A study on quantitative computation for prosodie strength of Mandarin speech

Prosodic strength refers to the relative prominence of each syllable, and it contributes to analysis of prosodic structure. A speaker would spend more effort to speak clearly in prosodically strong positions, and spend less effort to produce reduced phonetic forms in prosodically weak positions. Based on this concept, we proposed a method for calculating the strength of each syllable in continuous Mandarin speech. We computed the extra effort that speaker spend for the syllable, and the deviation degree of the F0 contour from its canonical tone. The calculated extra effort, deviation degree and duration were combined in a linear model for prosodic strength computation. The experiments were conducted on the Annotated Speech Corpus of Chinese Discourse. Results showed that: a) Strength values are related to several grammatical and prosodic features such as part of speech, word stress level, and break index; b) The average strength values of syllables in continuous speech are significantly smaller than that of isolated mono-syllable. The results showed the effectiveness of the model

[1]  A. Prince,et al.  On stress and linguistic rhythm , 1977 .

[2]  Mark Liberman,et al.  Prosodic strength intrinsic to lexical items: A corpus study on tone reduction in Tone4+Tone4 words in Mandarin Chinese , 2016, 2016 10th International Symposium on Chinese Spoken Language Processing (ISCSLP).

[3]  Keikichi Hirose,et al.  Tone nucleus modeling for Chinese lexical tone recognition , 2004, Speech Commun..

[4]  Chilin Shih.  Linking phonology and phonetics: An implementation model of tones , 2008 .

[5]  Lianhong Cai,et al.  A New Prosodic Strength Calculation Method for Prosody Reduction Modeling , 2008, 2008 6th International Symposium on Chinese Spoken Language Processing.

[6]  C. En.  A SYSTEM OF "TONE-LETTERS" , 1980 .

[7]  Keikichi Hirose,et al.  Discriminating Chinese lexical tones by anchoring F0 features , 2000, INTERSPEECH.

[8]  Chilin Shih,et al.  Quantitative measurement of prosodic strength in Mandarin , 2003, Speech Commun..

[9]  趙 元任,et al.  A grammar of spoken Chinese = 中國話的文法 , 1968 .

[10]  Keikichi Hirose,et al.  Tone nucleus-based multi-level robust acoustic tonal modeling of sentential F0 variations for Chinese continuous speech tone recognition , 2005, Speech Commun..

[11]  John J. Ohala,et al.  Gesture, Segment, Prosody: The segment: primitive or derived? , 1992 .

[12]  Chilin Shih,et al.  Prosody modeling with soft templates , 2003, Speech Commun..

[13]  Wu Hua,et al.  Speech corpus of Chinese discourse and the phonetic research , 2000, INTERSPEECH.

[14]  Huang,et al.  Experimental study on downstep in Chinese intonation , 2007 .

[15]  Min Liu,et al.  A preliminary study on acoustic correlates of tone2+tone2 disyllabic word stress in Mandarin , 2014, INTERSPEECH.

[16]  Yi Xu,et al.  Contextual tonal variation in Mandarin Chinese , 1993 .

[17]  Björn Lindblom,et al.  Explaining Phonetic Variation: A Sketch of the H&H Theory , 1990 .