Speech-to-Animation
Speech-to-Animation (STA) is a technology developed exclusively by FaceUnity, which uses voice message to drive avatar to speak with emotions and movements. The technology combines deep learning neural networks with computer graphics to enable computers to understand the content of speech and fine-tune the avatar's lip movements, facial expressions, and body gestures to create realistic avatar animations.
Demo
Person
Background
Action
Emotion
Chinese Input Only
159/60
仅支持上传 30 秒以内的 .mp3 文件
为了保证合成效果，配音请勿添加背景音乐
Technical Advantage
-
Real-time Speech Analysis
Accurately analyze information in the speech.and transform it into the phoneme script that can drive the avatar.
-
Precise Lip Animation
It defines dozens of basic lip styles and lip coefficients to realize the lip shape transition of word-to-word pronunciation.
-
Rich Expression&Emotion Feedback
Using 55 blendshapes and expression coefficients to show all common expressions of human.
-
Limb Movement Animation
With mature specifications of skeleton and movement production. Models can be driven by audio content, including the head and the body.
-
Multi-style Avatars
Supporting the rendering and driving of various types of avatars,including cartoon characters, animal style, real human style, and high-rendering simulation style.
Application Scenario
-
AI Virtual Assistant
Combined with NLP and TTS technology, it can provide customers an AI virtual assistant with real-time conversation and rich facial expression feedback, which making the interaction between human and machine more vivid and interesting.
MORE -
AI Virtual Anchor
With STA technology, editors can quickly generate videos of AI virtual anchor with synchronized lip shapes, rich facial expressions and model actions by simply inputting text content.
MORE
