Descript’s Research team builds the generative video and audio models, and the multimodal understanding systems, behind Descripts most distinctive features: Video Regenerate and lipsync, video translation, and zero-shot voice and roomtone cloning. This isnt research for