Sony AI

Sony AI, a research subsidiary of Sony Group Corporation whose work spans Sony Pictures Entertainment's entertainment assets, is developing AI systems for dubbing and translating entertainment content, including a lip-sync synchronization method called DubWise and an emotional voice conversion framework called EmoReg. These tools are at the research stage and have not been confirmed as deployed in SPE productions.

Sony AI's 'AI for Creators' research project includes work on dubbing, translation, and voice synthesis aimed at the entertainment sector. DubWise, presented in a June 2024 research paper, is a method that uses large language models and visual cues from video (such as lip movements) to synchronize dubbed audio with on-screen performance. EmoReg, presented at AAAI 2025, is a diffusion-based emotional voice conversion framework designed to replicate and control emotional intensity in AI-generated speech for dubbing applications. The CASAT framework was also presented at AAAI 2025 for context- and style-aware machine translation for Indian-language entertainment. These are active research projects; deployment in commercial SPE productions has not been confirmed by primary sources.