英文摘要 |
隨著語音合成技術的發展,我們隨意輸入的詞句能夠以東京方言的方式進行語音合成。合成出來的語音質感生動逼真,足以達到實用化的水準,將來能期待運用在日語教育之中並帶給學習者正面影響。本論文將探討OJAD語音合成技術將如何合成—在表達情感上不可或缺卻未被重視的—文末句調,並提出語音合成技術尚待解決的問題點,最後提出語音合成技術導入日語語音教育的可能性及極限。調查主要結果為,凡合成之文末句調含有「音高上升兩次」或「詞調無效化」的時候依目前語音合成技術無法順利合成。在尚未有完善的一套日語語音教育之中,本研究藉由語音合成的實現度之調查與分析,闡述語音合成技術導入日語語音教育的可能性及日語教師的定位。 With rapid advances in technology, many improvements have been made on speech synthesis. We can input text in a way that can be synthesized in Tokyo dialect. The synthesized speech is vivid and realistic enough to reach a practical level. We can look forward to using it in Japanese education in the future. This study investigated the relationship between input text and speech synthesis of sentence-final intonation. We identified the problem in speech synthesis technology. We also identified the possibility and limit of teaching Japanese pronunciation using speech synthesis. In conclusion, in case of ''two pitch rises'' or ''invalidate accent'', speech synthesis will be wrong. Among the incomplete day teaching Japanese pronunciation, we investigation and analysis of the realization of speech synthesis, to reports on the feasibility and value of an approach to teaching Japanese pronunciation. |