Recognition of time-distorted sentences by normal-hearing and cochlear-implant listeners

This study evaluated the effects of time compression and expansion on sentence recognition by normal-hearing (NH) listeners and cochlear-implant (CI) recipients of the Nucleus-22 device. Sentence recognition was measured in five CI users using custom 4-channel continuous interleaved sampler (CIS) processors and five NH listeners using either 4-channel or 32-channel noise-band processors. For NH listeners, recognition was largely unaffected by time expansion, regardless of spectral resolution. However, recognition of time-compressed speech varied significantly with spectral resolution. When fine spectral resolution (32 channels) was available, speech recognition was unaffected even when the duration of sentences was shortened to 40% of their original length (equivalent to a mean duration of 40 ms/phoneme). However, a mean duration of 60 ms/phoneme was required to achieve the same level of recognition when only coarse spectral resolution (4 channels) was available. Recognition patterns were highly variable across CI listeners. The best CI listener performed as well as NH subjects listening to corresponding spectral conditions; however, three out of five CI listeners performed significantly poorer in recognizing time-compressed speech. Further investigation revealed that these three poorer-performing CI users also had more difficulty with simple temporal gap-detection tasks. The results indicate that limited spectral resolution reduces the ability to recognize time-compressed speech. Some CI listeners have more difficulty with time-compressed speech, as produced by rapid speakers, because of reduced spectral resolution and deficits in auditory temporal processing.     

Auditory-visual speech perception in normal-hearing and cochlear-implant listeners

The present study evaluated auditory-visual speech perception in cochlear-implant users as well as normal-hearing and simulated-implant controls to delineate relative contributions of sensory experience and cues. Auditory-only, visual-only, or auditory-visual speech perception was examined in the context of categorical perception, in which an animated face mouthing ba, da, or ga was paired with synthesized phonemes from an 11-token auditory continuum. A three-alternative, forced-choice method was used to yield percent identification scores. Normal-hearing listeners showed sharp phoneme boundaries and strong reliance on the auditory cue, whereas actual and simulated implant listeners showed much weaker categorical perception but stronger dependence on the visual cue. The implant users were able to integrate both congruent and incongruent acoustic and optical cues to derive relatively weak but significant auditory-visual integration. This auditory-visual integration was correlated with the duration of the implant experience but not the duration of deafness. Compared with the actual implant performance, acoustic simulations of the cochlear implant could predict the auditory-only performance but not the auditory-visual integration. These results suggest that both altered sensory experience and improvised acoustic cues contribute to the auditory-visual speech perception in cochlear-implant users.     

Spectral shape discrimination by hearing-impaired and normal-hearing listeners

The ability to discriminate between sounds with different spectral shapes was evaluated for normal-hearing and hearing-impaired listeners. Listeners discriminated between a standard stimulus and a signal stimulus in which half of the standard components were decreased in level and half were increased in level. In one condition, the standard stimulus was the sum of six equal-amplitude tones (equal-SPL), and in another the standard stimulus was the sum of six tones at equal sensation levels re: audiometric thresholds for individual subjects (equal-SL). Spectral weights were estimated in conditions where the amplitudes of the individual tones were perturbed slightly on every presentation. Sensitivity was similar in all conditions for normal-hearing and hearing-impaired listeners. The presence of perturbation and equal-SL components increased thresholds for both groups, but only small differences in weighting strategy were measured between the groups depending on whether the equal-SPL or equal-SL condition was tested. The average data suggest that normal-hearing listeners may rely more on the central components of the spectrum whereas hearing-impaired listeners may have been more likely to use the edges. However, individual weighting functions were quite variable, especially for the HI listeners, perhaps reflecting difficulty in processing changes in spectral shape due to hearing loss. Differences in weighting strategy without changes in sensitivity suggest that factors other than spectral weights, such as internal noise or difficulty encoding a reference stimulus, also may dominate performance.     

Sample discrimination of frequency by hearing-impaired and normal-hearing listeners

In a multiple observation, sample discrimination experiment normal-hearing (NH) and hearing-impaired (HI) listeners heard two multitone complexes each consisting of six simultaneous tones with nominal frequencies spaced evenly on an ERB(N) logarithmic scale between 257 and 6930 Hz. On every trial, the frequency of each tone was sampled from a normal distribution centered near its nominal frequency. In one interval of a 2IFC task, all tones were sampled from distributions lower in mean frequency and in the other interval from distributions higher in mean frequency. Listeners had to identify the latter interval. Decision weights were obtained from multiple regression analysis of the between- interval frequency differences for each tone and listeners' responses. Frequency difference limens (an index of sensorineural resolution) and decision weights for each tone were used to predict the sensitivity of different decision-theoretic models. Results indicate that low-frequency tones were given much greater perceptual weight than high-frequency tones by both groups of listeners. This tendency increased as hearing loss increased and as sensorineural resolution decreased, resulting in significantly less efficient weighting strategies for the HI listeners. Overall, results indicate that HI listeners integrated frequency information less optimally than NH listeners, even after accounting for differences in sensorineural resolution.     

Sentence recognition in noise promoting or suppressing masking release by normal-hearing and cochlear-implant listeners

Normal-hearing (NH) listeners maintain robust speech understanding in modulated noise by "glimpsing" portions of speech from a partially masked waveform--a phenomenon known as masking release (MR). Cochlear implant (CI) users, however, generally lack such resiliency. In previous studies, temporal masking of speech by noise occurred randomly, obscuring to what degree MR is attributable to the temporal overlap of speech and masker. In the present study, masker conditions were constructed to either promote (+MR) or suppress (-MR) masking release by controlling the degree of temporal overlap. Sentence recognition was measured in 14 CI subjects and 22 young-adult NH subjects. Normal-hearing subjects showed large amounts of masking release in the +MR condition and a marked difference between +MR and -MR conditions. In contrast, CI subjects demonstrated less effect of MR overall, and some displayed modulation interference as reflected by poorer performance in modulated maskers. These results suggest that the poor performance of typical CI users in noise might be accounted for by factors that extend beyond peripheral masking, such as reduced segmental boundaries between syllables or words. Encouragingly, the best CI users tested here could take advantage of masker fluctuations to better segregate the speech from the background.     

Spectral-peak selection in spectral-shape discrimination by normal-hearing and hearing-impaired listeners

Spectral-shape discrimination thresholds were measured in the presence and absence of noise to determine whether normal-hearing and hearing-impaired listeners rely primarily on spectral peaks in the excitation pattern when discriminating between stimuli with different spectral shapes. Standard stimuli were the sum of 2, 4, 6, 8, 10, 20, or 30 equal-amplitude tones with frequencies fixed between 200 and 4000 Hz. Signal stimuli were generated by increasing and decreasing the levels of every other standard component. The function relating the spectral-shape discrimination threshold to the number of components (N) showed an initial decrease in threshold with increasing N and then an increase in threshold when the number of components reached 10 and 6, for normal-hearing and hearing-impaired listeners, respectively. The presence of a 50-dB SPL/Hz noise led to a 1.7 dB increase in threshold for normal-hearing listeners and a 3.5 dB increase for hearing-impaired listeners. Multichannel modeling and the relatively small influence of noise suggest that both normal-hearing and hearing-impaired listeners rely on the peaks in the excitation pattern for spectral-shape discrimination. The greater influence of noise in the data from hearing-impaired listeners is attributed to a poorer representation of spectral peaks.     

Level discrimination of single tones in a multitone complex by normal-hearing and hearing-impaired listeners

A conditional-on-a-single-stimulus (COSS) analysis procedure [B. G. Berg, J. Acoust. Soc. Am. 86, 1743-1746 (1989)] was used to estimate how well normal-hearing and hearing-impaired listeners selectively attend to individual spectral components of a broadband signal in a level discrimination task. On each trial, two multitone complexes consisting of six octave frequencies from 250 to 8000 Hz were presented to listeners. The levels of the individual tones were chosen independently and at random on each presentation. The target tone was selected, within a block of trials, as the 250-, 1000-, or 4000-Hz component. On each trial, listeners were asked to indicate which of the two complex sounds contained the higher level target. As a group, normal-hearing listeners exhibited greater selectivity than hearing-impaired listeners to the 250-Hz target, while hearing-impaired listeners showed greater selectivity than normal-hearing listeners to the 4000-Hz target, which is in the region of their hearing loss. Both groups of listeners displayed large variability in their ability to selectively weight the 1000-Hz target. Trial-by-trial analysis showed a decrease in weighting efficiency with increasing frequency for normal-hearing listeners, but a relatively constant weighting efficiency across frequency for hearing-impaired listeners. Interestingly, hearing-impaired listeners selectively weighted the 4000-Hz target, which was in the region of their hearing loss, more efficiently than did the normal-hearing listeners.     

Discrimination of interaural temporal disparities by normal-hearing listeners and listeners with high-frequency sensorineural hearing loss

Thresholds of ongoing interaural time difference (ITD) were obtained from normal-hearing and hearing-impaired listeners who had high-frequency, sensorineural hearing loss. Several stimuli (a 500-Hz sinusoid, a narrow-band noise centered at 500 Hz, a sinusoidally amplitude-modulated 4000-Hz tone, and a narrow-band noise centered at 4000 Hz) and two criteria [equal sound-pressure level (Eq SPL) and equal sensation level (Eq SL)] for determining the level of stimuli presented to each listener were employed. The ITD thresholds and slopes of the psychometric functions were elevated for hearing-impaired listeners for the two high-frequency stimuli in comparison to: the listener's own low-frequency thresholds; and data obtained from normal-hearing listeners for stimuli presented with Eq SPL interaurally. The two groups of listeners required similar ITDs to reach threshold when stimuli were presented at Eq SLs to each ear. For low-frequency stimuli, the ITD thresholds of the hearing-impaired listener were generally slightly greater than those obtained from the normal-hearing listeners. Whether these stimuli were presented at either Eq SPL or Eq SL did not differentially affect the ITD thresholds across groups.     

Spectral-shape discrimination. I. Results from normal-hearing listeners for stationary broadband noises

This research is concerned with the ability of normal-hearing listeners to discriminate broadband signals on the basis of spectral shape. The signals were six broadband noises whose spectral shapes were modeled after the spectra of unvoiced fricative and plosive consonants. The difficulty of the discriminations was controlled by the addition of noise filtered to match the long-term speech spectrum. Two-interval discrimination measurements were made in which loudness cues were eliminated by randomizing (roving) the overall stimulus level between presentation intervals. Experimental results, examined as a function of intensity rove width, stimulus duration, and stimulus pair, were related to the predictions of a simple filter-bank model whose fitting parameter provides an estimate of internal noise. Most results, with the notable exception of duration effects, were predicted by the model. Estimates of internal noise in each frequency channel averaged roughly 7 dB for long-duration stimuli and 13 dB for short-duration stimuli. Results and predictions are compared to results of other studies concerned with the discrimination of spectral shape.     

Spectro-temporal analysis in normal-hearing and cochlear-impaired listeners

Detection thresholds for a 1.0-kHz pure tone were determined in unmodulated noise and in noise modulated by a 15-Hz square wave. Comodulation masking release (CMR) was calculated as the difference in threshold between the modulated and unmodulated conditions. The noise bandwidth varied between 100 and 1000 Hz. Frequency selectivity was also examined using an abbreviated notched-noise masking method. The subjects in the main experiment consisted of 12 normal-hearing and 12 hearing-impaired subjects with hearing loss of cochlear origin. The most discriminating conditions were repeated on 16 additional hearing-impaired subjects. The CMR of the hearing-impaired group was reduced for the 1000-Hz noise bandwidth. The reduced CMR at this bandwidth correlated significantly with reduced frequency selectivity, consistent with the hypothesis that the across-frequency difference cue used in CMR is diminished by poor frequency selectivity. The results indicated that good frequency selectivity is a prerequisite, but not a guarantee, of large CMR.     

Frequency dependence of binaural performance in listeners with impaired binaural hearing.

Binaural performance was measured as a function of stimulus frequency for four impaired listeners, each with bilaterally symmetric audiograms. The subjects had various degrees and configurations of audiometric losses: two had high-frequency, sensorineural losses; one had a flat sensorineural loss; and one had multiple sclerosis with normal audiometric thresholds. Just noticeable differences (jnd's) in interaural time, interaural intensity, and interaural correlation as well as detection thresholds for NoSo and NoS pi conditions were obtained for narrow-band noise stimuli at octave frequencies from 250-4000 Hz. Performance of the impaired listeners was generally poorer than that of normal-hearing listeners, although it was comparable to normal in a few instances. The patterns of binaural performance showed no apparent relation to the audiometric patterns; even the two subjects with similar degree and configuration of hearing loss have very different binaural performance, both in the level and frequency dependence of their performance. The frequency dependence of performance on individual tests is irregular enough that one cannot confidently interpolate between octaves. In addition, it appears that no subset of the measurements is adequate to characterize the performance in the rest of the measurements with the exception that, within limits, interaural correlation discrimination and NoS pi detection performance are related.     

Perception of clear fricatives by normal-hearing and simulated hearing-impaired listeners

Speakers may adapt the phonetic details of their productions when they anticipate perceptual difficulty or comprehension failure on the part of a listener. Previous research suggests that a speaking style known as clear speech is more intelligible overall than casual, conversational speech for a variety of listener populations. However, it is unknown whether clear speech improves the intelligibility of fricative consonants specifically, or how its effects on fricative perception might differ depending on listener population. The primary goal of this study was to determine whether clear speech enhances fricative intelligibility for normal-hearing listeners and listeners with simulated impairment. Two experiments measured babble signal-to-noise ratio thresholds for fricative minimal pair distinctions for 14 normal-hearing listeners and 14 listeners with simulated sloping, recruiting impairment. Results indicated that clear speech helped both groups overall. However, for impaired listeners, reliable clear speech intelligibility advantages were not found for non-sibilant pairs. Correlation analyses comparing acoustic and perceptual data indicated that a shift of energy concentration toward higher frequency regions and greater source strength contributed to the clear speech effect for normal-hearing listeners. Correlations between acoustic and perceptual data were less consistent for listeners with simulated impairment, and suggested that lower-frequency information may play a role.     

Minimum spectral contrast for vowel identification by normal-hearing and hearing-impaired listeners

To determine the minimum difference in amplitude between spectral peaks and troughs sufficient for vowel identification by normal-hearing and hearing-impaired listeners, four vowel-like complex sounds were created by summing the first 30 harmonics of a 100-Hz tone. The amplitudes of all harmonics were equal, except for two consecutive harmonics located at each of three "formant" locations. The amplitudes of these harmonics were equal and ranged from 1-8 dB more than the remaining components. Normal-hearing listeners achieved greater than 75% accuracy when peak-to-trough differences were 1-2 dB. Normal-hearing listeners who were tested in a noise background sufficient to raise their thresholds to the level of a flat, moderate hearing loss needed a 4-dB difference for identification. Listeners with a moderate, flat hearing loss required a 6- to 7-dB difference for identification. The results suggest, for normal-hearing listeners, that the peak-to-trough amplitude difference required for identification of this set of vowels is very near the threshold for detection of a change in the amplitude spectrum of a complex signal. Hearing-impaired listeners may have difficulty using closely spaced formants for vowel identification due to abnormal smoothing of the internal representation of the spectrum by broadened auditory filters.     

Stop-consonant recognition for normal-hearing listeners and listeners with high-frequency hearing loss. II: Articulation index predictions

Articulation index (AI) theory was used to evaluate stop-consonant recognition of normal-hearing listeners and listeners with high-frequency hearing loss. From results reported in a companion article [Dubno et al., J. Acoust. Soc. Am. 85, 347-354 (1989)], a transfer function relating the AI to stop-consonant recognition was established, and a frequency importance function was determined for the nine stop-consonant-vowel syllables used as test stimuli. The calculations included the rms and peak levels of the speech that had been measured in 1/3 octave bands; the internal noise was estimated from the thresholds for each subject. The AI model was then used to predict performance for the hearing-impaired listeners. A majority of the AI predictions for the hearing-impaired subjects fell within +/- 2 standard deviations of the normal-hearing listeners' results. However, as observed in previous data, the AI tended to overestimate performance of the hearing-impaired listeners. The accuracy of the predictions decreased with the magnitude of high-frequency hearing loss. Thus, with the exception of performance for listeners with severe high-frequency hearing loss, the results suggest that poorer speech recognition among hearing-impaired listeners results from reduced audibility within critical spectral regions of the speech stimuli.     

The resolution of complex spectral patterns by cochlear implant and normal-hearing listeners

The differences in spectral shape resolution abilities among cochlear implant (CI) listeners, and between CI and normal-hearing (NH) listeners, when listening with the same number of channels (12), was investigated. In addition, the effect of the number of channels on spectral shape resolution was examined. The stimuli were rippled noise signals with various ripple frequency-spacings. An adaptive 41FC procedure was used to determine the threshold for resolvable ripple spacing, which was the spacing at which an interchange in peak and valley positions could be discriminated. The results showed poorer spectral shape resolution in CI compared to NH listeners (average thresholds of approximately 3000 and 400 Hz, respectively), and wide variability among CI listeners (range of approximately 800 to 8000 Hz). There was a significant relationship between spectral shape resolution and vowel recognition. The spectral shape resolution thresholds of NH listeners increased as the number of channels increased from 1 to 16, while the CI listeners showed a performance plateau at 4-6 channels, which is consistent with previous results using speech recognition measures. These results indicate that this test may provide a measure of CI performance which is time efficient and non-linguistic, and therefore, if verified, may provide a useful contribution to the prediction of speech perception in adults and children who use CIs.     

Speech recognition with varying numbers and types of competing talkers by normal-hearing, cochlear-implant, and implant simulation subjects

The goals of this study were to measure sensitivity to the direct-to-reverberant energy ratio (D/R) across a wide range of D/R values and to gain insight into which cues are used in the discrimination process. The main finding is that changes in D/R are discriminated primarily based on spectral cues. Temporal cues may be used but only when spectral cues are diminished or not available, while sensitivity to interaural cross-correlation is too low to be useful in any of the conditions tested. These findings are based on an acoustic analysis of these variables and the results of two psychophysical experiments. The first experiment employs wideband noise with two values for onset and offset times to determine the D/R just-noticeable difference at -10, 0, 10, and 20 dB D/R. This yielded substantially higher sensitivity to D/R at 0 and 10 dB D/R (2-3 dB) than has been reported previously, while sensitivity is much lower at -10 and 20 dB D/R. The second experiment consists of three parts where specific cues to D/R are reduced or removed, which enabled the specified rank ordering of the cues. The acoustic analysis and psychophysical experiments also provide an explanation for the "auditory horizon effect."     

The effect of competing melodies on melody recognition by hearing-impaired and normal-hearing listeners

For a group of 30 hearing-impaired subjects and a matched group of 15 normal-hearing subjects (age range 13-17) the following data were collected: the tone audiogram, the auditory bandwidth at 1000 Hz, and the recognition threshold of a short melody presented simultaneously with two other melodies, lower and higher in frequency, respectively. The threshold was defined as the frequency distance required to recognize the test melody. It was found that, whereas the mean recognition threshold for the normal-hearing subjects was five semitones, it was, on the average, 27 semitones for the hearing-impaired subjects. Although the interindividual spread for the latter group was large, it did not correlate with the subjects' auditory bandwidth, nor with their musical experience or education.     

Evaluation of two voice-separation algorithms using normal-hearing and hearing-impaired listeners

Two signal-processing algorithms, designed to separate the voiced speech of two talkers speaking simultaneously at similar intensities in a single channel, were compared and evaluated. Both algorithms exploit the harmonic structure of voiced speech and require a difference in fundamental frequency (F0) between the voices to operate successfully. One attenuates the interfering voice by filtering the cepstrum of the combined signal. The other uses the method of harmonic selection [T. W. Parsons, J. Acoust. Soc. Am. 60, 911-918 (1976)] to resynthesize the target voice from fragmentary spectral information. Two perceptual evaluations were carried out. One involved the separation of pairs of vowels synthesized on static F0's; the other involved the recovery of consonant-vowel (CV) words masked by a synthesized vowel. Normal-hearing listeners and four listeners with moderate-to-severe, bilateral, symmetrical, sensorineural hearing impairments were tested. All listeners showed increased accuracy of identification when the target voice was enhanced by processing. The vowel-identification data show that intelligibility enhancement is possible over a range of F0 separations between the target and interfering voice. The recovery of CV words demonstrates that the processing is valid not only for spectrally static vowels but also for less intense time-varying voiced consonants. The results for the impaired listeners suggest that the algorithms may be applicable as components of a noise-reduction system in future digital signal-processing hearing aids. The vowel-separation test, and subjective listening, suggest that harmonic selection, which is the more computationally expensive method, produces the more effective voice separation.     

VCVs vs CVCs for stop/fricative distinctions by hearing-impaired and normal-hearing listeners

Moderately to profoundly hearing-impaired (n = 30) and normal-hearing (n = 6) listeners identified [p, k, t, f, theta, s] in [symbol; see text], and [symbol; see text]s tokens extracted from spoken sentences. The [symbol; see text]s were also identified in the sentences. The hearing-impaired group distinguished stop/fricative manner more poorly for [symbol; see text] in sentences than when extracted. Further, the group's performance for extracted [symbol; see text] was poorer than for extracted [symbol; see text] and [symbol; see text]. For the normal-hearing group, consonant identification was similar among the syllable and sentence contexts.     

Spectral integration of speech bands in normal-hearing and hearing-impaired listeners

This investigation examined whether listeners with mild-moderate sensorineural hearing impairment have a deficit in the ability to integrate synchronous spectral information in the perception of speech. In stage 1, the bandwidth of filtered speech centered either on 500 or 2500 Hz was varied adaptively to determine the width required for approximately 15%-25% correct recognition. In stage 2, these criterion bandwidths were presented simultaneously and percent correct performance was determined in fixed block trials. Experiment 1 tested normal-hearing listeners in quiet and in masking noise. The main findings were (1) there was no correlation between the criterion bandwidths at 500 and 2500 Hz; (2) listeners achieved a high percent correct in stage 2 (approximately 80%); and (3) performance in quiet and noise was similar. Experiment 2 tested listeners with mild-moderate sensorineural hearing impairment. The main findings were (1) the impaired listeners showed high variability in stage 1, with some listeners requiring narrower and others requiring wider bandwidths than normal, and (2) hearing-impaired listeners achieved percent correct performance in stage 2 that was comparable to normal. The results indicate that listeners with mild-moderate sensorineural hearing loss do not have an essential deficit in the ability to integrate across-frequency speech information.