Vowel Identification Scores Research Articles

Minimum spectral contrast needed for vowel identification by normal hearing and cochlear implant listeners.

The minimum spectral contrast needed for vowel identification by normal-hearing and cochlear implant listeners was determined in this study. In experiment 1, a spectral modification algorithm was used that manipulated the channel amplitudes extracted from a 6-channel continuous interleaved sampling (CIS) processor to have a 1-10 dB spectral contrast. The spectrally modified amplitudes of eight natural vowels were presented to six Med-EI/CIS-link users for identification. Results showed that subjects required a 4-6 dB contrast to identify vowels with relatively high accuracy. A 4-6 dB contrast was needed independent of the individual subject's dynamic range (range 9-28 dB). Some cochlear implant (CI) users obtained significantly higher scores with vowels enhanced to 6 dB contrast compared to the original, unenhanced vowels, suggesting that spectral contrast enhancement can improve the vowel identification scores for some CI users. To determine whether the minimum spectral contrast needed for vowel identification was dependent on spectral resolution (number of channels available), vowels were processed in experiment 2 through n (n =4, 6, 8, 12) channels, and synthesized as a linear combination of n sine waves with amplitudes manipulated to have a 1-20 dB spectral contrast. For vowels processed through 4 channels, normal-hearing listeners needed a 6 dB contrast, for 6 and 8 channels a 4 dB contrast was needed, consistent with our findings with CI listeners, and for 12 channels a 1 dB contrast was sufficient to achieve high accuracy (>80%). The above-mentioned findings with normal-hearing listeners suggest that when the spectral resolution is poor, a larger spectral contrast is needed for vowel identification. Conversely, when the spectral resolution is fine, a small spectral contrast (1 dB) is sufficient. The high identification score (82%) achieved with 1 dB contrast was significantly higher than any of the scores reported in the literature using synthetic vowels, and this can be attributed to the fact that we used natural vowels which contained duration and spectral cues (e.g., formant movements) present in fluent speech. The outcomes of experiments 1 and 2, taken together, suggest that CI listeners need a larger spectral contrast (4-6 dB) than normal-hearing listeners to achieve high recognition accuracy, not because of the limited dynamic range, but because of the limited spectral resolution.

Auditory Learning and Adaptation after Cochlear Implantation: A Preliminary Study of Discrimination and Labeling of Vowel Sounds by Cochlear Implant Users

This study examined two possible reasons underlying longitudinal increases in vowel identification by cochlear implant users: improved labeling of vowel sounds and improved electrode discrimination. The Multidimensional Phoneme Identification (MPI) model was used to obtain ceiling estimates of vowel identification for each subject, given his/her electrode discrimination skills. Vowel identification scores were initially lower than the ceiling estimates, but they gradually approached them over the first few months post-implant. Taken together, the present results suggest that improved labeling is the main mechanism explaining post-implant increases in vowel identification.

Childrens’ Phoneme Identification in Reverberation and Noise

This study assessed the effects of reverberation, noise, and their combination on listeners' identification of consonants and vowels in naturally produced nonsense syllables presented at different sensation levels (re: speech recognition threshold). A secondary purpose of this study was to assess listeners' identification of voicing, manner, and place of articulation for consonants at 50 dB SL in the reverberation, noise, and combined conditions. Listeners, aged 6-30 years, identified consonant-vowel-consonant-vowel (CVCV) stimuli presented at four different sensation levels (re: speech recognition threshold) of 30, 40, 50, and 60 dB SL in 4 listening conditions: (a) an optimal listening situation (no reverberation, no noise), (b) reverberation only (1.3 seconds), (c) noise only (+13 dB S/N against a multitalker babble), and (d) reverberation plus noise. Results showed that all listener groups achieved maximum consonant identification performance at 50 dB SL. Vowel identification scores were unaffected by SL. Statistical analyses revealed that children's ability to identify consonants varied according to listening condition. For example, children's consonant identification abilities reached adult-like levels of performance at about age 14 years in the reverberation-only and noise-only listening conditions. However, in the reverberation-plus-noise listening condition, children's consonant identification abilities do not mature until the late teenage years. The ability to identify vowels, on the other hand, develops much earlier. A feature analysis of the consonant data showed that for all 3 features (voicing, manner, and place), identification scores were highest in the control condition, similar for the reverberation-only and noise-only conditions, and lowest in the reverberation-plus-noise condition. Voicing was easier for listeners to identify than manner or place of articulation features in reverberation and noise. Taken together, these results suggest that the ability to identify speech in reverberation and noise reaches adult-like level of performance at different ages for different components of the speech signal.

Factors affecting vowel identification in hearing-impaired listeners

Vowel identification generally poses little difficulty for listeners with mild to moderate hearing impairment despite evidence for poorer frequency resolution compared to normal-hearing listeners. Normal-hearing listeners use formant movement and duration cues as well as spectral target cues to identify vowels. This study examines use of formant movement and duration cues by hearing-impaired listeners. Sets of very natural-sounding vowel stimuli with and without formant movement cues (dynamic and static) and with and without duration cues (appropriate and fixed) were made with a new speech resynthesis technique [H. Kawahara, Proc. ICASSP, 1–4 (1997)] using vowels produced by male and female talkers in /dVd/ context. Two groups of young normal-hearing listeners, one with simulated hearing loss typical of 70 year-old males and one control group, displayed no difference in overall vowel identification scores. Both groups obtained significant benefit from tokens containing formant movement cues. Duration was not as effective a cue for vowel identification as formant movement. Results for elderly hearing-impaired and elderly normal-hearing listeners will be presented, and the impact of talker variability upon vowel identification will be discussed. [Work supported by NIH-NIDCD 02229.]

"Minimal" High-Frequency Hearing Loss and School-Age Children

The purpose of this study was to assess the consonant and vowel identification abilities of 12 children with minimal high-frequency hearing loss, 12 children with normal hearing, and 12 young adults with normal hearing using nonsense syllables recorded in a classroom with reverberation time of 0.7 s in two conditions of: (1) quiet and (2) noise (+13 dB S/N against a multi-talker babble). The young adults achieved significantly higher mean consonant and vowel identification scores than both groups of children. The children with normal hearing had significantly higher mean consonant identification scores in quiet than the children with minimal high-frequency hearing loss, but the groups performances did not differ in noise. Further, the two groups of children did not differ in vowel identification performance. Listeners’ responses to consonant stimuli were converted to confusion matrices and submitted to a sequential information analysis (SINFA, Wang & Bilger, 1973). The SINFA determined that the amount of information transmitted, both overall and for individual features, differed as a function of listener group ad listening condition.

The influence of talker differences on vowel identification by normal-hearing and hearing-impaired listeners.

Vowel identification was tested in quiet, noise, and reverberation with 20 normal-hearing subjects and 20 hearing-impaired subjects. Stimuli were 15 English vowels spoken in a /b-t/context by six male talkers. Each talker produced five tokens of each vowel. In quiet, all stimuli were identified by two judges as the intended targets. The stimuli were degraded by reverberation or speech-spectrum noise. Vowel identification scores depended upon talker, listening condition, and subject type. The relationship between identification errors and spectral details of the vowels is discussed.

Speech of cochlear implant patients: a longitudinal study of vowel production.

Acoustic parameters were measured for vowels spoken in /hVd/ context by four postlingually deafened recipients of multichannel (Ineraid) cochlear implants. Three of the subjects became totally deaf in adulthood after varying periods of partial hearing loss; the fourth became totally deaf at age four. The subjects received different degrees of perceptual benefit from the prosthesis. Recordings were made before, and at intervals following speech processor activation. The measured parameters included F1, F2, F0, SPL, duration, and amplitude difference between the first two harmonic peaks in the log magnitude spectrum (H 1-H2). Numerous changes in parameter values were observed from pre- to post-implant, with differences among subjects. Many changes, but not all, were in the direction of normative data, and most changes were consistent with hypotheses about relations among the parameters. Some of the changes tended to enhance phonemic contrasts; others had the opposite effect. For three subjects, H 1-H2 changed in a direction consistent with measurements of their average air flow when reading; that relation was more complex for the fourth subject. The results are interpreted with respect to: characteristics of the individual subjects, including vowel identification scores; mechanical interactions among glottal and supraglottal articulations; and hypotheses about the role of auditory feedback in the control of speech production. Almost all the observed differences could be attributed to changes in the average settings of speaking rate, F0 and SPL, which presumably can be perceived without the need for spectral place information. Some observed F2 realignment may be attributable to the reception of spectral cues.

Identification of vowels produced by various talkers.

Differences in identification of vowels in degraded listening conditions, produced by various talkers, were investigated. Fifteen American monophthongs and diphthongs (vowels) were produced by six male talkers in a /b–t/ context. Each syllable was produced five times. The order of production of the 75 syllables was random. All vowels were highly identifiable in a quiet listening condition as assessed by two speech pathologists. Confusion matrices for the vowels degraded by either speech-spectrum noise (S/N=0 dB) or reverberation ( T=1.5 s) were collected with two groups composed of 20 normal-hearing and 20 hearing-impaired subjects. In each group, there were 100 responses for each vowel. Vowel identification scores depended on talkers, listening conditions, and type of subjects, indicating a relationship between production of the vowels and their vulnerability to degradations. Errors for selected vowels will be discussed. [Research supported by National Institute of Deafness and Other Communication Disorders.]

Loudspeaker‐to‐listener distance: Effect on the phoneme identification scores of normal‐hearing and hearing‐impaired listeners

Loudspeaker‐to‐listener distance was assessed for consonant and vowel identification scores of ten adults with normal‐hearing and ten adults with high‐frequency hearing loss (i.e., pure‐tone thresholds >25 dB HL at 2 kHz and higher) in a 415‐m3 (18×7.2×3.2 m) lecture hall, with a T = 1.5s (average T at 500, 1000, and 2000 Hz), and an estimated critical distance of 3.3 m. Each listener was seated at two distances within the critical distance (1 and 2 m) and three distances beyond the critical distance (4, 8, and 15 m) at a 0‐deg azimuth to the separation between two loudspeakers placed side‐by‐side. One loudspeaker presented a 25‐item CVCV nonsense syllable test at 77 dBA at 1 m. The 20 listeners were tested at all five distances in two speech‐to‐noise ratio (S/N) conditions: (1) varied—in which the speech level decreased with distance in relation to a 45‐dBA ambient noise level resulting in a + 32‐dB S/N at 1 m to a + 23‐dB S/N at 15 m, and (2) fixed—in which a multitalker babble was presented through the other loudspeaker and adjusted to maintain a + 15 dB S/N at all five distances. Scores were significantly better: (1) for normal‐hearing listeners than for hearing‐impaired listeners; and (2) in the varied S/N condition than in the fixed S/N condition. While vowel scores were not significantly affected by distance, consonant scores were significantly better within the critical distance than beyond the critical distance. Further, there was no significant decrease in consonant scores as loudspeaker‐to‐listener distance increased outside the critical distance.

Identification of vowels in quiet, noise, and reverberation: relationships with age and hearing loss.

Vowel identification in quiet, noise, and reverberation was tested with 40 subjects who varied in age and hearing level. Stimuli were 15 English vowels spoken in a (b-t) context in a carrier sentence, which were degraded by reverberation or noise (a babble of 12 voices). Vowel identification scores were correlated with various measures of hearing loss and with age. The mean of four hearing levels at 0.5, 1, 2, and 4 kHz, termed HTL4, produced the highest correlation coefficients in all three listening conditions. The correlation with age was smaller than with HTL4 and significant only for the degraded vowels. Further analyses were performed for subjects assigned to four groups on the basis of the amount of hearing loss. In noise, performance of all four groups was significantly different, whereas, in both quiet and reverberation, only the group with the greatest hearing loss performed differently from the other groups. The relationship among hearing loss, age, and number and type of errors is discussed in light of acoustic cues available for vowel identification.

On the role of spectral transition for speech perception.

This paper examines the relationship between dynamic spectral features and the identification of Japanese syllables modified by initial and/or final truncation. The experiments confirm several main points. "Perceptual critical points," where the percent correct identification of the truncated syllable as a function of the truncation position changes abruptly, are related to maximum spectral transition positions. A speech wave of approximately 10 ms in duration that includes the maximum spectral transition position bears the most important information for consonant and syllable perception. Consonant and vowel identification scores simultaneously change as a function of the truncation position in the short period, including the 10-ms period for final truncation. This suggests that crucial information for both vowel and consonant identification is contained across the same initial part of each syllable. The spectral transition is more crucial than unvoiced and buzz bar periods for consonant (syllable) perception, although the latter features are of some perceptual importance. Also, vowel nuclei are not necessary for either vowel or syllable perception.

Recognition of vowels from information in fricatives: perceptual evidence of fricative-vowel coarticulation.

Four studies investigated the perceptual effects of spectral variations in fricatives produced in different vowel contexts. The alveolar and palatal fricatives, [s, z, integral of, 3], were produced by two talkers in the context of the vowels [a, i, u], generating 12 fricative-vowel combinations. A computer-controlled editing procedure was used to excise fricative segments of 150-ms duration, as measured back from vowel onset. These excised segments were used as test stimuli in the four experiments. In the first experiment, fricative identification was highly accurate, especially for segments produced in the [a] context. The results of the subsequent three vowel identification experiments, revealed that the high vowels [i] and [u] were identified 60%--80% of the time in all fricative contexts, with the exception of [i] produced in the context of [integral of]. In contrast, identification scores for [a] were close to chance in all fricative contexts. Acoustic analyses of the stimuli revealed that the fricative segments with high vowel identification scores exhibited clear evidence of spectral changes associated with the vowels, while those segments with the highest fricative identification scores exhibited spectra most similar to fricatives produced in isolation. These results, in combination with more extensive acoustic analyses [S. D. Soli, J. Acoust. Soc. Am. 70, 976--984 (1981)] are discussed in terms of variations in the articulatory compatibility of tongue movements required to produce fricative-vowel sequences.

Order effect of acoustic segments of VC and CV syllables on stop and vowel identification.

The purpose of this study was to determine the importance of the vocalic transition relative to other acoustic cues in the perception of both CV and VC syllables. Subjects identified stops, /p, t, k, b, d, g/ and vowels, /i, u, 3/ from the aperiodic, aperiodic + vocalic transition, vocalic transition, and vocalic transition + vowel segments edited from VC and CV syllables. The major findings were as follows: (1) no significant difference was found in consonant identification scores between aperiodic, aperiodic + vocalic transition, and vocalic transition segments in CV syllables compared to those in VC syllables; (2) consonant identifications from vocalic transition + vowel segments in VC syllables were significantly greater than those from vocalic transition + vowel segments in CV syllables; (3) no significant difference was found in vowel identification scores between aperiodic + vocalic transition, vocalic transition + vowel, and vocalic transition segments in CV syllables compared to those in VC syllables; and (4) vowel identifications from aperiodic segments were significantly greater in CV syllables than in VC syllables. The results revealed a significant right-to-left perceptual advantage for both vowels and consonants. The findings for consonant identification emphasize the importance of the temporal order relationship between the vocalic transition and steady-state vowels in the perceptual process.

Vowel and speaker identification in natural and synthetic speech.

The study describes the results of a listening test in which subjects were required to identify both the vowel and the sex of the speaker. Ten monophthongal American English vowels were selected: /i I ε æ a u Λ 3/. These vowels were produced by a male speaker, a female speaker, and a child. Formant values were read from spectrograms made of these vowels and used for synthesizing the same 30 vowels on a Glace-Holmes synthesizer. A third set of vowels was generated synthetically using formant and fundamental frequency values reported by Peterson and Barney (1952). In this set, the fundamental frequency characteristic of male, female and children's voices was combined with male, female and children's formants. The set of 150 stimuli was presented to 60 listeners, who were asked to identify the speakers and vowels. Overall speaker identification scores were higher than overall vowel identification scores for both normal speakers and vowels synthesized from measurements; for the Peterson-Barney set, the scores were approximately the same. The highest overall correct score was 88.3%, obtained for the normal male speaker; the lowest correct score was 31.0%, obtained for children's vowels synthesized from measurements.

Effects of distance on the visual reception of speech.

Two talkers with normal hearing and speech presented with voice 240 common nouns (80 monosyllables, 80 trochees, 80 spondees) to six profoundly deaf children whose task was to lipread without acoustic cues at distances from 5-100 ft. Under bright, shadow-free illumination, lipreading performance diminished from 75% correct at 5 ft to 11% correct at 100 ft. Scores varied with distance similarly for both talkers. The stress patterns of the stimulus words influenced their intelligibility, with scores decreasing from spondees to trochees to monosyllables. In a supplementary study, one talker presented two tests of phoneme recognition to the same six deaf children whose task was to lipread from 5, 20, or 70 ft. Identification of consonants in VCV context depended on their place of articulation (front superior to back) and on the surrounding vowel (/a–a/ superior to /i–i/ or /u–u/). Vowel-identification scores were less dependent on distance than were consonant-identification scores. In general, tense (stressed) vowels were more easily identified in /b/-V-/b/ context than were lax (unstressed) vowels.