Baidu has launched MuseSteamer 2.0, an upgraded image-to-video
AI model that adds natural human voices and ambient audio, removing the need for dubbing. Speech, lip, and body movements are reportedly synced across characters with millisecond precision.
