An Integrated Video Transcription and Translation System

The workflow is broken down into steps that're really time consuming and do not work very well. This is because translating and transcribing videos often requires using than one tool. Some video translation platforms only offer transcription services for videos while other platforms only offer translation services for videos. This can result in a translation outcome for the videos because of potential problems or inconsistencies in the translation of the videos. The video translation system described in this paper offers a platform for translation and transcription services for videos in many languages. It uses a speech-to-text converter and an audio extractor for the videos. To use the system, you need to either upload a video file or put a link to a video into the system. When you upload a video file the system extracts the audio from the video. Saves it in formats like WAV or MP3. When you add a video link to the system yt-dlp gets the audio from the video in quality. The Whisper Base Model processes the extracted audio from the video to make a text transcription in the same language as the audio from the video. If the text from the video needs to be translated the mT5 multilingual transformer model is used to translate the text from the video while keeping the meaning and sentence structure of the text, from the video.