Sound and picture in sync
ByteDance builds Seedance 2.0 to generate audio and video in one pass, so music, footsteps and voices can land on the same beat as the motion. Describe the sound the way you describe the picture: what is heard, when it starts and how loud it is. A singer holding a note or a drummer hitting a cymbal is a good first test, because any drift between sound and movement is easy to spot.

Cinematic 16:9 video keyframe, a tight close-up of a singer at a vintage chrome ribbon microphone, mouth open mid-note, throat tensed, eyes shut, breath faintly visible in cool air, a warm key light on one side of the face and a magenta rim light on the other. The frame captures the exact beat of a sung note, intimate and emotional.






