Breathing changes during listening and subsequent speech according to the speaker and the loudness level

Amélie Rochet-Capellan (1, 2) , Susanne Fuchs (2) , Leonardo Lancia (3) , Pascal Perrier (1)

1 GIPSA-MAGIC - GIPSA - Machines parlantes, Gestes oro-faciaux, Interaction Face-à-face, Communication augmentée
2 ZAS - Zentrum für Allgemeine Sprachwissenschaft [Berlin]
3 Max Planck Institute for Evolutionary Anthropology [Leipzig]

In the context of the interactive alignment model for multi-level adaptations in dialogue situations (Pickering & Garrod, 2004), our work focuses on the adaptation of listener's breathing to a speaker's breathing. Our general goal is to understand if, and how, listeners adapt their breathing to speakers' breathing: what are the information that the listener's breathing system is "catching" from the speakers' behavior to eventually change accordingly? Previous works showed that, in dialogue situation, listeners and speakers tend to synchronize their breathing at the time of turn-taking (Guaïtella, 1993; McFarland, 2001). Moreover, perception studies found that when breathing noise (e.g. when speakers inhale) is added to speech synthesis, the listeners' recall performance increases (Whalen et al., 1995). From the acoustical signal, listeners are also able to discriminate between speech produced starting at high lung volume versus speech started at a low lung volume level (Milstein & Watson, 2004). Breathing profiles during listening are also different than during breathing at rest, and could be an indicator of the perceptual process. In this context, Brown (1962) hypothesised that poor listeners may be less capable to adapt their breathing to the speaker's breathing when compared to better listeners. Recently, Stephens et al. (2010) pre-recorded speakers' acoustical productions and the co-occurring brain activity. Then, they monitored listeners' brain activity during the audio playback of the speakers' productions. They found some speaker-listener neural coupling, even if the speaker was not present. Using a situation analogous to Stephens et al. (2010), the present study investigates if listeners' breathing changes according to the speaker they listen to (male vs. female) and to the loudness level (normal vs. loud) of the speaker's voice. We also evaluate if breathing during speech produced right after listening differs according to the speaker and to the loudness level of the signal heard during the listening task.

Mots clés: Dialogue; breathing; coordination

Domaines: Psychologie et comportements

Amélie Rochet-Capellan, Susanne Fuchs, Leonardo Lancia, Pascal Perrier. Breathing changes during listening and subsequent speech according to the speaker and the loudness level. ISICS 2012: International Symposium on Imitation and Convergence in Speech, Sep 2012, Aix-en-Provence, France. ⟨hal-00743785⟩