Does sound healing work? What the science shows

What is sound healing?

Sound healing is a name for many practices, not one method. It covers sound baths, where singing bowls, gongs and chimes are played while people lie still; recorded tones at chosen frequencies, such as 285 Hz or 528 Hz; binaural beats, two slightly different tones, one in each ear, heard as a slow beat; and plain listening to music. Each has been tested to a very different degree, so whether sound healing works depends on which part you mean.

Sound healing practices and their testing

Practice	What you hear	Tested how far
Listening to music	Recorded music	More than 100 randomised trials
Music therapy	Music with a registered therapist	47 studies on stress alone
Singing bowls and sound baths	Bowls, gongs and chimes, played live	About 20 studies, most of them small
Binaural beats	A different tone in each ear	Dozens of studies, mixed results
Single frequencies	One tone, such as 285 Hz	A few small studies of 432 and 528 Hz; none of 285 Hz

Music therapy is the one part that is a profession. The American Music Therapy Association defines it as music used "within a therapeutic relationship by a credentialed professional who has completed an approved music therapy program" (AMTA). In the UK, music therapist is a title protected by law: to use it, a person must be registered with the Health and Care Professions Council (HCPC). The US National Center for Complementary and Integrative Health makes the same point from the other side: the term names "the education, training, and credentials of the therapist", not a type of sound (NCCIH). A sound bath or a frequency video is not music therapy, however it is labelled.

The word healing is the one people search for, so Harmonance uses it to name the field, never as a promise. The history of sound healing traces where the practices come from, and what sound healing is sets out how Harmonance approaches it.

Does sound healing work?

For music in general, there is good evidence of a modest effect on how stressed people feel; for exact frequencies, there is almost none. The gap is mostly a matter of how many people have been tested, and how fairly.

The numbers behind the chart

What was tested	Studies	People
Music and stress	104 trials	9,617
Music and sleep	13 trials	1,007
432 Hz against 440 Hz	8 trials	250
528 Hz, people who heard it	5 studies	213
285 Hz	none found	none

Download the numbers (CSV)

The two largest rows are reviews that pool many randomised trials of music. The rows below them come from small studies that Harmonance's 432 Hz guide and 528 Hz guide list one by one, and 285 Hz has none. Singing bowls and binaural beats are left off the chart because the summaries of their reviews count studies, 19 of bowls and 22 of beats, but not people; both are covered below.

What does research show about music and stress?

On average, people given music report less stress than people who are not, by a modest amount. A review in Health Psychology Review pooled 104 randomised controlled trials with 9,617 participants and found an effect on psychological measures of stress of d = 0.545, and on physical measures, such as heart rate, of d = 0.380 (de Witte and colleagues). The comparison groups had their usual care or another activity instead of the music (the review in full). A second review, of 47 studies with 2,747 people that kept only music therapy, found larger effects on stress-related measures, NCCIH reports.

What does d = 0.545 mean in plain numbers? Statisticians translate an effect size into a "common language" figure: the probability that a person picked at random from one group scores higher than a person picked at random from the other (Lakens). Divide d by the square root of two, 1.414, and read the normal curve: 0.545 becomes 0.385, and 65 per cent of the curve lies below that point. So pick one person who had music and one who did not, and the music listener reported less stress about 65 times in 100, where a coin toss would give 50. For heart rate, d = 0.456, the same sum gives about 63 in 100. That is a real shift, and a modest one.

These trials tested music, not a frequency, so what they found is about listening in general, not about any number. People also always know when they are hearing music, so no trial like this can hide from its listeners which group they are in, and expectation can move how people say they feel.

What does research show about music and sleep?

For adults who sleep poorly, a Cochrane review found that listening to music made a moderate difference to how they rated their sleep. It pooled 13 randomised trials with 1,007 participants, who listened to recorded music for 25 to 60 minutes a day for between three days and three months (Jespersen and colleagues). On the Pittsburgh Sleep Quality Index, a questionnaire scored from 0 to 21 where lower means sleeping well, the music groups scored 2.79 points lower than comparison groups who had their usual care or nothing, across 10 studies and 708 people.

The reviewers put that at about one standard deviation, which in common language terms is about 76 in 100. They rated the certainty moderate and named the limits plainly: every trial was at high risk of bias because participants could not be blinded to music, and in the three trials that measured sleep with instruments rather than questionnaires, 136 people, music may have made no difference. No trial reported any harm. Like the stress trials, they tested music, not a frequency.

Do singing bowls and sound baths work?

Singing bowls have a small body of research with encouraging early results and weak designs. The newest systematic review found 19 clinical studies published between 2008 and 2024, nine of them randomised trials and seven case series with no comparison group, and concluded that singing bowls "may have potential"; most of the studies measured mental health outcomes (Cai and colleagues). Another review, of 14 studies, could not pool them at all: "Due to heterogeneity in study design and limited articles, no meta-analysis was conducted" (Lin and colleagues). An earlier review found only four peer-reviewed studies and wrote, "we cannot recommend singing bowl therapies at this stage" (Stanhope and Weinstein).

Two studies show what the evidence looks like up close. In California, 62 adults took part in a singing bowl meditation of about 60 minutes and afterwards reported less tension, anger and fatigue than before. The authors named the main limit themselves: "this was a nonrandomized observational study without a control group" (Goldsby and colleagues).

A stronger design followed. Fifty adults who scored high on a questionnaire about how tense and worried they felt at that moment were randomly given a singing bowl session, progressive muscle relaxation or a place on a waiting list. Both the bowls and the muscle relaxation left people calmer than waiting did, and the bowls more clearly so (Rio-Alamos and colleagues). That is one session with 50 people, and it compared bowls with another way to relax, not with other pleasant sounds. The guide to sound baths covers what to expect in a session and how to keep the volume safe.

Do specific frequencies like 285 Hz, 432 Hz or 528 Hz work?

No large or repeated study has shown that one frequency does something another does not. The studies that exist are small, and many compare music at a chosen frequency with silence, which cannot separate the tone from the music.

432 Hz has the most direct tests: eight small trials have played the same music at 432 and 440 Hz, 250 people in all, and their results point different ways. The 432 Hz guide reads each one, and tells the tuning's history. One of the eight shows the design at its tightest: in a double-blind crossover pilot, 33 volunteers heard the same film soundtracks tuned to 440 Hz on one day and to 432 Hz on another. With 432 Hz their heart rate averaged 4.79 beats a minute lower, right at the edge of statistical significance, p = 0.05, and the authors called for larger trials (Calamassi and Pomponi).

528 Hz has been played to people in five small studies, 213 people in all, each measuring something different; the 528 Hz guide reads every one. The study that also tested 432 Hz gave 162 university students 20 minutes of no music, 528 Hz music or 432 Hz music. Some saliva markers differed between the groups, but an attention test showed no difference (Bozok and colleagues). Each music group was compared with silence, not with the same music at ordinary tuning, so the study cannot say whether the frequency or simply the music made the difference.

No study has asked what 285 Hz does for the people who hear it. PubMed finds nothing for 285 Hz with music or Solfeggio, and for 285 Hz with listening it finds two hearing studies that used the number only as a test tone: one played tones up to 285 Hz to 126 young adults to record their hearing pathway (Prendergast and colleagues), and in the other, zebra finches and budgerigars caught a mistuned harmonic at 285 Hz more finely than people did (Lohr and Dooling).

The number belongs to the Solfeggio set, which grew from six numbers first printed in a 1999 book, Healing codes for the biological apocalypse, whose Library of Congress record files it under subjects including prophecy and numerology (Library of Congress). 285 was not one of the six: as the Solfeggio frequencies guide sets out, 174, 285 and 963 were added later, and no record names who added them. On Harmonance, 285 Hz carries the tradition's intention, restoration, as something to listen with, not as an effect.

Binaural beats have more research, and it points both ways. A meta-analysis of 22 studies found a medium average effect, g = 0.45 or about 62 in 100, which depended on the beat rate, the listening time and when people listened (Garcia-Argibay and colleagues). A systematic review of 14 brain recording studies found five in line with the idea that brain rhythms follow the beat, eight against it and one mixed (Ingendoh and colleagues). And when 221 people each fell asleep to three versions of the same sound, with a binaural beat, with a monaural beat and with no beat, the time they took to fall asleep did not depend on the version (Shumov and colleagues). The binaural beats guide goes further.

How could a study show that a frequency matters?

It would have to change one thing and keep everything else the same. Lying still, closing your eyes, a quiet room, time to yourself and the belief that something is working can all leave people calmer, so a fair test of a frequency compares it with the same sound minus that one difference, ideally with listeners not told which they hear.

What each kind of test can show

Test	Can show	Example
Before and after	People felt different afterwards	62 people, singing bowls
Against no music	Listening beats doing nothing	1,007 people, music and sleep
Against another way to relax	The sound did more than another calm activity	50 people, bowls against muscle relaxation
One thing changed	Whether that one change matters	33 people, 432 Hz against 440 Hz

Most sound healing research stands on the lower steps. Before and after studies show that people feel different afterwards; trials against no music show that music beats nothing; trials against another way to relax show that the sound did more than that activity. Only the top step can isolate a frequency, and it holds very little: the eight small 432 Hz trials and the 221-person beat comparison. None of them tested 285 Hz.

What does acoustics describe?

Acoustics describes the sound itself, and for a single tone it is exact. A tone of 285 Hz makes the air press and release 285 times a second. Sound moves through air at about 1,100 feet, or 335 metres, a second at sea level (NASA), so each of those waves is about 1.2 metres long. On a piano tuned to concert pitch, with the A above middle C at 440 Hz, 285 Hz falls almost exactly between two keys: 48 cents above the C sharp at 277.18 Hz and 52 cents below the D at 293.66 Hz, where a semitone is 100 cents. Nothing in its physics sets it apart from 280 or 290 Hz.

Harmonance measured what its own player sounds when you press play at the top of this page. The pure tone, which plays unless you switch on Strings, is a sine wave at 285 Hz with two quieter companions added for warmth: 97.5 per cent of its power is at 285 Hz, 2.1 per cent at 142.5 Hz an octave below, and 0.35 per cent at 570 Hz an octave above. Switch on Strings and it plays strings tuned to octaves of 285 Hz instead, with a quiet 285 Hz tone underneath.

Acoustics also explains beating. Two tones a few hertz apart, heard together, swell and fade at the difference between them, so 285 and 289 Hz pulse four times a second. A binaural beat sends one tone to each ear instead, and listeners hear a third tone pulsing at the difference, though the air at either ear carries no beat (Ingendoh and colleagues). What acoustics does not describe is an effect on health: it measures the sound, not what the sound does for the person hearing it.

Why might sound feel relaxing?

Tempo and silence seem to matter more than pitch. In a study of 24 adults, breathing and heart rate rose with faster music, slow or meditative music had a calming effect, and a two-minute pause of silence brought heart rate and breathing below where they had been before any music: "relaxation is particularly evident during a pause" (Bernardi and colleagues).

A steady tone has no tempo at all, which puts it nearer the slow end of that study than the fast end. No study has tested a steady tone this way, so that is a reason to try one, not a finding.

Why does Harmonance use the words sound healing?

Because it is what people call the field, and Harmonance would rather meet people there and tell them what is known. Every tone on the site plays free, with no ads, and every tone page says what studies do and do not show. Traditions are told whole, with where their numbers come from; tones are offered as intentions to listen with; nothing is promised.

How can you try sound healing at home?

• Choose a time when you can sit or lie still for ten minutes.
• Keep the volume low: listening to music at too high a volume can contribute to noise-induced hearing loss, NCCIH notes. The same care applies to loud gongs at a sound bath.
• Notice how you feel before you press play and again when the sound stops. It will not tell you why you feel different, but it will tell you whether you do.
• Headphones are needed only for binaural beats; a speaker is fine for a single tone.
• Stop if anything feels uncomfortable, take care if you have tinnitus or sensitive hearing, and never listen while driving.

Press play at the top of the page to hear 285 Hz, or open Need to unwind?, which pairs it with 432 Hz and Neptune at 211.44 Hz. In Harmonance Studio you can mix tones of your own. Play one quietly for ten minutes while you do nothing in particular, and notice how you feel.

Sources and limits

The effect sizes are the reviews' own; the "in 100" figures are Harmonance's arithmetic from them, which assumes the scores are spread on a normal curve, as the method does. Most studies were read through their abstracts; the stress review's methods, the singing bowl observational study, the tempo study and the effect size guide were read in full. The chart counts the people the music reviews report and the people in the studies that Harmonance's 432 Hz and 528 Hz guides list, and some trials may sit in more than one review. The player's sound was measured by rendering its own tone and reading its spectrum. PubMed holds no study of what 285 Hz does for listeners, which does not prove that none exists elsewhere.

1. de Witte and colleagues, music interventions and stress-related outcomes, Health Psychology Review: 104 trials, 9,617 participants.
2. de Witte and colleagues, the same review in full, University of Amsterdam repository: how the comparison groups were chosen.
3. Jespersen and colleagues, listening to music for sleep in adults, Cochrane: 13 trials, 1,007 participants.
4. Lakens, calculating and reporting effect sizes, Frontiers in Psychology: the common language effect size.
5. NCCIH, Music and Health: What You Need To Know: music therapy as a credential, the music therapy review, and listening volume.
6. American Music Therapy Association, what is music therapy.
7. Health and Care Professions Council, the professions it regulates: protected titles, music therapist among them.
8. Cai and colleagues, singing bowls: a systematic review of clinical studies, Integrative Medicine Research.
9. Lin and colleagues, Tibetan singing bowls in adults: a systematic review, Healthcare.
10. Stanhope and Weinstein, singing bowls: a systematic review, Complementary Therapies in Medicine.
11. Goldsby and colleagues, singing bowl sound meditation: an observational study, full text.
12. Rio-Alamos and colleagues, Tibetan singing bowl sounds: a randomised controlled trial.
13. Calamassi and Pomponi, music tuned to 440 Hz versus 432 Hz: a double-blind crossover pilot, Explore.
14. Bozok and colleagues, 528 Hz and 432 Hz music in 162 students, Brain and Behavior.
15. Prendergast and colleagues, noise exposure in young adults with normal hearing, Hearing Research: tones up to 285 Hz as test sounds.
16. Lohr and Dooling, mistuned harmonics heard by zebra finches, budgerigars and people, Journal of Comparative Psychology.
17. Library of Congress, the 1999 book's catalogue record.
18. Garcia-Argibay and colleagues, binaural beats: a meta-analysis, Psychological Research.
19. Ingendoh and colleagues, binaural beats and brain rhythms: a systematic review, PLOS ONE.
20. Shumov and colleagues, binaural or monaural beats and time to fall asleep in 221 people.
21. NASA Glenn Research Center, the speed of sound.
22. Bernardi and colleagues, music, tempo and the pause, Heart, full text.