Auditory-Visual Speech Processing (AVSP) 2013
This paper reports on the extraction and analysis of head nods taken from motion capture data of spontaneous dialogue in Swedish. The head nods were extracted automatically and then manually classified in terms of gestures having a beat function or multifunctional gestures. Prosodic features were extracted from syllables co-occurring with the beat gestures. While the peak rotation of the nod is on average aligned with the stressed syllable, the results show considerable variation in fine temporal synchronization. The syllables co-occurring with the gestures generally show greater intensity, higher F0, and greater F0 range when compared to the mean across the entire dialogue. A functional analysis shows that the majority of the syllables belong to words bearing a focal accent.
Index Terms: Gestures, prosody, motion capture, beats, head nods, stressed syllable
Bibliographic reference. Alexanderson, Simon / House, David / Beskow, Jonas (2013): "Aspects of co-occurring syllables and head nods in spontaneous dialogue", In AVSP-2013, 169-172.