8th European Conference on Speech Communication and Technology

Geneva, Switzerland
September 1-4, 2003


Acoustic Variations of Focused Disyllabic Words in Mandarin Chinese: Analysis, Synthesis and Perception

Zhenglai Gu, Hiroki Mori, Hideki Kasuya

Utsunomiya University, Japan

The focus effects on acoustic correlates include both prosodic and segmental modifications. Analysis of 35 focused words in a carrier sentences uttered by 2 male and 3 female speakers has shown that: (1) there is a significant asymmetry of vowel duration as well as F0 range between the pre-stressed and post-stressed syllables, implying that different strategies are employed in the task of focusing disyllabic words, i.e., emphasizing the first syllable as well as weakening the second syllable for the former, but emphasizing the second syllable only for the latter; (2) the tonal combinations significantly affect the variations of both the vowel duration and F0 range; (3) the formant frequencies (F1, F2) are changed systematically in a way that that the formants of the vowels plotted in the (F1, F2) plane were stretched outwards. Perceptual validation of the relative importance of these acoustic cues for signaling a focal word has been accomplished. Results of the perception experiment indicate that F0 is the dominant cue closely related to the judgment of focused word and the other two cues, duration and formant frequencies contribute less to the judgment.

Full Paper

Bibliographic reference.  Gu, Zhenglai / Mori, Hiroki / Kasuya, Hideki (2003): "Acoustic variations of focused disyllabic words in Mandarin Chinese: analysis, synthesis and perception", In EUROSPEECH-2003, 2429-2432.