Developing intelligible public address announcements in emergencies based on speech production and perception

2021 Fiscal Year Final Research Report

• Project/Area Number: 18K04666
• Research Category: Grant-in-Aid for Scientific Research (C)
• Allocation Type: Multi-year Fund
• Section: 一般
• Review Section: Basic Section 25030:Disaster prevention engineering-related
• Research Institution: Tokai University
• Principal Investigator: Hodoshima Nao 東海大学, 情報通信学部, 准教授 (40453609)
• Project Period (FY): 2018-04-01 – 2022-03-31
• Keywords: 音声 / 音声生成 / 音声知覚 / 明瞭度 / 残響 / 公共空間 / 高齢者 / 非母語話者

Outline of Final Research Achievements

This study aims to develop intelligible public address announcements in reverberant public spaces such as train stations and airports. Since people modify the way they speak in challenging acoustic environments and by speakers' intentions, this study proposed two speech modification approaches; speech spoken in the presence of reverberation and speech spoken in urgent style. Listening tests and acoustical analyses were conducted in reverberant environments simulating public spaces. Both proposed approaches were more intelligible than normally spoken speech for younger and older adults under the specific speaker and listening conditions.

Free Research Field

音声処理

Academic Significance and Societal Importance of the Research Achievements

駅・空港・大型商業施設などの屋内では、残響過多で災害時の緊急放送が不明瞭となる。残響は若年者よりも高齢者や非母語話者の音声明瞭度を低下させるため、我が国の超高齢化社会と在留・訪日外国人の増加に及ぼす影響は大きい。本研究結果を受けて、公共空間の音声案内を提案法で録音・合成した音声を用いる、もしくは緊急放送・誘導のための音声ガイドラインの作成により、災害時により多くの人に適切な情報伝達が可能になることが考えられる。