Research Article

SubGPT: An Automated Tool for Subtitle-Template Translation

Authors

  • Hussein Abu-Rayyash Kent State University, Department of Modern & Classical Language Studies, USA
  • Ali Al-Ramadan Kent State University, Department of Computer Science, USA

Abstract

SubGPT is an automated subtitle translation tool that integrates OpenAI's GPT to generate translations from Source Language Subtitle Template (SRT) files. The tool employs customized prompts to produce translations adapted for different dialects and vernaculars, while maintaining the spatial and temporal aspects of the original subtitles. A mixed-methods approach is used to evaluate SubGPT's performance, combining cosine similarity metrics and qualitative assessments from professional subtitlers. The results demonstrate the tool's effectiveness in handling diverse linguistic challenges and its potential for improving efficiency and accessibility in the audio-visual translation (AVT) industry. The tool contributes to the advancement of the AVT field through the integration of large language models for culturally and linguistically automated translation, with potential applications in different domains such as entertainment and education.

Article information

Journal

International Journal of Linguistics, Literature and Translation

Volume (Issue)

7 (6)

Pages

126-134

Published

2024-06-22

How to Cite

Abu-Rayyash, H., & Al-Ramadan, A. (2024). SubGPT: An Automated Tool for Subtitle-Template Translation. International Journal of Linguistics, Literature and Translation, 7(6), 126–134. https://doi.org/10.32996/ijllt.2024.7.6.14

Downloads

Keywords:

Audio-visual translation, Automated translation, Translation technology, Subtitling tools, GPT, English-Arabic, Subtitle templates