Definition, scope and levels of evidence
Earplug noise-reduction performance testing measures how much an earplug attenuates sound reaching the ear at specified frequencies and under declared fitting conditions. It produces frequency-band data, statistics or a single-number rating defined by a particular regulatory system. The subject of this entry is the testing, evaluation and reporting service; it is neither the earplug product itself nor a sleep-outcome trial. Three evidence levels must be kept separate. A controlled laboratory type test describes group attenuation for specified samples fitted according to a defined procedure. A subject-fit test using inexperienced participants is intended to better represent attenuation attainable by a typical user group. An individual fit test estimates the attenuation achieved by one person with one model in one fitting condition. None substitutes for the others. The applicable standard must also match the market, intended use and claim, whether the product is sold for occupational hearing protection, routine sleep, music, travel or a special noise environment. “Laboratory attenuation,” “wearer attenuation” and “better sleep” are not interchangeable conclusions.
Samples, fitting conditions and study design
Before commissioning a test, the product name, model, material, construction, size range, formable or premoulded design, custom status, single-use or reusable status, acoustic filter, cord and packaging version should be frozen. Production lot, sampling plan, sample count, storage and conditioning should be recorded. The protocol should identify the standard and edition, target market, participant inclusion and exclusion criteria, ear-canal and hearing screening, number of ears tested, fitting instructions, whether the participant or experimenter fits the plug, training level, repetitions, test-frequency range, sound-field and background-noise requirements, instrument calibration and invalid-data rules. Rolling, insertion depth and expansion time can materially change foam-plug results; size selection matters for premoulded plugs; and left-right identity and seal inspection matter for custom plugs. A different size, shortened or perforated plug, added filter, lubricant or previously used sample is not automatically equivalent to the original and should not be pooled unless the protocol expressly allows it.
REAT, frequency-band results and single-number ratings
The common subjective method is real-ear attenuation at threshold, or REAT. Hearing thresholds are measured for the same participant with the ear open and with the plug properly fitted, and their difference provides attenuation at each test band. Reports should retain the band means, standard deviations or other required statistics rather than publishing only the largest decibel value. Low-frequency results can be affected by masking from physiological noise in the occluded ear, and the measured values represent users only to the extent that users fit the device in the same manner as the test participants. SNR, H/M/L, octave-band estimates and the United States NRR label are produced under different rules, weightings, statistical assumptions and legal contexts. Their numerical values should not be directly ranked or relabelled. Any estimate of the A-weighted level under the plug should use a method matched to the field-noise spectrum and measurement weighting. A rating developed for steady noise does not by itself establish protection against high-peak impulsive noise such as gunfire or explosions.
Individual fit testing and sleep-scenario extensions
Individual fit testing asks whether a particular wearer can obtain sufficient attenuation with the selected plug and fitting technique and commonly reports a personal attenuation rating. Subjective systems may compare open and occluded hearing thresholds, while objective systems may compare microphone signals inside and outside a modified plug. Test frequencies, conversions and pass criteria differ among systems, so results must stay within each system’s validation scope. Testing should address both ears, refitting, targeted instruction and a size or model change where needed. It should not allow an operator to repeatedly optimise a fit and then present the best score as the user’s unaided, routine performance. For sleep use, separate extensions may assess seal after side-lying pressure, turning and jaw movement, overnight retention, attenuation after cleaning or ageing, and comfort. These are scenario-specific studies and should not be folded into a standard type rating. A short test in a quiet laboratory cannot establish all-night stability and cannot alone demonstrate faster sleep onset, fewer awakenings or improved sleep architecture.
Special products, signal audibility and safety boundaries
Earplugs with acoustic filters, level-dependent paths, communication, playback or active noise control require states to be declared separately, including passive, powered, unpowered, low-battery, each operating mode and maximum output. A passive-seal test does not cover electronic behaviour, and consumer active noise cancellation should not be called hearing protection without the corresponding evidence and label. More attenuation is not always more appropriate: over-attenuation can interfere with speech, doorbells, infant crying, fire or carbon-monoxide alarms and other signals that a user must perceive. Alarm audibility requires a separate evaluation using the actual alarm spectrum and level, room, distance, user hearing and fitting condition. It cannot be inferred directly from NRR, SNR or one frequency-band value, and a test cannot support a guarantee that an alarm will never be missed. Acoustic attenuation testing also does not cover material biocompatibility, ear-canal injury, infection, ingestion, cord entanglement, cleaning residue, flammability or battery safety. Those hazards require product-specific assessment.
Laboratory quality, production consistency and retesting
The laboratory should demonstrate that the method is within its competence scope and retain personnel authorisation, sound-field qualification, audiometric and acoustic calibration, environmental monitoring, software and calculation versions, blinded sample identifiers, raw thresholds, deviations and review records. Measurement uncertainty and inter-laboratory differences should be considered. Participant variability is not inconvenient data to be deleted; it is part of earplug fit performance. Predefined rules for invalid runs, outliers and retests should be applied transparently. Type-test samples should be traceable to normal production, while procurement acceptance should connect the result to production consistency through multi-lot sampling and control of critical dimensions and materials. Changes to formulation, density, geometry, size range, filter, surface finish, tooling, supplier, production site, package instructions or fitting diagrams require an impact assessment and, where seal, recovery or user technique may change, retesting within the applicable scope. If cleaning cycles, temperature, humidity, storage ageing or reuse are claimed conditions, conditioned or post-durability attenuation should also be verified.
Reports, procurement acceptance and claim limits
A complete report should identify the client and laboratory, sample and lot, receipt and test dates, applicable standard and edition, deviations, participant and sample counts, fitting and training conditions, equipment and calibration status, frequency-band results and dispersion, the calculation system for any single-number rating, uncertainty or limitations, authorised signatory and traceable report number. A purchaser should confirm that the reported model, size, material, instructions and production goods match and that the laboratory’s accreditation scope actually covers the method used. “Tested to” a standard does not mean every requirement of that standard was met, and a test report is not automatically a certification. Rankings across laboratories or products are meaningful only when method, fit condition, sample and rating system are comparable. Marketing may accurately repeat results and limitations from the applicable method, but must not assemble a higher number from the best band, best participant, better ear or conversions between schemes. An attenuation rating must not be extended into claims of treating insomnia, protecting every wearer, covering every noise, or guaranteeing that alarms will be heard.