ShengShu Technology Unveils Vidu S2, Splitting Real-Time AI Video Into an Interactive Avatar Model and a Live Stream Editor
- ShengShu Technology unveiled Vidu S2 on September 15, 2026 as two models: Vidu S2-Avatar for continuous real-time digital-character interaction, and Vidu S2-Editing for real-time editing of an incoming video stream.
- S2-Avatar raises real-time output resolution to 720p from 540p in Vidu S1 and lets users introduce a new reference image mid-stream, so a character can pick up a shown object, change into a specified outfit, or enter a new scene while preserving state after the action ends.
- S2-Editing covers four real-time tasks on live video, style transfer, outfit change, subject replacement, and background replacement, each following the source video's motion, camera movement, and occlusions.
- S2-Avatar's output can feed a spatial video pipeline that renders synchronized left- and right-eye views for VR headsets, and S2-Editing can either edit monocular video before conversion or edit existing spatial video directly.
- ShengShu says end-to-end development of Vidu S2 was led by Jintao Zhang, a PhD student advised by Professor Jun Zhu.