preprint · arXiv (2026) · Jul 22, 2026
Super Star: Towards Streaming Real-time Interactive Agents for Digital Humans
Wentao Jiang, Youchen Xie, Haidi Fan, Yajing Chen, Xin Wang, Ye Shi, Jingya Wang
In plain English
A framework for generating body gestures online for interactive digital humans using only the speech available so far. It pairs a streaming speech module with a causal gesture model and reports a better latency-quality trade-off and higher user preference than baselines.
gesture generation
digital humans
real-time avatars
audio-driven animation