• doors;
• voices;
• barking dogs;
• alarms; or
• hospital equipment.

That mechanism does not require white noise to directly “deepen” sleep. It simply makes disruptive peaks less salient.

This distinction matters because a sound can be useful as a masker without improving every sleep-stage metric.

Why the older evidence was so cautious

A 2020 systematic review evaluated 38 studies of continuous broadband noise and sleep.[1]

Studies varied substantially in sound spectrum, volume, population, outcome measurement, adherence and control conditions. Some reported benefits, some no effect and some potential disruption.

The reviewers rated the evidence for continuous noise improving sleep as very low quality.[1]

A broader 2022 review of auditory stimulation was also cautious. Only a minority of white-noise studies reported clearly positive sleep outcomes, and pink-noise studies were limited by methodological weaknesses and risk of bias.[3]

That older literature should remain visible even after newer positive trials.

The 2025 white-noise meta-analysis moves the evidence forward

A 2025 systematic review/meta-analysis restricted the evidence to randomized controlled trials and included 12 RCTs with 1,301 participants.[2]

Among adults ages 18–64, white noise reduced Pittsburgh Sleep Quality Index scores by about 3.70 points compared with control.[2]

Among older adults, the pooled difference was about 2.71 points in the favorable direction.[2]

The review also found favorable PSQI effects in ICU and non-ICU groups.[2]

That is a real update. White noise has more direct randomized evidence than the 2020 review could evaluate.

The newer result is still heterogeneous

The 2025 authors also reported substantial heterogeneity and uneven methodological quality.[2]

A sound machine in an ICU is not the same intervention as a consumer device in a quiet bedroom.

Clinical environments may have more room for improvement because they contain:

• alarms;
• staff conversations;
• equipment noise;
• hallway activity; and
• repeated sound peaks.

This is why a favorable hospital signal should not be generalized automatically to every home sleeper.

Hospital sound-machine evidence fits the masking hypothesis

A 2025 systematic review of sound machines in hospitalized adults suggested possible improvements in sleep efficiency, but the included studies were too heterogeneous for a clean pooled verdict.[4]

That is consistent with the idea that the intervention works best when a real environmental-noise problem exists.

If a bedroom is already quiet, adding continuous sound may have less upside and more potential to become another disturbance.

Pink noise is not automatically better

Pink noise contains relatively more low-frequency energy than white noise and is often perceived as softer or more natural.

That has led to marketing claims that pink noise:

• increases deep sleep;
• enhances memory;
• improves recovery; or
• is biologically superior to white noise.

The newer objective evidence complicates that story.

A 2026 randomized polysomnography experiment exposed healthy adults to intermittent environmental noise, continuous pink noise, earplugs and combinations.[5]

Environmental noise reduced N3 sleep. Pink noise partly mitigated some noise-related disruption—but continuous pink noise itself reduced REM sleep compared with the quiet condition.[5]

That is an important counterexample to “softer spectrum = better sleep.”

Solving one metric can worsen another

Sleep architecture is not a scoreboard where maximizing one stage always means better sleep.

A sound intervention might:

• reduce awakenings;
• improve subjective sleep;
• alter N3;
• reduce REM;
• change sleep efficiency; or
• have little effect on total sleep time.

Those outcomes should not be flattened into “improves sleep.”

See Subjective vs Objective Sleep and Sleep Onset vs Sleep Maintenance for the measurement framework.

White, pink, brown and other colors are different interventions

Consumer devices now include white, pink, brown, green and custom “sleep” spectra.

A trial of white noise does not validate brown noise. A pink-noise study does not validate every pink-noise app. A proprietary adaptive sound algorithm is another intervention again.

Sound spectrum, volume, speaker placement and whether the audio is continuous or responsive can all matter.

The tested intervention should control the claim.

Volume matters

The objective is masking, not maximum loudness.

A sound source loud enough to overwhelm every possible external noise can become an environmental stressor itself.

The older review specifically called for better characterization of acoustic exposure before widespread promotion.[1]

Consumer devices vary in:

• maximum output;
• distance from the sleeper;
• speaker quality;
• frequency response; and
• all-night versus timed playback.

There is no evidence-based universal sleep-machine volume in this literature.

Individual noise sensitivity matters

People differ dramatically in sound sensitivity.

A person who wakes to every vehicle outside may benefit from masking. Someone who finds continuous noise irritating may sleep worse with it.

This is another reason group averages should not become a universal recommendation.

Earplugs answer a different question

If the problem is external sound, reducing incoming sound and masking it are distinct strategies.

The 2026 pink-noise experiment also evaluated earplugs.[5]

For some sleepers, blocking the source may be simpler than adding another sound. For others, earplugs are uncomfortable or unsafe because they need to hear alarms, a child or a caregiving signal.

See Eye Masks, Earplugs, and Sleep.

Hearing and safety still matter

Continuous sound is an hours-long exposure.

The absence of an obvious adverse event in a small sleep trial is not proof that arbitrary high-volume playback near the ear is harmless.

Practical use should prioritize the lowest effective masking level rather than treating louder as more therapeutic.

What the evidence does not establish

Current evidence does not justify saying that white or pink noise:

• treats chronic insomnia in everyone;
• reliably increases deep sleep;
• reliably increases REM;
• works better than CBT-I;
• has one universal optimal volume;
• works equally well in quiet and noisy rooms;
• makes pink noise superior to white noise; or
• validates every colored-noise app or sound machine.

Where sound masking fits

Sound masking is best viewed as a low-complexity environmental experiment when unpredictable noise is a real bottleneck.

If the actual problem is chronic insomnia, circadian delay, sleep apnea, restless legs, pain, caffeine, alcohol or insufficient sleep opportunity, changing the color of background noise is unlikely to solve the main problem.

The broader room-level decision guide is Sleep Environment Evidence Guide.

Bottom line

The sound evidence has evolved without becoming simple.

The strongest current interpretation is:

• older reviews found inconsistent, very-low-certainty evidence for continuous noise;
• a 2025 RCT meta-analysis found meaningful subjective sleep-quality improvements with white noise in selected adult and clinical populations;
• those studies remain heterogeneous;
• a 2026 PSG study shows that pink noise can reduce REM sleep even while masking some environmental disruption; and
• benefit is most plausible when unpredictable environmental noise is the problem being solved.

White and pink noise are tools, not sleep-stage optimization frequencies.

Related reading

• Sleep Environment Evidence Guide
• Eye Masks, Earplugs, and Sleep
• Subjective vs Objective Sleep
• Why Sleep Studies Disagree