Databases for the Creation of Voice Driven Teleservices
Swedish telephone speech database containing 5000 speakers over the landline telephone network and 1000 speakers recorded over mobile telephone networks
Due to the progress reached in speech processing technology more and more powerful voice driven teleservices can be implemented, which allow easy access to: information services (train table information), transaction services (home shopping), call processing services (voice mail handling) via the telenetwork. To implement these language specific spoken language resources, i.e. speech databases, lexica and related tools are needed. In order to be competitive with American companies starting with a large monolingual market the consortium of SpeechDat will lay the ground for European companies to be competitive when starting with a multilingual environment. The project aims at producing speech databases realising a large coverage of languages and applications. The main features are: coverage of applications (application-oriented words, phonetically rich sentences, spontaneous utterances, speaker verification), coverage of the 11 official European languages and variants, coverage of speaking styles (commands, carefully pronounced and spontaneous speech), coverage of environmental influences (mobile and fixed telephone network). Around 5000 speakers will be recorded for the official languages over the fixed network, while there will be 1000 speakers for the language variants, the mobile recordings (5 languages) and the speaker verification recordings (3 languages). For validation and distribution of the data bases the European Language Resource Association (ELRA) will be involved in the project. The following languages (and variants) will be covered: Danish, Dutch, Flemish, British English, Welsh, Finnish, French, Belgian French, Swiss French, Luxembourgish French, German, Swiss German, Luxembourgish German, Greek, Italian, Portuguese, Slovenian, Spanish, Swedish and Finnish Swedish.
Group: Speech Communication and Technology
Kjell Elenius (Project leader)
Funding: EU (FP5 LE2-4001)
Duration: 1996-03-01 - 1998-02-28
Keywords: speech database telephone mobile fixed landline
Int Journal of Speech Technology, 3, 119-127. (2000). Experiences from collecting two Swedish telephone speech databases..
Proc. of LREC 2000, 2nd Intl Conf on Language Resources and Evaluation (pp. 1351-1356). Athens, Greece. [pdf] (2000). The COST 249 SpeechDat multilingual reference recogniser. In Gavrilidou, M., Caryannis, G., Markantonatou, S., Piperidis, S., & Stainhaouer, G. (Eds.),
Proc of ICSLP 2000, 6th Intl Conf on Spoken Language Processing (pp. 370-373). Beijing. [pdf] (2000). A noise robust multilingual reference recogniser based on SpeechDat(II). In
Proc of ICPhS-99 (pp. 1741-1744). (1999). Experiences from building two large telephone speech databases for Swedish.. In
Proc of Eurospeech 99 (pp. 2243-2246). (1999). Two Swedish SpeechDat databases - some experiences and results. In
Proc of Fonetik 99 (pp. 45-48). (1999). Two Swedish telephone speech databases.. In
TMH-QPSR, 40(1-2), 051-056. [pdf] (1999). Experiences from building two large telephone speech databases for Swedish.
Proc of Fonetik -97, Dept of Phonetics, Phonum 4 (pp. 61-64). Lövånger/Umeå. [pdf]
(1997). SpeechDat - Speech databases for creation of voice driven teleservices. In Bannert, R., Heldner, M., Sullivan, K., & Wretling, P. (Eds.),