Sixth International Conference on Spoken Language Processing
(ICSLP 2000)

Beijing, China
October 16-20, 2000

A Comparative Study on Acoustic and Linguistic Characteristics Using Speech from Human-To-Human and Human-To-Machine Conversations

Toshiyuki Takezawa, Fumiaki Sugaya, Masaki Naito (1), Seiichi Yamamoto

ATR Spoken Language Translation Research Laboratories, Seika-cho, Soraku-gun, Kyoto, Japan
(1) currently with KDD R&D Laboratories, Inc.

Speech translation and dialogue systems must accept conversational speech. In this paper, we discuss acoustic and linguistic characteristics based on results of speech recognition experiments using speech from human-to-human and human-to-machine conversations. Conversational speech inputs to machines consist of frozen expressions such as greetings and yes/no statements, and informative individual expressions like numerical data such as dates and telephone numbers. The former has a lower perplexity and acoustic characteristics close to spontaneous speech. The latter has a higher perplexity and acoustic characteristics close to read speech. Each utterance or each inter-pausal unit can be classified into the former or the latter. This new knowledge will help future research on speech translation and dialogue systems.

Full Paper

Bibliographic reference.  Takezawa, Toshiyuki / Sugaya, Fumiaki / Naito, Masaki / Yamamoto, Seiichi (2000): "A comparative study on acoustic and linguistic characteristics using speech from human-to-human and human-to-machine conversations", In ICSLP-2000, vol.3, 522-525.