Cue-based annotation and analysis of prosodic boundary events

Alejna Brugos, Mara Breen, Nanette Veilleux, Jonathan Barnes, Stefanie Shattuck-Hufnagel

Prosodic categories, like other grammatical categories, are realized with wide variability, yet listeners interpret linguistic meaning with apparent ease. ToBI aims to capture the linguistically meaningful prosodic elements of utterances, but does not capture the variability in acoustic cues that the labeller (and listener) must interpret in order to assign distinctive categories. Despite a long history of empirical work demonstrating the importance of individual cues in signalling prosodic events, little previous work has explored how these cues might systematically combine, trade off and vary in magnitude across ToBI categories. Following recent proposals of cue-based annotation of segments, and disfluent and stuttered speech prosody, the current paper assesses the role of six timing, pitch, and voice quality cues in signalling boundary events in a corpus of independently ToBI-labelled fluent American English speech. Results demonstrate that each cue accounts for unique variance in break index level, and that break index levels increase with the number of cues present per token. The data further suggest that labeller uncertainty is more frequent for tokens that exhibit only a subset of the cues. Future work will investigate annotation of cues to prominence, as well as additional cues to boundaries and grouping.

 DOI: 10.21437/SpeechProsody.2018-50

Cite as: Brugos, A., Breen, M., Veilleux, N., Barnes, J., Shattuck-Hufnagel, S. (2018) Cue-based annotation and analysis of prosodic boundary events. Proc. 9th International Conference on Speech Prosody 2018, 245-249, DOI: 10.21437/SpeechProsody.2018-50.

  author={Alejna Brugos and Mara Breen and Nanette Veilleux and Jonathan Barnes and Stefanie {Shattuck-Hufnagel}},
  title={Cue-based annotation and analysis of prosodic boundary events},
  booktitle={Proc. 9th International Conference on Speech Prosody 2018},