ESCA Tutorial and Research Workshop on Speech Input/Output Assessment and Speech Databases Noordwijkerhout, The Netherlands September 20-23, 1989

Speech Database Projects in Japan: Present and Future

Katsuhiko Shirai (1), Hiroya Fujisaki (2), S. Itahashi (3)

(1) Department of Electrical Engineering, Waseda University, Tokyo, Japan (2) Dept. of Electronic Eng., Faculty of Engineering, University of Tokyo Hongo, Bunkyo-ku, Tokyo, Japan (3) Institute of Information Sciences and Electronics, University of Tsukuba, Sakura-mura, Niihari, Ibaraki, Japan

Current status of main Japanese speech database projects will be summarized. The first one is the JEIDA Japanese Common Speech Data Corpus. The database is composed of 323 words and aimed mainly for use in design and evaluation of speech recognition algorithms. A large-scale Japanese speech database has been constructed by ATR. The database consists of word speech and continuous speech. Another effort to make a speech database for common use has been made in the Japanese National Project on Advanced Man-Machine Interface Through Spoken Language supported by a Grant-in-Aid for Scientific Research from the Ministry of Education. The database consists of isolated utterances and sentence/discourse materials.

Full Paper

Bibliographic reference.  Shirai, Katsuhiko / Fujisaki, Hiroya / Itahashi, S. (1989): "Speech database projects in Japan: present and future", In SIOA-1989, Vol.2, 47-50.