What problem does it solve? AI-generated video clips often feel lifeless because sound is forgotten until the end, and dialogue frequently desyncs from lip movement. This Skill structures audio direction for Seedance 2.0 so dialogue, sound effects, music, and ambience are planned as part of the prompt rather than patched in later. ## Core Features & Use Cases - Dialogue & Lip-Sync Direction: Writes short quoted lines assigned to named speakers with locked framing constraints that improve mouth accuracy, including per-language reliability guidance. - Sound Layer Composition: Builds compact layered audio blocks (Dialogue, Sound, SFX, Music, Silence) so each scene gets room tone, action evidence, and dramatic beats. - Audio Reference & Beat-Sync: Maps @Audio1 references to tempo, mood, or voice-tone roles, and resolves conflicts when audio and video references compete. - Desync Troubleshooting: Diagnoses failed lip-sync, wrong-speaker output, or ignored audio and applies fixes like shortening lines, locking the camera, and cleaning reference roles. - Use Case: A creator producing a two-character product ad in Mandarin uses this Skill to assign one speaker per clip, attach a voice reference for lip-sync, and layer a magnetic click SFX with low room tone. ## Quick Start Ask the agent to write a Seedance 2.0 audio prompt with lip-synced dialogue and layered sound effects for your scene.
