Kyoko Nagao

Phase transitions in a repetitive speech task as gestural recomposition

Tuller and Kelso (JSHR, 1991) show the relative timing of peak glottal opening and oral closure in repetitive /ip/ production differs between /ip/ and /pi/ productions. Also, as repetition rate increases, the timing in /ip/ merges with that in /pi/. The authors interpret this to indicate two stable regions in glottal‐to‐oral phasing, one for codas and one for onsets. The current paper examines a similar corpus of glottal transillumination traces of repeated syllables, which includes both labial and lingual stops. Preliminary examination of peak glottal timing with respect to acoustic landmarks indicating closure shows the same pattern of timing differences at slow rates and neutralization at fast rates. However, peaks in the transillumination traces often do not indicate glottal abduction for the consonant, but rather indicate a relatively open glottis for vocalic voicing, happening in the context of glottal stoppage at the onset of the vowel. In addition, coronal codas exhibit cycle‐to‐cycle variation from glottalized to abducted voiceless forms. Thus while rate does induce stepwise changes in glottal activity, phasing, in the sense of relative timing, does not account for many of the rate effects. Rather, rate appears to be inducing a recomposition of the gestures which comprise the speech.

The Perception of Syllable Affiliation of Singleton Stops in Repetitive Speech

Stetson (1951) noted that repeating singleton coda consonants at fast speech rates makes them be perceived as onset consonants affiliated with a following vowel. The current study documents the perception of rate-induced resyllabification, as well as what temporal properties give rise to the perception of syllable affiliation. Stimuli were extracted from a previous study of repeated stop + vowel and vowel + stop syllables (de Jong, 2001a, 2001b). Forced-choice identification tasks show that slow repetitions are clearly distinguished. As speakers increase rate, they reach a point after which listeners disagree as to the affiliation of the stop. This pattern is found for voiced and voiceless consonants using different stimulus extraction techniques. Acoustic models of the identifications indicate that the sudden shift in syllabification occurs with the loss of an acoustic hiatus between successive syllables. Acoustic models of the fast rate identifications indicate various other qualities, such as consonant voicing, affect the probability that the consonants will be perceived as onsets. These results indicate a model of syllabic affiliation where specific juncture-marking aspects of the signal dominate parsing, and in their absence other differences provide additional, weaker cues to syllabic affiliation.

Effects of speaking style on the regularity of mora timing in Japanese

The regularity of mora timing in Japanese has remained controversial over the years [Warner and Arai, Phonetica 58, 1–25 (2001)]. It is possible that the degree of regularity varies with speaking style. Four Japanese subjects spoke six‐mora proper names with varying syllable structures where the second name had either three simple syllables (e.g., Tomiko) or two syllables with a long vowel (e.g., Tooko) or a long consonant (e.g., Tokko). These were read either in formal sentences, read in conversational style sentences, or used in spontaneous description of pictures of characters having these names. Moras were measured as the intervals between vowel onsets using an automatic vowel‐onset detection algorithm. Although all styles suggested regular mora timing, the results show that the styles differ in the degree of temporal compensation for the moras constituted by long vowels and long consonants which are shorter than consonant‐vowel syllables. The compensation was clearest in the most formal style of speech.

Ear effect and gender difference of spontaneous otoacoustic emissions in children with auditory processing disorder

Spontaneous otoacoustic emissions (SOAEs) are found in most healthy ears, and can be used to measure the health of the cochlear structures and feedback mechanism. According to existing literature, right ears tend to exhibit greater numbers of SOAEs than left ears (Bilger et al., 1990) and females tend to show higher incidence of SOAEs than males (Moulin et al., 1993). The SOAE prevalence has not been extensively studied in children with Auditory Processing Disorder (APD), a disorder with unknown etiology that reduces ones ability to process auditory information. This study examined the prevalence and ear advantage of SOAEs between genders in children diagnosed with APD. SOAEs were investigated in 19 children (7 girls and 12 boys) with APD and 24 typically developing children (14 girls and 10 boys) aged 7-12. Right ear advantage was more prevalent in control (71%) than APD subjects (42%). However, over 30% more females exhibited a right ear advantage than males in each group. Although the results are not significant, our findings indicate that the lack of right ear advantage for SOAE is more prevalent in children with APD, particularly in males, suggesting that cochlear mechanisms or their control might be somehow affected in APD.

Relationship between articulation and mispronunciation detection in children with speech delay: Perception of unfamiliar speech vs. their own speech

We examined the relationship between speech production and mispronunciation detection ability in children with speech delay (SD). Thirty-three SD children aged between 6;0 and 10;0 participated in a mispronunciation detection test using three types of stimuli: words pronounced correctly by other unfamiliar children (OTHERS); words mispronounced by OTHERS; and the participant?s own speech (SELF) pronounced either correctly or incorrectly. The participant?s articulation was assessed by the standardized GFTA-2 scores. Results indicated that SD children made significantly more errors when judging SELF speech than when judging OTHERS speech. Multiple regression analyses revealed that accuracy of detecting OTHERS mispronounced words was a significant predictor of GFTA-2 scores in these SD children. Interestingly, in the regression model, accuracy for detecting SELF mispronunciations made a significant independent contribution in addition to accuracy at detecting OTHERS mispronunciations. Overall these two measures accounted for a significant proportion of the variance in GFTA-2 scores (R-squared = 0.45). These findings suggest that children with SD may have more coarse phonological representations of their own speech than the speech of other children.

Effects of listener experience with foreign accent on perception of accentedness and speaker age.

The current study examined the effects of foreign accent and listener experience on the perception of a speaker’s age and native language. Ten audio stimuli were prepared from the recording of five Arabic speakers and five English speakers (18–79 years old) from the Speech Accent Archive [Weinberger (2009)] for the perception experiment. Thirty native speakers of English participated in the perception experiment through Amazon’s Mechanical Turk website, estimated the speaker age, rated the speaker’s accentedness, and estimated the native language of the speaker. The listeners were divided into two groups based on their experience with foreign accented English (experienced and inexperienced groups). Higher correlation was found between perceived age (PA) and actual chronological age (CA) for the native (English) stimuli than for the Arabic‐accented stimuli in both listener groups. The correlation between PA and CA was higher in the experienced listeners than in the inexperienced listeners. Accentedness rat...

Lexical and phonotactic effects on the perception of rate induced resyllabification

Stetson (1951) noted that, when repeated, singleton coda consonants (VC) appear to modulate into onset consonants (CV) as the rate of repetition increases. de Jong et al. (2001) found that naEe listeners robustly perceive such resyllabifications with labial consonants, and in a later study, that such perceptions broadly corresponded to changes in glottal timing. In the current study, stimuli included labial, coronal, and velar stops, creating mixtures of real words (such as ‘‘eat’’), and nonwords (such as ‘‘ead’’). A comparison of the perception of real and nonreal words reveals no robust effect of lexical status. In addition, vowels in the corpus were either tense or lax, so that the CV combination is phonotactically illegal in half of the corpus. The perception of resyllabification also does occur with these lax vowels, though only for voiced coronal and labial stops. Other stops did not exhibit resyllabification. Analyses of glottal and acoustic recordings are currently underway. [Work supported by NIDCD and NSF.]

Cross‐language perception of rate induced resyllabification

This paper investigates the degree to which speakers of languages with different syllabic inventories are similar in their perceptions of syllabic affiliation. Stetson (1951) noted that repeated coda (VC) structures become perceived as onset (CV) structures as repetition rates increase. Stimuli were extracted from a repetitive production experiment in which English talkers produced voiced and voiceless labial stop onsets and codas at tempi controlled by a rate varying metronome. Native English, Japanese, and Korean listeners were asked to label the repeated syllable with one of four choices: ‘‘bee,’’ ‘‘pea,’’ ‘‘eeb,’’ ‘‘eep.’’ All three language groups showed perceptual shifts from VC to CV as the speech rate increased. However, Japanese listeners are biased toward identifying VC productions as CVs, and very rarely identified CV productions as VCs. These results indicate a very consistent cross‐language perception of syllabic affiliation, regardless of a language’s syllabic inventory, although the languag...

The perception of rate induced resyllabification in English

Perceptual resyllabification is a phenomenon in which English listeners start to perceive repeated coda (VC) structures as onset (CV) structures as repetition rates increase (Stetson 1951). This paper investigates whether native listeners exhibit perceptual resyllabification.

A role of fundamental frequencies in the perception of emphasized words

Many previous studies have shown the role of the fundamental frequency (F0) on focus structure in speech production. The current study examined the role of F0 on focus perception. Eleven subjects of Tokyo Japanese were asked to listen to 889 F0‐manipulated sentences and to select the word that they thought the speaker emphasized. Each stimulus included a noun phrase such as ‘‘W1‐no W2,’’ where both W1 and W2 are a noun with an accent on the first mora. The F0 contour of each stimulus was varied by changing values of F0 at the three points: the F0 minimum point (i.e., dip) in the noun phrase and the F0 maximum point (peak) in each word. The result showed that F0 functions as a phonetic cue for a listener to perceive the emphasis on the word. We found that the relative F0 differences between the two peaks largely determined the subject’s focus perception. Furthermore, logistic regression analysis suggests that the dip plays an important role. We will discuss the relationship between the dip and peaks for th...