AI-Powered Multilingual Dubbing

Artificial Intelligence (AI) is revolutionizing multilingual media by enabling efficient, automated video transcription, translation, and dubbing. This paper presents an end-to-end pipeline integrating state-of-the-art tools—Whisper for speech recognition, MarianMT for translation, and IndexTTS2 for expressive text-to-speech synthesis. These components together support the seamless localization of video content across languages. The system enhances accuracy, reduces human labor, and maintains voice fidelity and timing. This paper includes an introduction to the pipeline, literature reviews, model architecture, process flow, research methodology, experimental outcomes, benefits, limitations, and conclusion.