[2W] 国際セッションⅠ

[W-04] Improving International Learners’ Engagement in Japanese Online Engineering Courses by Adding Automatically English Dubbed Videos

○Sasipa Boonyubol¹, Do Ngoc Khanh¹, Toru Nagahama¹, Jeffrey S. Cross¹ (1. Tokyo Institute of Technology)

キーワード：MOOCs, online course, dubbed video, text-to-speech, speaking rate, Google API, edX

Massive open online courses (MOOCs) have become an alternative learning platform in recent years and the number of learners has increased recently due to the COVID-19 pandemic. While the majority of MOOCs are taught in English, some non-English MOOCs, e.g., those offered in Japanese, usually provide some content in English, such as subtitles, in order to attract more learners. With the aim of increasing the number of international learners from English-speaking countries, a time-efficient and low-cost process for video lecture dubbing into English was developed to circumvent the learners having to read subtitles. Automatic dubbing was done using Google Text-to-Speech (TTS) API for two different online engineering courses originally recorded with Japanese lectures. In this paper, the stepwise video dubbing process is described in detail to produce an automatically English dubbed lecture video from a video transcript. The algorithm to mitigate the problem of the audio-video synchronization is introduced. Preliminary results from the post-course survey on learners’ satisfaction towards the English automatically dubbed videos are also presented.