Summary: | 碩士 === 文藻外語大學 === 多國語複譯研究所 === 107 === Interpreting is a highly demanded job which requires concentration, excellent language abilities and the capability to overcome high pressure that accompanies the job. Interpreter training institutes tend to use interpretation exams to assess interpreting students’ performance and to ensure that students can meet the needs of the industry. Without a doubt, one must invest a formidable amount of time in practising interpreting if one wants to pass the exam. In hopes of assisting students in practising interpreting, this study aims to implement an assistance system for interpreter training which takes usability and affordability into account. This system utilizes online speech recognition services and JavaScript. Via this system, the student's output for the pre-set video will be transcribed and converted into a subtitle file nearly on-the-fly. The subtitle file will then be merged with professional interpreter's output and later presented to the students on the same video. Therefore, students can visually compare both outputs and learn how to improve their performance. A single-group pretest-posttest design, a semi-structured interview and text analysis were employed to explore whether the system has achieved the purpose set forth in this study. The experiment and text analysis show that the result from pretest to posttest is statistically significantly improved. The interviewees also recognise the usability of the system and its suitability on practising. However, with the hindrance of the poor performance of Mandarin speech recognition as well as the small sample size, it cannot be suggested that the speech-recognition-based assistance system for interpreter training can practically improve its users' output.
|