Jerry L. Punch

Methods and Applications of the Audibility Index in Hearing Aid Selection and Fitting

During the first half of the 20th century, communications engineers at Bell Telephone Laboratories developed the articulation model for predicting speech intelligibility transmitted through different telecommunication devices under varying electroacoustic conditions. The profession of audiology adopted this model and its quantitative aspects, known as the Articulation Index and Speech Intelligibility Index, and applied these indices to the prediction of unaided and aided speech intelligibility in hearing-impaired listeners. Over time, the calculation methods of these indices-referred to collectively in this paper as the Audibility Index-have been continually refined and simplified for clinical use. This article provides (1) an overview of the basic principles and the calculation methods of the Audibility Index, the Speech Transmission Index and related indices, as well as the Speech Recognition Sensitivity Model, (2) a review of the literature on using the Audibility Index to predict speech intelligibility of hearing-impaired listeners, (3) a review of the literature on the applicability of the Audibility Index to the selection and fitting of hearing aids, and (4) a discussion of future scientific needs and clinical applications of the Audibility Index.

Nonsense-syllable recognition in noise using monaural and binaural listening strategies

Using a binaurally equipped KEMAR manikin, syllables of the CUNY Nonsense Syllable Test were recorded in sound field at 0-degree azimuth against a background of cafeteria noise at 270-degrees azimuth, at several signal-to-noise (S/N) ratios. The combination of inputs recorded at each ear was delivered to ten normal-hearing (NH) and eight sensorineurally hearing impaired (HI) listeners through insert ear phones to produce five experimental listening conditions: (1) binaural head shadow (HS), in which ear presentation was analogous to the original stimulus recording, (2) binaural favorable (BF), in which the noise-shadowed (right-ear) recording was presented to both ears, (3) monaural favorable (MF), in which the noise-shadowed recording was presented only to the right ear, (4) monoaural unfavorable (MU), in which the noise-unshadowed (left ear) recording was presented only to the left ear, and (5) simulated monoaural aided (SMA), in which the noise-shadowed recording was presented to the right ear and the noise-unshadowed recording--attenuated by 20 dB relative to the HS condition--was presented to the left ear. All main effects (subject type, listening condition, and S/N ratio) were statistically significant. Normal listeners showed 3.3- and 3.2-dB advantages, respectively, due to head-shadow and binaural squelch, over hearing-impaired listeners. Some hearing-impaired listeners performed better under the SMA or BF conditions than under the HS condition. Potential digital signal processing strategies designed to optimize speech understanding under binaurally aided listening conditions are discussed.

Evaluation of three strategies for fitting hearing aids binaurally.

Three strategies for evaluating optimum frequency shaping and noise reduction in binaural digital hearing aids were compared in a repeated-measures design, using a new preference-based prescriptive fitting method. These strategies consisted of using preferred frequency shaping and noise reduction values binaurally: (1) based on monaural testing; (2) based on separate evaluations of each ear; and (3) based on evaluation of a second ear while subjects wore an aid programmed with the preferred values in the first ear. Individually preferred characteristics were programmed for 17 hearing-impaired subjects, most of whom exhibited symmetrical sensorineural hearing loss. Each subject was administered intelligibility estimation and midplane localization measurements in the laboratory, as well as a questionnaire survey based on situational listening in the real world. No statistically significant differences in preferences for either frequency shaping or noise reduction were found for the three fitting strategies, suggesting that monaural testing is sufficient in symmetrical cases to provide information for binaural fitting. Related to this finding, differences across binaural conditions were minimal for both intelligibility estimation and localization results. A significant improvement in localization performance under binaural conditions over monaural listening, however, was documented by both the laboratory and the real world data. A strong overall preference for binaural over monaural amplification was also documented under real world conditions.

Speech-clarity judgments of hearing-aid-processed speech in noise: Differing polar patterns and acoustic environments

This investigation assessed the extent to which listeners’ preferences for hearing aid microphone polar patterns vary across listening environments, and whether normal-hearing and inexperienced and experienced hearing-impaired listeners differ in such preferences. Paired-comparison judgments of speech clarity (i.e. subjective speech intelligibility) were made monaurally for recordings of speech in noise processed by a commercially available hearing aid programmed with an omnidirectional and two directional polar patterns (cardioid and hypercardioid). Testing environments included a sound-treated room, a living room, and a classroom. Polar-pattern preferences were highly reliable and agreed closely across all three groups of listeners. All groups preferred listening in the sound-treated room over listening in the living room, and preferred listening in the living room over listening in the classroom. Each group preferred the directional patterns to the omnidirectional pattern in all room conditions. We observed no differences in preference judgments between the two directional patterns or between hearing-impaired listeners’ extent of amplification experience. Overall, findings indicate that listeners perceived qualitative benefits from microphones having directional polar patterns.

The effects of training format on earplug performance

This experiment investigated the effect of small-group versus individual hearing loss prevention (HLP) training on the attenuation performance of passive insert-type hearing protection devices (HPDs). A subject-fit (SF) methodology, which gave naive listeners access only to the instructions printed on the HPD product label, was used to determine real-ear attenuation at threshold (REAT) at third-octave noise bands between 125–8000 Hz. REAT measurements were augmented by use of the Hearing Loss Prevention Attitude-Belief (HLPAB) survey, a field-tested self-assessment tool developed by the National Institute for Occupational Safety and Health (NIOSH). Participants were randomly assigned to one of four experimental groups, consisting of 25 listeners each, in a controlled behavioral-intervention trial. There were two types of HPDs (formable and premolded) and two training formats (individual and small group). A short multimedia program, including a practice session, was presented to all 100 listeners. Results showed training to have a significant effect, for both HPDs on real-ear attenuation and attitude, but, importantly, there was no difference between small-group and individual training.

Listener‐assessed intelligibility of hearing‐aid‐processed speech

The capacity of listeners to yield reliable and valid preferences for hearing aid-processed speech was evaluated by the method of paired comparisons. Discourse by a male talker with General American dialect was processed by five hearing aids under conditions of quiet and a background of multitalker babble. Hearing aid-processed Revised Central Institute for the Deaf Sentences spoken by the same male talker and embedded in the same multitalker background were used as stimuli in establishing criterion intelligibility performance. All stimuli were tape-recorded and delivered via monaural earphone to 90 normal listeners. Test-retest results revealed reliability coefficients of 0.94 and 0.65 on the paired-comparison condition in quiet and in babble, respectively. A low-positive relationship was observed between responses on the two subjective tasks. Essentially negligible correlations were found between subjective preferences on these individual conditions and performance on the sentence task. General conditions are hypothesized under which paired-comparison judgments of intelligibility might be expected to correspond favorably with objective performance.

Hearing International Project on High-Quality, Low-Cost Hearing Aids

Within the Hearing International (HI) organization, a subcommittee of the Committee on Management and Rehabilitation of Hearing Loss was formed several years ago to facilitate the manufacture and commercial availability of one or more high-quality, low-cost hearing aids. The project is named the High- Quality Low-Cost Hearing Aid Project, and its aim is to facilitate the development of such aids for use in developing countries. Actions taken thus far by the subcommittee include meetings and discussions with representatives of leading hearing aid manufacturers, an analysis of hearing aid features and their costs, and an analysis of the opinions of hearing aid manufacturers toward the development of fully functional low-cost hearing aids. The latter two activities involved surveys of the -characteristics of contemporary hearing aids and of hearing aid manufacturers. Findings from these surveys provide the bases for much of the information in this chapter. While this information may be of interest to many laypersons, it should be of special interest to those involved in the pursuit of the goal of facilitating the manufacture of high-quality, low-cost hearing aids. Background information is provided in the previous chapter.

High-Quality, Low-Cost Hearing Aids

At least one-tenth of the world’s population has significant hearing impairment. The vast majority of cases of hearing impairment are caused by disorders of the inner ear, a condition referred to as sensorineural hearing loss. This common form of hearing loss is typically the result of inner-ear damage or degeneration, is usually not resolved through medical treatment, and is usually permanent in nature. Although sensorineural hearing loss accompanying aging, or presbycusis, is primarily responsible for the high prevalence of hearing impairment, it may also be caused by other factors. For example, it may be genetic, appearing at birth or later in life, or it may be caused by a variety of other factors, including birth trauma, traumatic injury after birth, intake of dangerous chemicals or drugs, or overexposure to loud noise. The impact of sensorineural hearing loss can be devastating. In fact, it has been linked to a variety of physical and psychosocial dysfunctions, including isolation, dependency, hypertension and stress, depression, and a greater prevalence of illness in general. In children, it can result in delayed language and speech development, as well as poor progress in school.

Signal averaging in real ear probe tube measurements.

In this study, signal averaging measurements were performed using a broadband noise as the test signal in two acoustic environments typifying relatively noisy situations in which real ear probe tube measurements are made clinically. The minimum time-domain signal averaging required to reduce the degrading effects of background noise to levels acceptable for the production of valid real ear unaided responses (REURs) was established for three normal listeners at various signal-to-noise (S/N) ratios. Generally, averaging was not required at S/Ns of 20 dB or higher. Acceptable responses were obtained at S/Ns as low as -5 dB when the number of averages was increased to 256 and higher. The function derived from the data in the two acoustic environments was in good agreement with a theoretical rule. Findings suggest that averaging of only a few sweeps of the test signal is required to obtain reliable and valid real ear measurements, even in somewhat noisy test environments. With a knowledge of the levels of the test signal and the background noise, data from this study can be used to determine the requisite number of signal sweeps to be averaged in acquiring real ear responses in a minimum amount of time.