Press releases from Microsoft or partner companies

Microsoft Unveils VASA-1 AI, Creates Real-Time Lifelike Talking Videos from Photos

Apr 18, 2024, 12:27 AM

Microsoft Research has unveiled VASA-1, a groundbreaking artificial intelligence model capable of generating hyper-realistic talking face videos from just a single portrait photo and a speech audio clip. This new technology features precise lip-audio synchronization, lifelike facial behavior, and naturalistic head movements, all produced in real time. VASA-1 stands out for its ability to create photorealistic, emotive talking heads that capture a wide spectrum of emotions and realistic facial expressions. The technology's real-time capability and its minimal input requirements mark a significant advancement in AI-driven media creation.