Industries / Industrial equipment and services / Industrial services / Industrial hygiene
Noise
Noise work measures how loud a workplace is and how much sound each worker takes in over a shift, tests workers' hearing year after year, and cuts exposure with quieter machines, enclosures, shorter shifts near loud equipment and hearing protectors.
7
Milestones achieved, from 1886 to 2014.
1
To watch for: announced, not yet confirmed outside.
2
Open deliverables. Nobody has shipped them.
The road
AchievedWatch forOpen
Achieved
1886
Boilermakers' hearing compared
Systematic studies of deafness caused by work began in the late 19th century. Atherley and Noble, writing in 1985, name two landmarks: Gottstein and Kayser's 1881 study of staff at a German railway works, and Barr's 1886 study of boilermakers in Scotland, which also asked the men what trouble they had hearing in everyday life. Both found the disease by comparing the hearing of whole trades.
LessonComparing the hearing of different trades was enough to show that noise at work causes deafness, long before anyone measured each worker's dose.
Credit
Barr; Gottstein and Kayser
1969
A limit for federal contractors
A 1969 rule under the Walsh-Healey Public Contracts Act set noise limits for firms holding US government supply contracts worth more than $10,000. OSHA, the federal safety regulator set up in 1970, adopted the same limit for general industry: 90 dBA averaged over 8 hours, where dBA (A-weighted decibels) is a sound level weighted the way the ear hears, with a 5 dB exchange rate, meaning each 5 dB rise halves the time allowed. NIOSH, the federal research institute for worker health, recommended 85 dBA in 1972, and in 1998 a 3 dB exchange rate, the rule that, according to a 1991 review by Alice Suter, the US EPA, most European countries and the UK already used; OSHA's limit has not changed.
Credit
US Department of Labor; NIOSH
1979
Labels that overstate protection
From 1979 the US Environmental Protection Agency required every hearing protector sold to carry a Noise Reduction Rating (NRR) measured in a laboratory. Field studies later found that the laboratory ratings overstated the protection workers actually got by 140% to 2,000% (Berger and colleagues, 1996, as summarized by NIOSH), and NIOSH now tells users to cut the labeled rating by 25% for earmuffs, 50% for foam earplugs and 70% for other earplugs.
LessonA rating measured with experts fitting the device does not tell you what a worker who fits it alone will get.
Credit
US EPA; field studies by Berger and others
1983
Hearing tests become routine
OSHA's hearing conservation amendment, made final on 8 March 1983 after a 1981 version was stayed for reconsideration, requires employers to monitor noise and to give a baseline hearing test and then yearly tests, training and fitted hearing protectors to every worker exposed at or above 85 dBA averaged over 8 hours. A standard threshold shift, an average loss of 10 dB or more at 2,000, 3,000 and 4,000 Hz in either ear compared with the baseline, means the worker must be refitted and must wear protectors.
Credit
OSHA
2003
Europe lowers the action levels
EU Directive 2003/10/EC, adopted on 6 February 2003 and due in national law by 15 February 2006, replaced a 1986 directive. Employers must offer hearing protectors above a daily exposure of 80 dB(A), make workers wear them at 85 dB(A), and keep the sound reaching the ear, counting the protectors, below 87 dB(A), with matching peak limits of 135, 137 and 140 dB(C).
Credit
European Parliament and Council of the EU
2011
Quieter machines lose to earplugs
OSHA's noise rules tell employers above the limit to use feasible engineering or administrative controls first, but OSHA's enforcement let employers rely on hearing tests and protectors instead. In October 2010 OSHA proposed to read "feasible" as "capable of being done"; it withdrew the proposal on 19 January 2011, three months later, citing resources, outreach and cost, although its head, David Michaels, had said that hearing protection alone cannot prevent preventable hearing loss.
LessonA rule that ranks quieter machines first still ends with earplugs first if enforcement accepts them.
Credit
OSHA
2014
Phone apps put to test
Kardous and Shaw of NIOSH tested smartphone sound-measuring apps against a reference sound level meter. Of more than 130 iOS apps, 10 met their criteria, and 4 Android apps were tested; two apps came within 0.07 dB (unweighted) and 0.52 dB (A-weighted) of the reference on average and two more within 2 dB. A 2016 follow-up found that external calibrated microphones brought four apps within 1 dB. NIOSH then released its own iOS Sound Level Meter app, which it says is accurate within 2 dBA.
Credit
C. A. Kardous and P. B. Shaw, NIOSH
Watch for
Earbuds as hearing protectors
Apple says the Hearing Protection feature of AirPods Pro 2 and AirPods Pro 3 reduces loud environmental noise by default in every listening mode, for sustained sound up to 110 dBA, and that the feature reached EU and UK users "following certification for those regions". It warns that protection depends on fit and asks users to run its seal test first.
Achieved whenan independent laboratory publishes the attenuation of AirPods Pro 3 measured with untrained users fitting their own earbuds, the subject-fit method of the ANSI S12.6 standard, and names the certificate and standard behind the EU and UK approval.
Claimed by
Apple, September 2025
Open
01
Test earbud seals on manikins
A bench test of consumer earbuds sold with a hearing-protection mode, starting with AirPods Pro 3. Fit each earbud in the ear of an acoustic test manikin at every tip size and at shallow, normal and deep insertion, play pink noise at 85 dBA with nobody in the room, and record both the noise reduction in each octave band and whether the phone's own seal test passed that fit.
The briefCredit · unclaimed02
Check tool makers' noise labels
An open table of the noise figures that makers declare in the manuals of current power tools, matched to independent measurements of the same models where they exist, with the gap between declared and measured noise for each.
The briefCredit · unclaimed
Open 01 · the brief
Test earbud seals on manikins
The first to publish is credited on this line, and the deliverable moves from Open to Achieved.
What to build
A bench test of consumer earbuds sold with a hearing-protection mode, starting with AirPods Pro 3. Fit each earbud in the ear of an acoustic test manikin at every tip size and at shallow, normal and deep insertion, play pink noise at 85 dBA with nobody in the room, and record both the noise reduction in each octave band and whether the phone's own seal test passed that fit.
Why it is still open
Apple says protection depends on fit but publishes no figures for poor fits. The closest independent test, by Chong-White, Mejia and Edwards in 2023, measured AirPods Pro 1 and 2 on a Brüel & Kjær head and torso simulator only after Apple's fit test had confirmed a seal, and found an average reduction of 27 dB for the Pro 2. A search found no independent test of the Pro 3, or of how often a passed seal test hides a poor fit.
Built with what exists
A head and torso simulator with an IEC 60318-4 ear simulator (Brüel & Kjær 4128C or GRAS KEMAR, found in many university hearing labs), a calibrated loudspeaker and sound level meter, the earbuds and their phone app, and Python with NumPy and SciPy for octave-band analysis.
First step
Book an afternoon on a university head and torso simulator and measure one AirPods Pro 3 pair at three insertion depths, logging the seal-test result for each.
Done looks like
At least three earbud models, each at three tip sizes and three insertion depths with five repeats, published with raw recordings, octave-band attenuation, and the share of fits that passed the seal test while losing more than 10 dB against the best fit.
Open 02 · the brief
Check tool makers' noise labels
The first to publish is credited on this line, and the deliverable moves from Open to Achieved.
What to build
An open table of the noise figures that makers declare in the manuals of current power tools, matched to independent measurements of the same models where they exist, with the gap between declared and measured noise for each.
Why it is still open
EU machinery law requires every manual to state the A-weighted sound pressure level at workstations when it exceeds 70 dB(A), and the sound power level when that pressure level exceeds 80 dB(A). NIOSH's power tools table measured sound power both cutting and running free, and most of its circular saws were louder cutting, one by 12 dB (108 against 96 dBA); by Wikipedia's count the database holds 166 tools from 18 makers. A European market surveillance survey, NOMAD, found that 80% of the noise declarations in more than 1,500 instruction manuals did not meet the Machinery Directive's requirements, but it judged what the manuals said against the law, not against measurements. A search found no public table that sets declared figures beside measured ones.
publications.europa.eudael.euracoustics.orgstacks.cdc.goven.wikipedia.org
Built with what exists
Makers' manual PDFs, which are public; a language model that reads documents and returns each figure with its page, such as Claude (Notes: "The jobs agents repeat"); the NIOSH table; Python for matching model numbers.
First step
Download the manuals of 30 angle grinders sold in the EU, copy their declared figures by hand, and compare them with what the model extracts.
Done looks like
Declared sound pressure, sound power, measurement uncertainty and test standard for at least 500 tool models, each linked to its manual page; at least 98% of 100 randomly checked values correct; and for every model NIOSH also measured, the declared minus measured difference.