Hearing protection devices (HPDs) are, as a practical matter, the first line of defense against hearing loss caused by excessive noise. Other ways to reduce exposure to loud sound (engineering control of noise sources, reduction of noise in the transmission path between a source and an individual) can indeed be more effective, but are often more costly or more difficult to manage.

HPDs can be classified by type (active versus passive), by form (e.g., earplugs and earmuffs of various types), and by effect. In all cases the goal is the same: to attenuate the magnitude of sound reaching the cochlea, thus limiting acoustical insult to the end-organ of hearing (see noise-induced hearing loss). Passive devices accomplish this through blockage of the airborne transmission path to the inner ear. Active devices seek to mechanically or electronically respond to noise to reduce signal amplitudes presented to the auditory system.

Earplugs fall into five categories (Berger, 2000): (1) closed-cell foam devices designed to be manually compressed, then inserted into the ear canal, where they expand to approximate their initial size (e.g., the Aearo Company E-A-R plug), (2) preformed devices available in different diameters to accommodate different ear canals (e.g., the PlastiMed V-51R plug), (3) malleable devices intended to fit a range of ear canals, (4) semi-insert devices held in the ear canal by means of a plastic or metal band, and (5) devices made from ear mold impressions taken from individual ear canals. Most earplugs are created from plastics (polyvinyl chloride, polyurethane, silicone, or acrylic); malleable earplugs often consist of wax-impregnated cotton or fiberglass enclosed in a thin plastic container. Most earplugs are passive, that is, they are not intended to respond differently to differing noise exposures. Some active earplugs employ metal slugs or other material intended to move within the plug when stimulated by sudden acoustic overpressures, thus increasing attenuation in response to impulsive noise. For various reasons, it is difficult or impossible to objectively measure the attenuation of such devices or to estimate their real-world benefit. A recent addition to the array of earplugs are those offered by Etymotic Research for users with specific needs (e.g., musicians) who seek flat attenuation across specified frequency ranges.

Earmuffs are designed as integral components of safety helmets or as separate devices that surround the outer ear and are held in place by headbands that extend over or behind the head or beneath the chin. Some of these are combined with communication systems intended to increase the signal-to-noise ratio of messages electronically routed to earphones placed within headsets. At present, most earmuffs are passive devices, but several have been developed as active systems. Indeed, most active hearing protectors are based on earmuffs, owing to the space required for sound sensing and processing components. Active protectors employ one of two (or both) methods to attenuate sound. One method senses sound with a microphone, then processes sound delivered through an earphone by means of automatic gain control circuitry: when incident sound exceeds a certain level, further increases are electronically clipped or otherwise squelched. The other method samples incident sound, reverses the phase of the signal, and electronically adds the reversed signal within the muff enclosure to partially cancel the incident sound. Because of incident signal changes and processing speed requirements, devices employing additive cancellation techniques are more effective at relatively low frequencies (e.g., below 500 Hz; see Nixon, McKinley, and Steuver, 1992). Some active noise reduction methods appear similar to (or may benefit from) methods employed in hearing aids.

Beyond type and form, HPDs differ in weight, comfort, uniformity of fit to individuals, compatibility with other protective or prosthetic devices, and compatibility with individual user health status. Using eyeglasses with earmuffs, for example, can create acoustic leaks that reduce attenuation performance. Similarly, a subject with excessive cerumen or a middle ear effusion should not use earplugs. Use of hearing protectors in hot, humid environments can be uncomfortable and can cause skin irritation. If hearing protectors (perhaps combined with hearing loss) render speech communication difficult, or if they limit audibility of other signals deemed important, users may reject them. For obvious reasons, earmuffs should not be used in conjunction with hearing aids. These and other issues are discussed in detail by Berger (2000).

Real environments in which hearing protectors might provide benefit differ tremendously in noise amplitude, spectrum, and duration. Noise exposure is normally indexed by time-weighted average (TWA) sound pressure levels sampled using integrating meters or personal noise dosimeters. In the United States, such measurements are specified by Federal regulation (Occupational Safety and Health Administration [OSHA], 1983, CFR Part 1910.95) and the subject of technical standards (American National Standards Institute [ANSI], S12.19 1996). Among other details, exposure is to be indexed using a slow meter ballistic characteristic and an A-weighting network (a high-pass filter useful in predicting the effects of broadband noise on hearing). TWA levels are single-number values used to describe noise exposure and determine actions to protect workers from noise-induced hearing loss in the workplace (OSHA, 1983).

In 1979, the Environmental Protection Agency (EPA) issued a regulation intended to promote laboratory measurement of hearing protector attenuation for the purpose of combining such information with exposure data to estimate protective effect. The EPA regulation (CFR40 Part 211) built on previous technical standards (ANSI, 1974), and invoked a single-number index, called the Noise Reduction Rating (NRR), to be included in hearing protector product labels. Computation of NRRs from averaged behavioral real-ear attenuation-at-threshold (REAT) data assume temporally continuous band-limited noise stimuli with equal energy per octave (pink noise), and address intersubject variability by doubling the standard deviation of threshold shifts, then subtracting that value from mean threshold shift for each noise band. Adjusted attenuation values are summed logarithmically across stimulus noise bands to yield an NRR in decibels. Because the NRR method also assumes measurement of unprotected levels indexed with a C-weighting network (which has a flatter frequency response than the A-weighting network used to measure exposure), an additional 7 dB must be subtracted from the NRR to estimate A-weighted noise levels when a hearing protector is in place (see OSHA, 1983, Appendix B). Finally, because it was recognized that how hearing protectors are placed in a subject's ears (plugs) or on a subject's head (muffs) could affect outcomes, the EPA method specified experimenter fitting of HPDs during laboratory testing.

Because real-ear attenuation methods performed following the procedures stipulated by EPA designate experimenter fitting of HPDs, it is to be expected that resulting NRRs will be larger than what would be found with subject fitting of HPDs. Because all extant methods for measuring REATs use temporally continuous noise, results of such measurements cannot be generalized to impulse noise (e.g., gunfire) or impact noise (e.g., forging).

Shortly after the inception of the current OSHA Hearing Conservation Rule (OSHA, 1983), the National Institute of Occupational Safety and Health (NIOSH) recommended that labeled NRRs be derated to estimate effectiveness in the field. Six schemes are noted in Appendix B of the Hearing Conservation Rule. These differ based on available measurement devices and data, but generally reduce the estimated benefit of hearing protectors. For example, if only A-weighted noise exposure data are available, 7 dB is subtracted from the NRR. For both A-weighted and C-weighted exposure data, the resulting corrected NRR is further reduced by 50%.

Subsequent research over two decades suggests that NRRs derated in this manner still overestimate the attenuation of hearing protectors in real-world situations. Various factors contribute to this inaccuracy, including overestimation associated with (1) experimenter fit, (2) highly trained test subjects (whose small standard deviations of REATs produce higher NRRs), and (3) differences in patterns of use of hearing protectors in laboratory and field settings (NIOSH, 1998).

Other pertinent generalizations include the following: (1) overall, earmuffs provide the most protection, foam and formable earplugs provide the next greatest protection, and all other insert types provide less, and (2) ideally, individuals should be fitted individually for hearing protectors (NIOSH, 1998). Generally, both earplugs and earmuffs provide greater attenuation at frequencies above 500 Hz than at lower frequencies (Berger, 2000).

Chapter 4 of the revised NIOSH criteria document (NIOSH, 1998) offers details about estimated real-world NRRs for 84% of wearers of hearing protectors, based on several independent studies. Labeled NRRs for single protectors range from 11 to 29 dB, while weighted mean NRR84 values range from 0.1 to 14.3 dB.

To address these problems, existing standards for measuring HPD attenuation were revised (Royster et al., 1996; Berger et al., 1998) to include subject-fit methods with audiometrically proficient listeners naive about HPDs. The resulting standard is ANSI S12.6 (1997). (A companion standard, ANSI S12.42 (1995), specifies a test fixture method and a microphone-in-real-ear method for measuring insertion loss useful for quality control and product development work with earmuffs.)

In 1995 the National Hearing Conservation Association proposed alternative labeling requirements in which only subject-fit real-ear attenuation data (ANSI S12.6-1997, Method B) are reported. The revised NRR(SF) information generally suggests less protection than NRRs based on experimenter fitting. Alternatively, the NHCA (1995) suggests labeling to include high, medium, and low NRRs based on statistical distributions of measured subject-fit REATs. This proposal has been endorsed by several other organizations. As of this writing, however, the EPA NRR labeling requirement remains based on the experimenter-fit method specified in ANSI S3.19 (1974).

If only experimenter-fit data are available, NIOSH (1998) currently recommends derating of NRRs based on type of hearing protector: 25% for earmuffs, 50% for formable earplugs, and 70% for all other earplugs. In the case of double protection (plugs and muffs), the OSHA Technical Manual (OSHA, 1999) recommends using the EPA NRR for the better protector, minus 7 dB, dividing the result by 2 (a 50% derating), then adding 5 dB to the field-adjusted NRR to account for the second protector.

Rather clearly, much work remains to be done to improve the prediction of real-world benefit of hearing protectors (Berger and Lindgren, 1992; Berger, 1999). One promising approach involves methods similar to in vivo real-ear gain measurements of hearing aids (now a common practice), together with modification of commonly used personal noise dosimeters. This approach requires the ability to simultaneously measure exposure level and the sound level generated within the ear canal of the wearer of a hearing protector. If both are measured with the same filtering schemes (preferably, the C-weighting network; ideally with both A and C networks), the signed difference between the two would index attenuation due to the hearing protector. If such measurements can be adapted to field use (e.g., with a two-channel noise dosimeter), it may be possible to add useful information to what otherwise can be determined about the performance of at least some HPDs. ANSI S12.42 (1995) addresses some of these issues for earmuffs and communication headsets, but only for laboratory measurements. Because the use of probe microphones with earplugs is likely to produce reactive measurement effects, this approach may not be suitable for insert devices.

It is generally recognized that effective use of hearing protectors in the workplace or elsewhere is influenced by factors that go beyond the physical performance of these devices. As summarized by NIOSH (1998), these factors include convenience and availability, comfort and ease of fit, compatibility with other safety equipment, and worker belief that the device can be worn effectively, will indeed prevent hearing loss, and will still permit hearing of important sounds.