Checked tone

Abstract: This dissertation investigates how the Taiwanese Tone Sandhi Groups are perceived, and the acoustic/phonetics correlates of listeners' judgments. A series of perception experiments have been conducted to scrutinize the following topics - Taiwanese tone neutralization, Tone Sandhi Group (TSG) as a prosodic domain, perceived boundary strength in Taiwanese and Taiwanese sentence disambiguation. Taiwanese tone neutralization was examined with a checked tone recognition experiment and a corpus study. In the tone recognition gating experiment, the difference between the accuracy rate on citation and sandhi tones suggested that the two tones maybe be incompletely neutralized but as members of the same tonal category. The corpus study found that duration, f0 range and voice quality in domain-final position preserve distinctions between all the tone pairs. More specifically, the citation tones are longer in duration, wider in F0 range and creakier in voice quality. The prosodic domain identification experiment revealed that participants were able to identify TSG as well as the other two prosodic levels, Word and Intonation Phrase even when the speech signal is low-pass filtered (i.e. only the prosodic cues - duration, f0, and voice quality - were retained). The perceived boundary strength rating experiment further showed that participants assigned different ratings for the three prosodic levels in spontaneous speech stimuli, and the acoustic analyses revealed that duration, f0 and voice quality cues in the last syllable in each stimulus were useful to the listeners. However, multiple linear regression indicated that these acoustic cues were far from sufficient to account for listeners' perception. In other words, to discriminate one prosodic domain from the others, listeners apparently need acoustic/prosodic cues not only from the syllables at the prosodic boundary, but also the syllables before the boundary within the same prosodic domain. In the sentence disambiguation experiment, participants were able to interpret ambiguous sentences accurately. The results suggest that prosodic cues such as final lengthening and pitch reset, found at the disambiguation points, provide a strong basis for sentence identification.

Abstract: This study investigates the relative importance of phonation and pitch cues in (White) Hmong tone identification. Hmong has seven productive tones, two of which involve non-modal phonation. The breathy tone is usually produced with a mid- or high-falling pitch contour similar to the high-falling modal tone. Similarly, aside from some pitch differences between the low modal tone and the low-falling creaky or checked tone, production studies have shown that the phonation differences between the two tones are large. Fifteen native listeners participated in two perception tasks, in which they were asked to indentify the word they heard. In the first task, participants heard natural stimuli with manipulated F0 and duration (phonation unchanged). Results indicate that the phonation of the stimulus is important in identifying the breathy tone, but not the creaky one. Duration and F0 were more closely tied to creaky tonal identification than phonation. In the second task, source spectrum components were manipulated to create stimuli ranging from modal to breathy sounding, with the F0 held constant. The results of this task indicate that changes in H1-H2 and H2-H4 are independently important for distinguishing breathy from modal phonation when F0 is held constant. [Work supported by NSF and NIH]

Abstract: It is well known that the duration reflects the difference between checked tone's syllables and other syllables in Cantonese. In this paper, aiming at improving the naturalness of the Cantonese TTS system, we use two concepts, the voice blank and pitch contour blank, to examine the particularity of the checked tone's duration with a serial designed experiments in Cantonese speech. The impact of the four factors, including the tone types of the Cantonese checked tone's syllable (CHKTS), the types of the occlusive stop of the CHKTS, the onset types of the syllables following the CHKTS, the location in the sentence of the CHKTS, have been investigated for the duration of check tones. The results are helpful to construct the context information for the prosody prediction in Cantonese TTS system. Moreover, with two perception experiments, the paper also proves that the silence segment which indicated as voice blank and extension of the onset duration following the CHKTSs in the Cantonese utterance is a kind of prosodic compensation of the CHKTSs.

Abstract: This paper selects normal checked characters and makes a comprehensive analysis of checked tone merging in Lingchuan dialect from rhyme, initial consonants and tone. In terms of rhyme, geng rhyme’s merging speed is fastest and shen rhyme is slowest. As for ancient consonants, nasal has the fastest speed, and voiced initial has the slowest speed. Judging from the attribution of the tones, the voiced and nasal initial checked characters are majorly classified into the departing tone, while the voiceless initial checked characters are attributed to four tone categories. The merging in Lingchuan dialect mostly follows the rules of mandarin, but from the perspective of the quantity, there are a few differences between Lingchuan dialect and Mandarin.