A rating system in search of hearing protectors.
Murphy WJ; Franks JR
J Acoust Soc Am 1999 Feb; 105(2):1131
Previous methods for rating the noise reduction of hearing protection devices (HPDs) have determined the mean attenuation and standard deviations at several octave-band frequencies using real-ear attenuation at threshold (REAT) measurements. Typically, some fraction of the standard deviation is subtracted from the mean attenuation to estimate an assumed protected value for a given percentage of the population. The subtractive correction assumes normally distributed numbers that are linear and not logarithmic. Data from several HPDs were analyzed with a statistical approach which first determined the A-weighted noise reduction for individual measurements. The cumulative distribution of A-weighted noise reduction (CDNRA) was then fit with a combination of two logistic regression curves. The Pearsons correlation coefficients for the model and the data were better than 0.99 for each of four devices analyzed. The CDNRA method yielded reasonable results from data collected from well-fit and poorly fit hearing protectors. In contrast, the NRR method would yield negative NRR values for poorly fit protectors. This paper describes the CDNRA method and looks at the results for an interlaboratory test of four HPDs, as well as additional data collected in the NIOSH Hearing Protector Laboratory.
Hearing-level; Hearing-loss; Hearing-acuity; Health-surveys; Age-factors; Age-groups; Information-retrieval-systems; Physical-examination; Racial-factors; Sex-factors; Audiological-testing; Audiometry
William J. Murphy, Bioacoustics and Occupational Vibration Section, NIOSH, MS C-27, 4676 Columbia Pkwy., Cincinnati, OH 45226-1998
Journal of the Acoustical Society of America