Third ESCA/COCOSDA Workshop on Speech Synthesis

November 26-29, 1998
Jenolan Caves House, Blue Mountains, NSW, Australia

Report on the Third ESCA TTS Workshop Evaluation Procedure

Jan P.H. van Santen (1), Louis C.W. Pols (2), Masanobu Abe (3), Dan Kahn (4), Eric Keller (5), Julie Vonwiller (6)

(1) Bell Labs, Lucent Technologies, Murray Hill, NJ, USA
(2) University of Amsterdam, The Netherlands
(3) NTT, Kanagawa, Japan
(4) E-Speech, Gillette, NJ, USA
(5) University of Lausanne, Switzerland
(6) APPEN Speech Technology, Gordon, NSW, Australia

This paper provides a description and rationale for the Evaluation Procedure taking place at the Workshop. The procedure has three goals. First, setting a precedent of providing conference participants with a more candid and thorough picture of the quality of current TTS systems than is usually available in the form of prepared conference demonstrations. Second, providing results that will be informative for TTS systems developers. Third, stimulating a discussion and contributing to a consensus building process on text-to-speech synthesis evaluation.


Full Paper

Bibliographic reference.  Santen, Jan P.H. van / Pols, Louis C.W. / Abe, Masanobu / Kahn, Dan / Keller, Eric / Vonwiller, Julie (1998): "Report on the Third ESCA TTS Workshop Evaluation Procedure", In SSW3-1998, 329-332.