Kathryn H. Arehart

Coherence and the Speech Intelligibility Index

The speech intelligibility index (SII) (ANSI S3.5-1997) provides a means for estimating speech intelligibility under conditions of additive stationary noise or bandwidth reduction. The SII concept for estimating intelligibility is extended in this paper to include broadband peak-clipping and center-clipping distortion, with the coherence between the input and output signals used to estimate the noise and distortion effects. The speech intelligibility predictions using the new procedure are compared with intelligibility scores obtained from normal-hearing and hearing-impaired subjects for conditions of additive noise and peak-clipping and center-clipping distortion. The most effective procedure divides the speech signal into low-, mid-, and high-level regions, computes the coherence SII separately for the signal segments in each region, and then estimates intelligibility from a weighted combination of the three coherence SII values.

Prevalence of spontaneous otoacoustic emissions in neonates

The prevalence of spontaneous otoacoustic emissions (SOAEs) was measured in a group of 100 neonates and in a group of 50 normal-hearing young adults. The prevalence of SOAEs in the adult group (0.62) is at the high end of the range of prevalences reported in other surveys of adult SOAEs based on measurements using similar microphones. The prevalence of SOAEs in neonates (0.64) is not significantly different from that in adults. The various tendencies that have been found to be significant in the pooled results of other surveys are also evident in our adult group: more SOAEs in right ears, a higher prevalence of SOAEs in females, and a dependence between ears for the occurrence of SOAEs. The above-mentioned tendencies are also significant in the infant data. The major differences between the infant and adult results are the predominant SOAE frequency range and the average levels of SOAEs. The majority of adult SOAEs are between 1.0 and 2.0 kHz, whereas the majority of neonatal SOAEs are between 2.5 and 5.0 kHz. The average SOAE level is -2.6 dB SPL for adults and 8.5 dB SPL for infants.

Working memory, age, and hearing loss: susceptibility to hearing aid distortion.

Objectives: Hearing aids use complex processing intended to improve speech recognition. Although many listeners benefit from such processing, it can also introduce distortion that offsets or cancels intended benefits for some individuals. The purpose of the present study was to determine the effects of cognitive ability (working memory) on individual listeners’ responses to distortion caused by frequency compression applied to noisy speech. Design: The present study analyzed a large data set of intelligibility scores for frequency-compressed speech presented in quiet and at a range of signal-to-babble ratios. The intelligibility data set was based on scores from 26 adults with hearing loss with ages ranging from 62 to 92 years. The listeners were grouped based on working memory ability. The amount of signal modification (distortion) caused by frequency compression and noise was measured using a sound quality metric. Analysis of variance and hierarchical linear modeling were used to identify meaningful differences between subject groups as a function of signal distortion caused by frequency compression and noise. Results: Working memory was a significant factor in listeners’ intelligibility of sentences presented in babble noise and processed with frequency compression based on sinusoidal modeling. At maximum signal modification (caused by both frequency compression and babble noise), the factor of working memory (when controlling for age and hearing loss) accounted for 29.3% of the variance in intelligibility scores. Combining working memory, age, and hearing loss accounted for a total of 47.5% of the variability in intelligibility scores. Furthermore, as the total amount of signal distortion increased, listeners with higher working memory performed better on the intelligibility task than listeners with lower working memory did. Conclusions: Working memory is a significant factor in listeners’ responses to total signal distortion caused by cumulative effects of babble noise and frequency compression implemented with sinusoidal modeling. These results, together with other studies focused on wide-dynamic range compression, suggest that older listeners with hearing loss and poor working memory are more susceptible to distortions caused by at least some types of hearing aid signal-processing algorithms and by noise, and that this increased susceptibility should be considered in the hearing aid fitting process.

Effects of noise and distortion on speech quality judgments in normal-hearing and hearing-impaired listeners.

Noise and distortion reduce speech intelligibility and quality in audio devices such as hearing aids. This study investigates the perception and prediction of sound quality by both normal-hearing and hearing-impaired subjects for conditions of noise and distortion related to those found in hearing aids. Stimuli were sentences subjected to three kinds of distortion (additive noise, peak clipping, and center clipping), with eight levels of degradation for each distortion type. The subjects performed paired comparisons for all possible pairs of 24 conditions. A one-dimensional coherence-based metric was used to analyze the quality judgments. This metric was an extension of a speech intelligibility metric presented in Kates and Arehart (2005) [J. Acoust. Soc. Am. 117, 2224-2237] and is based on dividing the speech signal into three amplitude regions, computing the coherence for each region, and then combining the three coherence values across frequency in a calculation based on the speech intelligibility index. The one-dimensional metric accurately predicted the quality judgments of normal-hearing listeners and listeners with mild-to-moderate hearing loss, although some systematic errors were present. A multidimensional analysis indicates that several dimensions are needed to describe the factors used by subjects to judge the effects of the three distortion types.

Longitudinal measurements of spontaneous otoacoustic emissions in infants

It has previously been shown [E. M. Burns, K. H. Arehart, and S. L. Campbell, J. Acoust. Soc. Am. 91, 1575-1581 (1992)] that both the overall prevalence of spontaneous otoacoustic emissions (SOAEs) and most of the various gender- and ear-related prevalence tendencies are not significantly different in 1-month-olds and adults. However, large differences were found between the neonates and adults in the distributions of the frequencies and levels of SOAEs. Both the average level and the median frequency were significantly higher in infants. To obtain longitudinal SOAE data, infants from this original group of 1-month-olds were tested at ages 3, 6, 12, and 24 months. In general, individual SOAEs decrease in level with age, and high-frequency SOAEs tend to show the largest decreases. No substantial shifts occur in the frequencies of individual SOAEs. The frequency and level distributions at 24 months of age are still not adult-like. SOAEs which show short-term instabilities in frequency and/or amplitude at 1 month of age typically continue to evidence such instabilities at later ages. These results suggest the cochlea is adult-like at birth, and imply that the observed SOAE changes reflect developmental changes in the external and middle ear.

Multichannel dynamic-range compression using digital frequency warping

A multichannel dynamic-range compressor system using digital frequency warping is described. A frequency-warped filter is realized by replacing the filter unit delays with all-pass filters. The appropriate design of the frequency warping gives a nonuniform frequency representation very close to the auditory Bark scale. The warped compressor is shown to have substantially reduced group delay in comparison with a conventional design having comparable frequency resolution. The warped compressor, however, has more delay at low than at high frequencies, which can lead to perceptible changes in the signal. The detection threshold for the compressor group delay was determined as a function of the number of all-pass filter sections in cascade needed for a detectible change in signal quality. The test signals included clicks, vowels, and speech, and results are presented for both normal-hearing and hearing-impaired subjects. Thresholds for clicks are lower than thresholds for vowels, and hearing-impaired subjects have higher thresholds than normal-hearing listeners. A frequency-warped compressor using a cascade of 31 all-pass filter sections offers a combination of low overall delay, good frequency resolution, and imperceptible frequency-dependent delay effects for most listening conditions.

Effects of harmonic content on complex-tone fundamental-frequency discrimination in hearing-impaired listeners.

Complex-tone (fundamental) frequency discrimination was measured in eight well-trained listeners with moderately severe sensorineural hearing impairments as a function of parametric variations in the rank, number, and sensation level of stimulus components. Results indicate substantial differences in the effects of harmonic content on complex-tone frequency discrimination among hearing-impaired listeners and between hearing-impaired and normal-hearing listeners. Despite large variability among the hearing-impaired subjects, several patterns of results emerged: Performance of two subjects grew worse as harmonic rank increased; performance of three subjects did not change substantially with changes in harmonic rank; and performance of three impaired subjects improved as harmonic rank increased. Performance of all but two subjects was significantly degraded for stimuli containing low-order harmonics. For stimuli containing only high-order harmonics, five subjects showed performance that was comparable to that of normal-hearing subjects, and three showed abnormally poor performance. Performance of impaired subjects generally improved as the number of stimulus components increased. The sensation level of stimulus components influenced the performance of several impaired subjects, but not in a uniform manner. To the extent that complex-tone fundamental-frequency discrimination can be assumed to be a pitch perception task, the present results suggest that, in contrast to normal-hearing subjects, hearing-impaired listeners rely primarily on periodicity cues in the perception of complex-tone pitch.

Speech Enhancement Based on Generalized Minimum Mean Square Error Estimators and Masking Properties of the Auditory System

In this paper, the family of conditional minimum mean square error (MMSE) spectral estimators is studied which take on the form (E(XEalphap/|Xp + Dp|))1alpha/, where Xp is the clean speech spectrum, and Dp is the noise spectrum, resulting in a generalized MMSE estimator (GMMSE). The degree of noise suppression versus musical tone artifacts of these estimators is studied. The tradeoffs in selection of (alpha), across noise spectral structure and signal-to-noise ratio (SNR) level, are also considered. Members of this family of estimators include the Ephraim-Malah (EM) amplitude estimator and, for high SNRs, the Wiener Filter. It is shown that the colorless residual noise observed in the EM estimator is a characteristic of this general family of estimators. An application of these estimators in an auditory enhancement scheme using the masking threshold of the human auditory system is formulated, resulting in the GMMSE-auditory masking threshold (AMT) enhancement method. Finally, a detailed evaluation of the proposed algorithms is performed over the phonetically balanced TIMIT database and the National Gallery of the Spoken Word (NGSW) audio archive using subjective and objective speech quality measures. Results show that the proposed GMMSE-AMT outperforms MMSE and log-MMSE enhancement methods using a detailed phoneme-based objective quality analysis

The Hearing-Aid Speech Perception Index (HASPI)

Abstract This paper presents a new index for predicting speech intelligibility for normal-hearing and hearing-impaired listeners. The Hearing-Aid Speech Perception Index (HASPI) is based on a model of the auditory periphery that incorporates changes due to hearing loss. The index compares the envelope and temporal fine structure outputs of the auditory model for a reference signal to the outputs of the model for the signal under test. The auditory model for the reference signal is set for normal hearing, while the model for the test signal incorporates the peripheral hearing loss. The new index is compared to indices based on measuring the coherence between the reference and test signals and based on measuring the envelope correlation between the two signals. HASPI is found to give accurate intelligibility predictions for a wide range of signal degradations including speech degraded by noise and nonlinear distortion, speech processed using frequency compression, noisy speech processed through a noise-suppression algorithm, and speech where the high frequencies are replaced by the output of a noise vocoder. The coherence and envelope metrics used for comparison give poor performance for at least one of these test conditions.

Evaluation of an auditory masked threshold noise suppression algorithm in normal-hearing and hearing-impaired listeners

While there have been numerous studies in the field of speech enhancement, the majority of these studies have focused on noise reduction for normal-hearing (NH) individuals. In addition, no speech enhancement algorithms reported in the signal processing community have reported an improvement in intelligibility, with the exception of a recent study by Tsoukalas et al. [IEEE Transactions of Speech and Audio Processing 5 (6) (1997) 497]. This study addresses the problem of speech enhancement for both NH and hearing-impaired (HI) subjects. A noise suppression algorithm based on auditory masked thresholds was implemented and evaluated for NH and HI subjects. Two different tests for intelligibility were used in the evaluation including the nonsense syllable test and the diagnostic rhyme test. Speech quality was evaluated using sentences from the hearing-in-noise test. Tests were performed using two types of noise (voice communications channel and automobile highway noise) at two different signal-to-noise ratios. Ten NH and 11 HI listeners were used to evaluate the enhancement algorithm. Results indicate that the enhancement algorithm yielded significantly better quality ratings and significantly better intelligibility scores in both NH and HI listeners in some but not all of the test conditions. The algorithm resulted in the greatest intelligibility improvements in the communications channels noise and for the nonsense syllable stimuli.