Automatic Analysis of Emotional Prosody in Mandarin Chinese: Applying the Momel Algorithm

Ting Wang, Hongwei Ding, Qiuwu Ma, Daniel Hirst


Based on the Momel algorithm, a set of acoustic parameters was analyzed automatically on Chinese emotional speech. Global prosodic features were calculated on the sentence level, which showed a concordance with the usual pattern reported in the literature. Local constraints were also considered on the syllable layer. An ANOVA showed that there were interactive effects among emotions, syllable positions and syllable tones on certain parameters. Further more, by examining the pitch movements, no significant difference was found between neutral speech and active emotional speech, which was different from the performance in non-tonal languages. However when reducing the tonal influence by using only tone 1 syllables in the utterance, this inverse effect disappeared. Hence we posited an interpretation that due to the existence of lexical tone in Mandarin Chinese, the paralinguistic use of pitch movements has been reduced.


 DOI: 10.21437/SpeechProsody.2014-12

Cite as: Wang, T., Ding, H., Ma, Q., Hirst, D. (2014) Automatic Analysis of Emotional Prosody in Mandarin Chinese: Applying the Momel Algorithm. Proc. 7th International Conference on Speech Prosody 2014, 120-124, DOI: 10.21437/SpeechProsody.2014-12.


@inproceedings{Wang2014,
  author={Ting Wang and Hongwei Ding and Qiuwu Ma and Daniel Hirst},
  title={{Automatic Analysis of Emotional Prosody in Mandarin Chinese: Applying the Momel Algorithm}},
  year=2014,
  booktitle={Proc. 7th International Conference on Speech Prosody 2014},
  pages={120--124},
  doi={10.21437/SpeechProsody.2014-12},
  url={http://dx.doi.org/10.21437/SpeechProsody.2014-12}
}