This note describes a framework used to simulate three basic emotional styles by means of prosodic transplantation techniques applied to the output of a corpus based speech synthesis system. The target pitch profiles together with duration and energy constraints have been obtained applying simple rules inferred from the analysis of a small corpus, recorded in three emotional styles. Results of perceptual tests show that styles are well recognized even if the acoustical quality, in some cases, degrades.
Cite as: Zovato, E., Pacchiotti, A., Quazza, S., Sandri, S. (2004) Towards emotional speech synthesis: a rule based approach. Proc. 5th ISCA Workshop on Speech Synthesis (SSW 5), 219-220
@inproceedings{zovato04_ssw,
author={Enrico Zovato and Alberto Pacchiotti and Silvia Quazza and Stefano Sandri},
title={{Towards emotional speech synthesis: a rule based approach}},
year=2004,
booktitle={Proc. 5th ISCA Workshop on Speech Synthesis (SSW 5)},
pages={219--220}
}