Multilingual Video Translator and Transcriber

Created by team Taleemabad on February 23, 2024

In this project, the goal is to translate an English video into Spanish using a multi-step process. The initial step involves transcribing the audio from the English video using OpenAI Whisper models on Colab. Subsequently, the generated transcript undergoes translation utilizing GPT-4. To produce the translated voiceover, OpenAI's TTS-1 model is employed, but due to a character limit of 4096, the translated transcript is divided into manageable chunks. Voiceovers for each segment are then obtained, and the individual audio files are merged using ffmpeg. The final challenge lies in synchronizing the source video and audio, an ongoing effort requiring significant time and dedication. Despite this synchronization hurdle, the team is submitting the completed sections of the project, showcasing progress in the intricate process of translating and dubbing the source content.

Category tags: