EXPLAINER

Damage starts around 85 decibels, and the official limits disagree about time

OSHA and NIOSH use different rules for how fast allowable exposure falls as sound gets louder. WHO's venue standard permits an average of 100 decibels. The headline 'one billion at risk' figure is a modelled range.

Estimated prevalence of unsafe listening among people aged 12-34Personal listening devices: 23.81%; Loud entertainment venues: 48.2%0%25%50%Personal listening devices23.81%Loud entertainment venues48.2%
Estimated prevalence of unsafe listening among people aged 12-34
GroupValue (%)
Personal listening devices23.81 (18.99 to 29.42)
Loud entertainment venues48.2
Estimated prevalence of unsafe listening among people aged 12-34 The personal listening device estimate is a pooled prevalence with 95% CI. The venue estimate comes from a fitted model, because the pooled estimate carried limited certainty. Source: BMJ Global Health

Noise exposure above 85 decibels can damage hearing, and how much damage depends on how long the exposure lasts [s1]. There is no single number that answers "how loud is too loud", because the answer is a pair — level and duration — and the two US agencies that set those pairs use different arithmetic to relate them. OSHA's permissible exposure limit is 90 dBA averaged over an eight-hour day; NIOSH's exposure calculation halves the allowable time for every 3 dBA increase, where OSHA's halves it for every 5 dBA [s1]. On the same sound, those two rules give materially different verdicts about safety.

The occupational numbers, which are the only enforceable ones

OSHA's legal limit on workplace noise is a time-weighted average of 90 dBA over an eight-hour day [s1]. Separately, employers in general industry must implement a hearing conservation programme when employee noise exposures equal or exceed 85 dBA as an eight-hour time-weighted average; in construction, the threshold for administrative or engineering controls plus a hearing conservation programme is 90 dBA over eight hours [s1]. Those programmes require measuring noise levels, providing free annual hearing tests and free hearing protection, and training workers [s1].

The exchange rate is the part that non-specialists miss and that changes everything. OSHA uses a 5 dBA exchange rate: each 5 dBA increase halves the time a person may be exposed for the same dose [s1]. NIOSH recommends a 3 dBA exchange rate, on the reasoning that every 3 dBA doubles the sound energy [s1]. OSHA's own illustration of the difference is stark: it allows 8 hours of exposure at 90 dBA but only 2 hours at 100 dBA [s1]. NIOSH has found that significant noise-induced hearing loss occurs at exposure levels equivalent to the OSHA permissible exposure limit, based on updated information from literature reviews [s1].

Two rules of thumb from the same source are more useful than most consumer advice. If you need to raise your voice to speak to someone three feet away, the noise level might be over 85 dBA [s1]. And in open space, every doubling of the distance between you and a noise source reduces exposure by 6 dBA [s1] — which is why moving away from a speaker stack does more than any earplug you are not wearing.

The recreational standard is looser than the workplace one

WHO issued a global standard for safe listening at venues and events in March 2022, applying to places and activities where amplified music is played [s2]. Its first of six recommendations is a maximum average sound level of 100 decibels [s2]. The other five are about implementation: live monitoring and recording of sound levels using calibrated equipment by designated staff; optimising venue acoustics and sound systems; making personal hearing protection available to audiences with instructions for use; providing quiet zones where people can rest their ears; and training and informing staff [s2].

Set that 100 dB average beside the occupational figures and the gap is obvious. A workplace averaging 100 dBA would, on OSHA's own example, permit two hours of exposure [s1]. A venue standard that permits a 100 dB average across an entire event is a compromise between hearing protection and the experience people paid for — and WHO says so, framing the standard as limiting risk "while preserving high-quality sound and an enjoyable listening experience" [s2].

For personal audio devices, the earlier WHO-ITU standard, issued in 2019, is voluntary and specifies device features rather than a level: a sound allowance function that tracks a user's cumulative exposure as a percentage of a reference, a personalised listening profile, volume limiting options including automatic volume reduction and parental control, and general guidance on safe listening [s4]. WHO recommends that governments and manufacturers adopt it [s4]; it does not require them to.

Where the "one billion" number comes from

The figure that anchors most coverage of headphone risk deserves its provenance stated. WHO's 2019 release put nearly 50% of people aged 12 to 35 — 1.1 billion young people — at risk of hearing loss from prolonged and excessive exposure to loud sounds [s4], and the 2022 venue standard repeats "over 1 billion people aged 12 to 35" [s2].

The systematic review those figures now rest on is more explicit about its uncertainty than the headline is. Published in BMJ Global Health in 2022, it pooled 33 studies covering 19,046 individuals aged 12 to 34, searching for peer-reviewed work published between 2000 and 2021 [s3]. The pooled prevalence of unsafe listening from personal listening devices was 23.81% (95% CI 18.99% to 29.42%) [s3]. For loud entertainment venues the authors had limited certainty in their pooled estimate, so they instead fitted a model as a function of intensity thresholds and exposure duration and derived a prevalence of 48.20% [s3]. Applying those to the global population aged 12 to 34 produced a range — 0.67 to 1.35 billion — rather than a point estimate [s3].

So: roughly a quarter of young people use personal audio devices unsafely, measured; roughly half are exposed unsafely at venues, modelled; and the resulting global count spans a factor of two. The number is not wrong. It is a modelled range reported as a round figure, and the modelled half of it is the larger half.

The practical translation

The evidence supports a small number of concrete statements. Exposure over 85 decibels can damage hearing, and the damage accumulates with time at level [s1]. Doubling your distance from a source buys 6 dBA [s1]. Venues operating to the WHO standard may still average 100 decibels, a level at which occupational rules would cap exposure in hours, not evenings [s1] [s2]. And hearing loss from loud sound is permanent but preventable — WHO's own framing [s2].

What the evidence does not provide is a validated consumer threshold for headphone volume, because the exposure that matters is the product of level and duration and almost nobody measures either. The device features WHO and ITU specified in 2019 exist precisely to close that gap [s4]. This article is informational and is not medical advice; new or worsening hearing difficulty, or ringing that persists after loud exposure, warrants assessment by a clinician or audiologist.

Sources

  1. Occupational Noise ExposureU.S. Occupational Safety and Health Administration , August 1, 2026
  2. WHO releases new standard to tackle rising threat of hearing lossWorld Health Organization , March 2, 2022
  3. Prevalence and global estimates of unsafe listening practices in adolescents and young adults: a systematic review and meta-analysisBMJ Global Health , November 15, 2022
  4. New WHO-ITU standard aims to prevent hearing loss among 1.1 billion young peopleWorld Health Organization , February 12, 2019

More on

Related coverage