Laughter matters! From an analysis of a very large corpus of naturally-occurring conversational speech we have confirmed that approximately one in ten utterances contains laughter. From among these laughing utterances, we were able to distinguish four types of laughter according to what each revealed about the speaker's affective state, and we were able to recognise these different types automatically, by use of Hidden Markov Models trained on laugh segments, with a success rate of greater than 75%. The paper also presents a speech synthesis interface that enables the control of such emotional expression for use in an interactive conversation.
Cite as: Campbell, N., Kashioka, H., Ohara, R. (2005) No laughing matter. Proc. Interspeech 2005, 465-468, doi: 10.21437/Interspeech.2005-317
@inproceedings{campbell05_interspeech,
author={Nick Campbell and Hideki Kashioka and Ryo Ohara},
title={{No laughing matter}},
year=2005,
booktitle={Proc. Interspeech 2005},
pages={465--468},
doi={10.21437/Interspeech.2005-317}
}