Abstract
(Englisch)
|
Due to the progress reached nowadays in speech processing technology, more and more powerful voice driven teleservices can be implemented. They allow, for instance, easy access to information services (e.g. train table information), transaction services (e.g. home shopping), and call processing services (e.g. voice mail handling) via the telephone network. Many European companies are active in the field of creating such services and delivering the required speech technology. However, for research purposes and for the implementation of the speech processing technology (speech recognition and speaker verification), spoken language reˆsources are necessary, including speech databases, lexica, and related tools.
The goal of the SpeechDat project was to produce, standardize, evaluate and disseminate large speech databases covering a wide range of languages (most of the European languages) and applications.
In the framework of this project, IDIAP participated to the development of annotation software tools, and the specification of the databases. IDIAP was responsible for collecting and labeling the data in Swiss French and Swiss Germa, in collaboration with ETH, and with Swisscom as an associated partner. This project was successfully terminated at the end of 1998. By that time, IDIAP had released their contribution to SpeechDat data, which represents a subset of the Swiss-French Polyphone database satisfying the European Speechdat format. Its distribution has been entrusted to ELRA.
More information about the contribution of IDIAP is given in the attached document and on to web site of the project (http://www.phonnetik.uni-munchen.de/SpeechDat.html).
|
