Implementation of Respiration in Articulatory Synthesis Using a Pressure-Volume Lung Model

Keisuke Tanihara, Shogo Yonekura, Yasuo Kuniyoshi


In previous studies of the 1D vocal tract model of articulatory synthesis, subglottal pressure is typically regarded as constant, ignoring its dynamics. However, human vocalization is initially generated by glottal airflow via subglottal pressure change. This change is caused by the expansion and contraction of the lungs. In the current study, we propose a new pressure-volume model that relates pressure changes to volume changes of the human lung. Using this model, the behavior of the human lung can be integrated with articulatory synthesis. This model produces positive and negative subglottal pressure corresponding to expiration and inspiration respectively. In addition, breathing could be implemented in the proposed model. This implementation would expand the possibilities for articulatory synthesis.


 DOI: 10.21437/Interspeech.2018-1080

Cite as: Tanihara, K., Yonekura, S., Kuniyoshi, Y. (2018) Implementation of Respiration in Articulatory Synthesis Using a Pressure-Volume Lung Model. Proc. Interspeech 2018, 2504-2508, DOI: 10.21437/Interspeech.2018-1080.


@inproceedings{Tanihara2018,
  author={Keisuke Tanihara and Shogo Yonekura and Yasuo Kuniyoshi},
  title={Implementation of Respiration in Articulatory Synthesis Using a Pressure-Volume Lung Model},
  year=2018,
  booktitle={Proc. Interspeech 2018},
  pages={2504--2508},
  doi={10.21437/Interspeech.2018-1080},
  url={http://dx.doi.org/10.21437/Interspeech.2018-1080}
}