對於英文作為第二語言ESL(English as a second language)的學習者,學習英文一直是一大挑戰,往往透過電腦輔助語言學習CALL(Computer Assisted Language Learning System)能降低在人與人對話時所產生的恐懼感,還能提升ESL學習者的學習效率。傳統CALL的測驗方式多半是以語音撥放讓使用者聽,聽完後去點選覺得正確的發音,這類的CALL無法即時辨識使用者的發音並且回饋給使用者正確的發音建議。本研究利用GOOGLE釋出的語音辨識SR(Speech Recognition)程式應用程式介面API(Application Programming Interface)和ResponsiveVoice.JS語音合成API去設計一套CAPT(Computer Assisted Pronunciation Training)讓使用者去測驗和練習,並開發的一套線上的語音、發音辨識和提供建議的CALL。使用者能透過系統測驗和練習提高對英文發音的正確性,也會記錄使用者的測驗過程並會回饋使用者個人的發音錯誤狀況統計以提供發音更正建議。經過邀請27位竹山高中學生進行實驗,能證實本研究的確能大大的提升使用者對英文發音的正確性。
Learning English as a second language (ESL) has always been a challenge. Various Computer Assisted Language Learning (CALL) systems help a lots. These Conventional CALL systems have focused on providing pre-recorded audio and let users repeatedly listen to the pre-recorded audios to enhance their listening comprehension; however, they do not provide a real-time voice recognition and feedback. Based on the art-of-the-state technologies of voice recognition and speech synthesis, we build an on-line CALL system that can promptly recognize users’ pronunciation, instantly give feedback and perform a thorough analysis of personal pronunciation. An experiment, including pre-test, exercise and post-test was conducted to examine the effectiveness of the system. 27 high school students were invited to take part in the experiment and have all accomplished the tasks. The results show that our system, which include the complete exercises and thorough analysis report, greatly enhances users’ pronunciation.