Generating Japanese-accented English voices of 3 types according to the listening proficiency of Japanese ESL learners

• Item type: SIG Technical Reports(1)
• 公開日: 2024-02-22
• タイトル: Generating Japanese-accented English voices of 3 types according to the listening proficiency of Japanese ESL learners
• タイトル / 言語: en
• タイトル: Generating Japanese-accented English voices of 3 types according to the listening proficiency of Japanese ESL learners
• 言語: eng
• キーワード / 主題Scheme: Other
• キーワード / 主題: SP3
• 資源タイプ / 資源タイプ識別子: http://purl.org/coar/resource_type/c_18gh
• 資源タイプ: technical report
• 著者所属: Department of Information Science, Faculty of Science, Kochi University
• 著者所属: Department of Information Science, Faculty of Science, Kochi University
• 著者所属: Department of Information Science, Faculty of Science, Kochi University
• 著者所属(英): en
• 著者所属(英): Department of Information Science, Faculty of Science, Kochi University
• 著者所属(英): en
• 著者所属(英): Department of Information Science, Faculty of Science, Kochi University
• 著者所属(英): en
• 著者所属(英): Department of Information Science, Faculty of Science, Kochi University
• 著者名: Kiyotada, Mori; Yasuo, Miyoshi; Ryo, Okamoto
• 著者名(英): Kiyotada, Mori; Yasuo, Miyoshi; Ryo, Okamoto
• 論文抄録 / 内容記述タイプ: Other
• 論文抄録 / 内容記述: We have developed a voice generator for Japanese-accented English according to Japanese ESL learners’ listening skills. Three rules of Japanese-accented English for the voice generator are constructed by operating SSML based on features of English spoken by Japanese. Two evaluation experiments were conducted to confirm the clarity of the Japanese-accented English. The first experiment includes phoneme recognition and listening comprehension questions in American English. The second experiment includes the same content questions as the first set in three rules of Japanese-accented English. As a result, nine subjects showed results opposite to predictions: two of three rules generated by Japanese-accented English are more difficult to hear than native English.
• 論文抄録(英) / 内容記述タイプ: Other
• 論文抄録(英) / 内容記述: We have developed a voice generator for Japanese-accented English according to Japanese ESL learners’ listening skills. Three rules of Japanese-accented English for the voice generator are constructed by operating SSML based on features of English spoken by Japanese. Two evaluation experiments were conducted to confirm the clarity of the Japanese-accented English. The first experiment includes phoneme recognition and listening comprehension questions in American English. The second experiment includes the same content questions as the first set in three rules of Japanese-accented English. As a result, nine subjects showed results opposite to predictions: two of three rules generated by Japanese-accented English are more difficult to hear than native English.
• 書誌レコードID / 収録物識別子タイプ: NCID
• 書誌レコードID / 収録物識別子: AN10442647
• 書誌情報: 研究報告音声言語情報処理（SLP）; 巻 2024-SLP-151, 号 87, p. 1-4, 発行日 2024-02-22
• ISSN / 収録物識別子タイプ: ISSN
• ISSN / 収録物識別子: 2188-8663
• Notice: SIG Technical Reports are nonrefereed and hence may later appear in any journals, conferences, symposia, etc.
• 出版者 / 言語: ja
• 出版者: 情報処理学会