This paper describes the new SPRUCE (SPeech Response from Unconstrained English) text-to-speech synthesis system being developed at the Universities of Bristol and Essex in the UKThe system exhibits a marked improvement in output naturalness over current systems by incorporating recent fundamental changes in the underlying theory of speech production. Based on a large scale dictionary the system performs a grammatical parse on the input text to produce improved prosodies. The final output is based on stored normalised parametric analyses of natural syllables to produce improved segmental rendering. To date we have been able to demonstrate an early version of the system which shows considerable promise. Keywords: text-to-speech synthesis - naturalness in synthesis - syllables
Cite as: Lewis, E., Tatham, M.A.A. (1991) A new text-to-speech synthesis system. Proc. 2nd European Conference on Speech Communication and Technology (Eurospeech 1991), 1235-1238, doi: 10.21437/Eurospeech.1991-283
@inproceedings{lewis91_eurospeech,
author={E. Lewis and Mark A. A. Tatham},
title={{A new text-to-speech synthesis system}},
year=1991,
booktitle={Proc. 2nd European Conference on Speech Communication and Technology (Eurospeech 1991)},
pages={1235--1238},
doi={10.21437/Eurospeech.1991-283}
}