The goal of this study is to evaluate to what extent listeners are capable of distinguishing between professional speaking styles based only on prosodic cues. The four speaking styles contemplated in the study – TV news, religious, political, and interview speech – were recorded in the context of normal use from Brazilian television and low-pass filtered to remove semantic information. The participants of a perception study were presented with one minute samples of the filtered speech and asked to choose which of two speaking styles each sample represents. The results of the perception experiment demonstrate that listeners are able to identify the speaking style with 90% accuracy, proving that even when semantic and lexical information is removed from the signal, there remains sufficient information in the prosodic cues to allow listeners to identify these speaking styles.
Index Terms: prosody, speech perception, speaking style, professional voice
Cite as: Castro, L., Freitas, M., Moraes, J.A.d., Serridge, B. (2010) Listeners' ability to identify professional speaking styles based on prosodic cues. Proc. Speech Prosody 2010, paper 326
@inproceedings{castro10b_speechprosody,
author={Luciana Castro and Myrian Freitas and João Antônio de Moraes and Ben Serridge},
title={{Listeners' ability to identify professional speaking styles based on prosodic cues}},
year=2010,
booktitle={Proc. Speech Prosody 2010},
pages={paper 326}
}