xADA - Expressive Audio Driven Animation for MetaHumans

Explore xADA, Epic Games' cutting-edge model for generating expressive and realistic facial, tongue, and head animation directly from speech audio in this 54-minute conference talk from Unreal Fest Orlando 2025. Learn how this innovative technology leverages state-of-the-art machine learning models to extract rich audio features and automatically create face and body animation for MetaHuman or MetaHuman-compatible models. Discover xADA's dual operational modes: a fully automatic system and a customizable mode that allows artists to specify and override emotion and blink timings for enhanced creative control. Watch a live demonstration of the technology in action while gaining insights into the underlying mechanisms of audio-driven animation and practical implementation strategies for both offline and streaming character speech animation in Unreal Engine. Master the integration of this advanced AI-powered animation system into your MetaHuman workflow to create more lifelike and emotionally expressive digital characters.