Manwa L. Ng

Consonant accuracy in Mandarin-speaking children with repaired cleft palate☆

OBJECTIVE To compare the consonant production of Chinese-speaking cleft palate children with perceived hypernasal resonance (PHR) after palatoplasty and those with perceived normal resonance (PNR), and to assess the possible influence of language on articulation. SETTING Two hospital cleft lip and palate centers in mainland China. PARTICIPANTS Thirty-one speakers were allocated into two groups based on perceptual judgment results of their resonance provided by three speech therapists: one group with PNR (n=20, average age=9.3 years), and another group with PHR (n=11, average age=8.3 years). All participants had no known hearing or cognitive deficits. INTERVENTION Articulation was evaluated using two Mandarin Chinese assessment tools, the Putonghua Segmental Phonology Test and the Deep Test for Cleft Palate Speakers in Putonghua. Speaker consonant accuracy was evaluated by two experienced speech therapists. RESULTS Compared to individuals with PNR, the PHR group exhibited more difficulties on production of unaspirated consonants, including/b/,/t/,/k/,/ts/,/tʂ/ and/tɕ/, than for aspirated consonants. CONCLUSION The distinctive feature of aspiration in Mandarin phonology brought a language specific pattern to consonant production among those speakers with PHR after primary palatal closure.

Differential effects of increased loudness on tongue kinematics in individuals with PD? Analyses of two cases

Previous studies have suggested significant variation in the perceptual speech deficit displayed by individuals with Parkinsons disease (PD). This heterogeneity is partly attributed to differential changes in the underlying articulatory kinematics across speakers. Loudness-based treatment has been frequently used in individuals with PD. An improved understanding of individual differences and the effects of loudness on tongue movement in individuals with PD may help optimize speech intervention. The present study aimed to (1) examine individual differences in tongue kinematics during habitual and loud sentence production in two dysarthric speakers with PD, and (2) examine the effect of loudness manipulations on tongue kinematics during sentence production within each dysarthric speaker with PD. Electromagnetic articulography was used to record tongue tip movement during habitual and loud sentence productions in two participants with PD. Their performance was first matched with the performance of an age and gender-matched healthy individual. Subsequently, a within-participant comparison between habitual and loud sentence production was carried out to examine the effect of increased loudness. When compared with an age and gender-matched healthy control, both PD individuals showed greater tongue kinematic parameters during habitual and loud sentence productions. Subsequently, a within-participant comparison showed that, from habitual to loud speech, one PD speaker increased distance and maximum velocity of tongue movement while the other reduced maximum acceleration of tongue movement. The observed individual differences in tongue kinematics may be predictive of the efficacy of loudness-based treatments for articulatory and consequent perceptual improvements for individual participants.

Surface Electromyographic Activity of Extrinsic Laryngeal Muscles in Cantonese Tone Production

Patients after total laryngectomy lose their ability to speak. Electrolarynx is a commonly used electronic device that helps these patients to verbally communicate. However, existing electrolarynx systems do not provide pitch control function, which is critical in speech communication especially for tonal languages. This study investigated the surface electromyographic (sEMG) activity of extrinsic laryngeal muscles in producing speech sounds of different pitches by normal speakers. In particular, the sEMG signals for producing different lexical tones of Cantonese were extracted and analyzed. The experimental results on Cantonese tone production confirmed that the sEMG signal from sternocleidomastoid muscle can be used to differentiate high-pitch tones from low-pitch tones. This reveals the potential of developing pitch-controlled EL systems for laryngectomees who speak Cantonese and other tonal languages.

The role of fundamental frequency and formants in voice gender identification

Abstract Purpose The present study examined how fundamental frequency (F0) and formant frequencies contribute to the identification of gender. Method Speech stimuli were synthesized from recorded voices of men and women using a formant scaling factor of 1.2 and F0 range of 100–250 Hz. Listeners who were native speakers of Cantonese were instructed to judge the perceived gender of the voice stimuli. Percent-correct gender identification of male and female stimuli at different F0–formant combinations was obtained. Results F0 was found to be the primary cue for gender perception and listeners showed a higher accuracy in identifying males than females voices. Conclusions The findings are consistent with results previously reported, although other acoustic cues such as voice quality may also affect gender perception.

A perceptual and acoustic study of voice gender in prepubertal children

Purpose: The present study attempted to investigate the sexual dimorphism in prepubertal childrens voice and their implicit knowledge on voice gender. Methods: Fundamental frequency (F0) and the first two formant frequencies (F1 and F2) of voice samples produced by 25 male and 26 female Cantonese-speaking prepubertal children produced under two speaking conditions: (1) natural condition, and (2) imitation condition when mimicking the opposite gender voice were examined. Results: Accuracy of gender identification by adult listeners for voices associated with the natural and imitation conditions was 81.7% and 59.5% respectively. In addition to the lack of difference in F0 between genders under the natural condition, F1 and F2 associated with boys were lower than those with girls, suggesting differences in vocal tract length and possible sex-specific articulatory behaviors between boys and girls, in an attempt to enhance sexual dimorphism for gender voice identification. Under the imitation condition, boys exhibited significantly higher F0 and F1 than girls. Conclusion: Prepubertal children possessed the implicit knowledge of sexually dimorphic acoustic correlates (F0 and F1) and were capable of altering the rate of vocal fold vibration and the effective vocal tract length at will to render vocal characteristics of the opposite sex.

Effects of L1 Tone on Perception of L2 Tone – A Study of Mandarin Tone Learning by Native Cantonese Children

In the present study, the Perceptual Assimilation Model (PAM) was tested on its applicability in child L2 lexical tone acquisition. The possible effect of L1 (Cantonese) lexical tones on L2 (Mandarin) lexical tone learning was explored. Accuracy rate and error patterns were examined with an AX discrimination task and a forced-choice identification task. Forty-nine native Cantonese-speaking students aged 8 years participated in the study. Results revealed that these children exhibited nearly perfect performance in the discrimination of Mandarin tones. However, significant tone differences were detected in the identification task. Tone 4 (T4) was identified with the lowest accuracy, and T1 with the highest. Error analysis revealed that Mandarin T2-T3 was the most confusing pair, followed by the T1-T4 pair. The inherent phonetic similarity between lexical tones in a language and the tone similarities across languages may also have contributed to perception difficulties, which could help to refine and supplement the PAM in the tonal/suprasegmental domain.

Effects of transcranial direct current stimulation over the Broca’s area on tongue twister production

Abstract Purpose: The present study aimed to explore the short-term effect of anodal transcranial direct current stimulation (tDCS) on tongue twister production. Method: Thirty healthy native Cantonese adult speakers were randomly assigned to the anodal tDCS group or the sham tDCS group. Anodal tDCS of 2 mA was applied over the Broca’s area of the brain. The stimulation lasted for 20 min for the anodal tDCS group and 30 s for the sham tDCS group. The participants were instructed to produce a list of tongue twisters before, immediately after and 4 h after tDCS. Result: Speech rate and response accuracy measured immediately after stimulation were significantly faster and higher, respectively, than before stimulation. Although there was no change in speech rate measured at 4 h after stimulation, response accuracy at that time point was significantly lower than that measured immediately after stimulation. However, there were no significant differences between the anodal tDCS and sham tDCS groups in either speech rate or response accuracy. Conclusion: The findings revealed that a single session of anodal tDCS over the Broca’s area did not significantly improve speech production during tongue twister production.

Differences in speech motor control between bilingual and monolingual speakers—An acoustic study

The present study attempted to acoustically examine the speech motor abilities associated with English produced by English monolingual speakers (MS), Cantonese-English bilingual speakers with superior (BS-SE) and inferior English (BS-IE). Articulation rate, formant frequencies (F1 and F2), and voice onset time (VOT) obtained from different speech tasks were compared across the three speaker groups to reveal the language influence on their speech motor control. Results indicated that: (1) the MS group exhibited the fastest articulation rate while the BS-IE group the slowest; (2) the three speaker groups had significantly different VOT values for the plosives /b-/, /g-/, /ph-/ and /th-/; and (3) bilingual speakers exhibited larger vowel spaces, demonstrating more posterior tongue position, based on F1 and F2 values of the three corner vowels /-i/, /-a/ and /-u/, than monolingual speakers. No systematic implications were obtained regarding the effect of bilingualism on speech motor control. Despite the incon...

The Influence of Language Experience on the Categorical Perception of Vowels: Evidence from Mandarin and Korean.

Previous research on categorical perception of speech sounds has demonstrated a strong influence of language experience on the categorical perception of consonants and lexical tones. In order to explore the influence of language experience on vowel perception, the present study investigated the perceptual performance for Mandarin and Korean listeners along a vowel continuum, which spanned three vowel categories /a/, /ɜ/, and /u/. The results showed that both language groups exhibited categorical features in vowel perception, with a sharper categorical boundary of /ɜ/-/u/ than that of /a/-/ɜ/. Moreover, the differences found between the two groups revealed that the Korean listeners’ perception tended to be more categorical along the /a/-/ɜ/-/u/ vowel continuum than that of the Mandarin listeners. Furthermore, the Mandarin listeners tended to label stimuli more often as /a/ and less often as /u/ than the Korean counterparts. These perceptual differences between the Mandarin and Korean groups might be attributed to the different acoustic distribution in the F1×F2 vowel space of the two different native languages.

Deep Neural Networks for Voice Quality Assessment Based on the GRBAS Scale.

Poster Presentation - Session: Learning, Education and Different Speech - no. Sun-P-7-3-3, paper ID 986