Speaking style prosodic variation: an 8-hour 9-style corpus study

Jean Philippe Goldman, Tea Prs̈ir, George Christodoulides, Antoine Auchlin


This paper presents the results of a prosodic and phonostylistic analysis based on C-PhonoGenre, a 9-hour-long spoken French corpus, including 10 speakers on average recorded in 10 speaking situations. The corpus was automatically segmented at phonetic, syllabic, word levels (EasyAlign), and larger pause- separated units. Part-of-speech annotation (DisMo) and prominent syllable detection (ProsoProm) was added automatically. The corpus was also manually annotated at the syllabic level for stylistic variants, such as post-tonic schwas, liaisons, elisions, disfluencies, audible breaths and noises. Acoustic analyses (ProsoReport, DurationAnalyser) provide more than 100 micro- and macro-prosodic measures, which we correlate with the phonostylistic and linguistic annotation. This analysis finally yields a contrastive, fine-grained prosometric description of phonostylistic and situational variation, over 4 situational gradual dimensions: audience, media, preparation, and interactivity. Further statistical analysis was carried out to explore the discriminative and explanatory power of combinations of prosodic measures.


 DOI: 10.21437/SpeechProsody.2014-9

Cite as: Goldman, J.P., Prs̈ir, T., Christodoulides, G., Auchlin, A. (2014) Speaking style prosodic variation: an 8-hour 9-style corpus study. Proc. 7th International Conference on Speech Prosody 2014, 105-109, DOI: 10.21437/SpeechProsody.2014-9.


@inproceedings{Goldman2014,
  author={Jean Philippe Goldman and Tea Prs̈ir and George Christodoulides and Antoine Auchlin},
  title={{Speaking style prosodic variation: an 8-hour 9-style corpus study}},
  year=2014,
  booktitle={Proc. 7th International Conference on Speech Prosody 2014},
  pages={105--109},
  doi={10.21437/SpeechProsody.2014-9},
  url={http://dx.doi.org/10.21437/SpeechProsody.2014-9}
}