Consonant Place Perception Research Articles

On the relation of apparent naturalness to phonetic perceptual identification of voicing

Does the qualitative experience of speech sounds influence phonetic perception? Our studies of consonant place perception have revealed a dichotomous relation between phonetic sensitivity and naturalness. Although natural quality and phonetic sensitivity sometimes co-vary, in other conditions phonetic sensitivity is indifferent to huge variation in naturalness. New tests are reported here extending the research to the dimension of voicing, a contrast correlated with qualitatively distinct acoustic constituents in normal production. Two acoustic methods were used to create naturalness variants: (1) variation in the excitation of a synthetic voicing source and (2) variation in the bandwidth of the formant centers. A naturalness tournament was composed of items drawn from the test series, and the sensitivity of perceivers to the voicing contrast was estimated with the cumulative d′ across the series in identification tests. Together, the findings show how intelligibility and naturalness can be either orthogonal or contingent aspects of consonant perception for the dimensions of place and voicing. These measures offer a tool to understand the contribution of normative functions in the perception of speech. [Research supported by NIH (DC00308).]

Effect of a Single-Channel Wide Dynamic Range Compression Circuit on Perception of Stop Consonant Place of Articulation

Previous studies have shown that altering the amplitude of a consonant in a specific frequency region relative to an adjacent vowel's amplitude in the same frequency region will affect listeners' perception of the consonant place of articulation. Hearing aids with single-channel, fast-acting wide dynamic range compression (WDRC) alter the overall consonant-vowel (CV) intensity ratio by increasing consonant energy. Perhaps one reason WDRC has had limited success in improving speech recognition performance is that the natural amplitude balances between consonant and vowel are altered in crucial frequency regions, thus disturbing the aforementioned amplitude cue for determining place of articulation. The current study investigated the effect of a WDRC circuit on listeners' perception of place of articulation when the relative amplitude of consonant and vowel was manipulated. The stimuli were a continuum of synthetic CV syllables stripped of all place cues except relative consonant amplitudes. Acoustic analysis of the CVs before and after hearing aid processing showed a predictable increase in high-frequency energy, particularly for the burst of the consonant. Alveolar bursts had more high-frequency energy than labial bursts. Twenty-five listeners with normal hearing and 5 listeners with sensorineural hearing loss labeled the consonant sound of the CV syllables in unaided form and after the syllables were recorded through a hearing aid with single-channel WDRC. There were significantly more listeners who were unable to produce a category boundary when labeling the aided stimuli. Of those listeners who did yield a category boundary for both aided and unaided stimuli, there were significantly more alveolar responses for the aided condition. These results can be explained by the acoustic analyses of the aided stimuli.

Locus equations: A partial solution to the problem of consonant place perception

In their important work on locus equations, Sussman and his colleagues have helped to simplify the theoretical problem of how human listeners identify place of articulation contrasts among consonants, but much work remains before this problem is solved.

Improved information transfer of phoneme features with a multiple-electrode cochlear implant.

The spectral maxima sound processor (SMSP), recently developed at the University of Melbourne, has been shown to provide improved speech discrimination in quiet and in noise for subjects using the Mini System 22 cochlear implant when compared with commercially available speech processors, including the Mini Speech Processor (MSP) [C. M. McKay and H. J. McDermott, J. Acoust. Soc. Am. 89, 1959 (1991)]. The ability of five subjects to perceive phoneme features using the SMSP has been compared with their ability using the MSP. Information transfer analysis has been applied to the results of open-set CNC word tests and closed-set vowel and consonant recognition tests. Phoneme features studied were consonant manner and place of articulation and voicing, vowel duration, formant frequencies, and formant transitions. The results show an overall improvement in the perception of all phoneme features, with large improvements seen in perception of consonant place of articulation. This consonant feature has been perceived poorly relative to other consonant features with all previous speech processors used with this implant. Further development of the SMSP is being guided by the results of this analysis.

Auditory filter width and consonant recognition for hearing‐impaired listeners

Abnormal frequency resolution is often assumed to contribute to reduced speech recognition by hearing‐impaired listeners. However, a strong association between these two conditions secondary to hearing impairment has not been established, perhaps because both are degraded by threshold elevation. Thus, in this experiment, which examined auditory filter effects on consonant recognition, presentation levels of speech stimuli for each subject were selected on the basis of predictions from Articulation Index theory, in order to control for threshold sensitivity differences among the normal‐hearing and hearing‐impaired listeners. Estimates of filter width for 25 hearing‐impaired subjects (calculated from masked thresholds measured in broadband noise as function of the width of a spectral notch) revealed auditory filter widths that increased with the threshold for the probe. Consonant place perception for most hearing‐impaired subjects was nearly equivalent to results for normal‐hearing listeners when threshold ...

Toward the specification of spectral content for consonant place perception

Recent results from this laboratory have reemphasized the importance of the perception of broad regions of spectral energy for consonant recognition by normal‐hearing and hearing‐impaired listeners. In order to establish the bandwidth that is essential for recognition of specific cues, the effect of spectral content on consonant place perception was examined. High‐pass and low‐pass filtered consonant‐vowel (CV) syllables were presented to normal‐hearing individuals and to young listeners with high‐frequency sensorineural hearing loss secondary to noise exposure. Stimuli included nine synthetic syllables formed by pairing three voiced plosive consonants with three vowels, and three CVs spoken by a male talker. Stimuli were presented at five speech presentation levels ranging from 24–88 dB SPL. Performance as a function of cutoff frequency for each syllable suggests substantial differences in spectral regions contributing to the perception of the three consonants. In addition, the relative importance of certain spectral regions for consonant recognition is differentially affected by the accompanying vowel. Performance by the hearing‐impaired listeners demonstrates the interactive effects of speech level, audiometric configuration, and consonant spectral shape. [Work supported by grants from NINCDS.]

Initial stop burst place perception for normal‐ and impaired‐hearing listeners

This study examined whether hearing‐impaired listeners could identify place of articulation for initial stop consonants, using information occurring early in the stop release; for normal‐hearing listeners this release information has been implicated as a highly important cue for place perception of initial stops. Place perception for the release bursts of initial stop consonants was tested for impaired‐(N = 23) and for normal‐hearing (N = 6) listeners. The burst were obtained from ten utterances each of the syllables /pad/, /kad/, /tad/, /bad/, /gad/, /dad/. Three test conditions were used: (1) the entire syllable; (2) the initial release transients alone; and (3) the aspirations following the transients of the voiceless consonants. For the normal‐hearing group, consonant perception was 100% when tested for the entire syllable; consonant place perception for the transients only, was 87%, and for the voiceless aspirations 67%. Among the hearing‐impaired listeners, mean stop place perception was about 60% when tested for the entire syllable. Tests of the burst segments only, yielded chance performance for the impaired listeners.

Stop consonant place perception with single-formant stimuli: Evidence for the role of the front-cavity resonance

The third formant and the second formant were found on average to cue the place of articulation of intervocalic stop consonants equally well when the stop consonants occurred before the vowel/i/. This result and others provide some support for the notion that the fundamental resonance of the front cavity plays an important role in the perception of the phonetic dimension of place of articulation.