Real-ear polar patterns and aided directional sensitivity.

Real-ear directional sensitivity patterns were obtained from a group of eight subjects wearing in-the-ear and completely-in-the-canal-placed microphones to determine whether depth of microphone placement within the ear affects electroacoustic estimates of directionality. Both conventional and extended versions of the Unidirectional Index and the Directivity Index were used to estimate directional sensitivity, and results indicated that depth of microphone placement affects directionality in a region-specific as well as a frequency-specific manner. Results also indicated that conclusions about directional sensitivity vary when assumptions about signal and noise source locations, the orientation of the listener relative to them, and human perceptual abilities are taken into consideration. Results are discussed in terms of the utility of modified estimates of directional sensitivity to quantify directional sensitivity when measurements are made in real ears and on possible localization and speech recognition effects that may vary with depth of microphone placement in a regionally specific manner.

Acoustic and hearing aid circuit variables affecting the tolerability of aided impulsive-type sounds.

The tolerability of aided impulsive-type stimuli was investigated in a group of 13 hearing-impaired listeners. Two linear circuits (one with a class A and one with a class D output stage) and one adaptive frequency response (AFR) circuit (with a class D output stage) were investigated. In a three-way paired-comparison task, subjects chose the hearing aid that was most tolerable when 75 dB sound pressure level (SPL) impulsive-type sounds were presented. Real-ear measurements of rms SPL, peak SPL, crest factor, and spectral distribution were made to determine which of these variables was most closely associated with behavioral tolerability scores. Results indicated significant differences across hearing aids for tolerability scores, rms sound pressure levels, and spectral peak frequencies. Highest tolerability scores were associated with the hearing aid that produced the lowest rms sound pressure levels in the ear canal (class D AFR). Significant correlations were found between tolerability and both rms SPL and peak SPL. Results are discussed in terms of circuit algorithm and in terms of the possible effects of hearing aid saturation.

A new technique for quantifying temporal envelope contrasts.

A new technique has been developed for precisely quantifying the temporal contrasts that exist between two sound samples. This technique is based on envelope subtraction, and generates an Envelope Difference Index that may be used to help clarify whether alteration of the natural speech envelope via amplification improves or degrades speech intelligibility. The Envelope Difference Index method may also be used to assess hearing aid saturation, and may have other applications as well. The technique is applicable whenever a precise quantification of the difference between two temporal envelopes is required, regardless of stimulus duration.

Automated Methods of Assessing Hearing Aid Performance in Listeners With Impaired Hearing.

The methods that have been described, although still under development, are intended to demonstrate that the issues of sound quality, speech intelligibility, and loudness may be addressed in a clinical setting in a reasonable amount of time. Being automated, these methods are relatively time-efficient, and may become more so as they are refined. Being database oriented, these methods make it easy to monitor individual clients over time, and also provide an efficient way to evaluate the success rate of individual circuits across listeners with similar hearing losses. Individually, these techniques provide only some of the information that may indicate whether a particular fitting may succeed. Taken together, these behavioral methods will provide a great deal of both subjective and objective information that will help the dispenser not only decide which of several hearing aids may be best for a user, but also how a given instrument may be adjusted to provide the maximum benefit. By including the formal assessment of sound quality, speech intelligibility, and loudness in the hearing aid evaluation, the needs of the user are more likely to be met, and the degree of satisfaction achieved is likely to be higher than if the evaluation consists solely of matching the real ear insertion response to a prescribed target.

Hearing aid saturation, coherence, and aided loudness discomfort.

Real-ear measurements of the aided Loudness Discomfort Level (LDL) were obtained from five hearing-impaired listeners who were fit with two Class A and two Class D linear hearing aids, each with a different saturation sound pressure level (HFA SSPL90). These measurements were obtained with 75 dB SPL continuous discourse to determine whether saturation-induced distortion contributes to the sensation of loudness. Real-ear coherence measurements made at LDL were used to determine the extent of saturation, and sound quality judgments were used to determine whether the distortion present at LDL affected sound quality. Results indicated that the SPL, coherence, and sound quality ratings obtained at LDL were all higher for Class D hearing aids with relatively high HFA SSPL90 than for Class A hearing aids with relatively low HFA SSPL90. Overall results were generally consistent with the hypothesis that distortion affects both sound quality and the perception of loudness.

Hearing aid saturation and aided loudness discomfort.

Clinical measurements of the loudness discomfort level (LDL) are generally performed while the subject listens to a particular stimulus presented from an audiometer through headphones (AUD-HP). The assumption in clinical practice has been that the sound pressure level (SPL) corresponding to the sensation of loudness discomfort under AUD-HP conditions will be the same as the corresponding to LDL with the hearing aid. This assumption ignores the fact that the distortion produced by a saturating hearing aid could have an influence on the sensation of loudness. To examine these issues, 5 hearing-impaired subjects were each fit with four linear hearing aids, each having a different saturation sound pressure level (SSPL90). Probe-tube microphone measurements of ear canal SPL at LDL were made while the subjects listened to continuous discourse in quiet under aided and AUD-HP conditions. Also using continuous discourse, real-ear coherence measures were made at various output sound pressure levels near LDL. All four hearing aid types produced mean LDLs that were lower than those obtained under AUD-HP conditions. Those hearing aids with higher SSPL90 produced significantly higher LDLs than hearing aids with lower SSPL90. A significant negative correlation was found between real-ear SPL and real-ear coherence. Quality judgments made at LDL indicated that sound quality of hearing aids with higher SSPL90 was preferred to that of hearing aids with lower SSPL90. Possible fitting implications regarding the setting of SSPL90 from AUD-HP LDL measures are discussed.

Strategies for enhancing the consonant to vowel intensity ratio with in the ear hearing aids.

Numerous investigators have suggested that increasing the consonant to vowel intensity ratio (CVR) may improve speech intelligibility. This investigation was designed to determine the extent to which analog circuits, small enough to fit into in the ear hearing aids, can increase the CVR, and whether CVR enhancement is of benefit to hearing-impaired listeners. Real ear CVRs, calculated from real ear recordings of nonsense syllables, were obtained from eight hearing-impaired listeners. Recordings from each listener were obtained through each of four hearing aid circuits: (1) an adaptive high-pass filter; (2) a faster acting adaptive high-pass filter; (3) the fast-acting adaptive high-pass filter with expansion; and (4) an infinite amplitude clipper. The amount of CVR enhancement was compared to performance of the subjects with a NST speech recognition task. Subjects also ranked the four circuits for amount of consonant emphasis provided. Results indicated that the four hearing aid circuits increased the real ear CVR by 4 to 6 dB, relative to unaided. Aided CVR varied, however, across circuits and between fricative and stop consonants. Performance on the NST recognition task was generally consistent with the amount of CVR increase provided. Rank ordering for consonant emphasis was consistent with aided CVR for stop consonants, but not for fricatives.

High-level psychophysical tuning curves: simultaneous masking by pure tones and 100-Hz-wide noise bands.

Simultaneous-masked psychophysical tuning curves were obtained from normal-hearing listeners using low-level (20-25 dB SPL) probe tones in quiet and high-level (60 dB SPL) probe tones, both in quiet and in the presence of a broad-band background noise. The background noise was introduced to eliminate combination tones or combination bands and other off-frequency listening cues that exist at high levels. Tuning curves were obtained using pure-tone maskers and 100-Hz-wide narrow-band noise maskers for probe tones at 1000 and 4000 Hz. High-level tuning curves for pure-tone maskers demonstrated large discontinuities or "notches" on the low-frequency sides of the tuning curves. Broad-band background noise eliminated those notches, indicating that the notches were due to the detection of off-frequency listening cues at combination-tone frequencies. High-level tuning curves for 100-Hz-wide narrow-band maskers also demonstrated notches on the low-frequency sides. Those notches were eliminated with broad-band background noise, which indicates that combination bands strongly influenced the shapes of high-level tuning curves obtained with narrow-band maskers. The influence of combination bands was dependent upon test frequency. At 1000 Hz, combination bands had very little influence on the shapes of high-level tuning curves. At 4000 Hz, where the masker bandwidth was substantially less than the critical bandwidth, combination bands strongly affected the low-frequency sides of the tuning curves. In 2 subjects tested at a probe frequency of 2000 Hz with 100-Hz-wide masking bands, combination bands also influenced the low-frequency sides of high-level tuning curves. The presence of combination-tone or combination-band cues essentially steepened the low-frequency slopes of tuning curves, resulting in sharper estimates of tuning. Comparisons of tuning curves obtained with pure-tone maskers and narrow-band maskers, in the same listeners, revealed that pure-tone maskers were more effective than narrow-band maskers when the masker frequencies were in the tail region of the tuning curve. The results of these experiments support the notion that tuning in the normal auditory system broadens notably with stimulus level, once off-frequency listening cues such as combination tones or combination bands are eliminated. The low-level simultaneously masked tuning curve demonstrates a sharp bandpass tuning characteristic, whereas the high-level simultaneously masked tuning curve in background noise demonstrates a broad low-pass tuning characteristic.(ABSTRACT TRUNCATED AT 400 WORDS)

High-level psychophysical tuning curves: simultaneous masking with different noise bandwidths.

Simultaneous-masked psychophysical tuning curves were measured with narrow-band noise maskers varying in bandwidth from 40 Hz to 800 Hz to determine the masker bandwidths at which combination-band detection cues no longer influence tuning-curve shapes. Tuning curves were obtained at 1000 and 4000 Hz from normal-hearing listeners using high-level (60 dB SPL) probe tones in quiet and in the presence of a broadband background noise to eliminate combination bands and other off-frequency listening cues that exist at high levels. High-level tuning curves revealed notches on the low-frequency sides. Those notches were eliminated with broad-band background noise, which indicates that combination bands can strongly influence the shapes of high-level tuning curves obtained with narrow-band maskers, primarily by steepening the low-frequency and tail slopes. Combination-band detection cues had a stronger influence at 4000 Hz than at 1000 Hz. As masker bandwidth increased, combination bands had less influence on tuning-curve shapes. These results suggest a possible relation between masker bandwidth and auditory critical bandwidth: combination bands affected the low-frequency sides of the tuning curves only when the masker bandwidth was less than the auditory critical bandwidth.