The Master Guide / 07 Video with Seedance · Writing the prompt
Sound inside the shot
Seedance's own sound is the film's in-shot layer (decided). Generate it, keep it and direct it.
Seedance's own sound is the film's in-shot layer (decided). Generate it, keep it and direct it.
The AUDIO block comes last before STYLE in every shot without speech. Write, in order:
the space and the distance;
constant room tone;
two to six sounds in shot order, each tied to an action or a time;
one off-screen layer;
the exclusions, as one short plain sentence at the end: no music, no voices, no subtitles, and nothing that swells, whooshes, stings or fades at the start, at a cut or at the end.
Write the exclusions as a sentence, not a list of words, because exclusion words can leak into the sound [official guidance].
The rest of the in-shot sound:
Space and off-screen [proven as craft]: an off-screen layer where it helps, muted where Seedance makes it too loud. Cultural sounds pass the cultural check first (see The brief and research → Culture).
Music only from a source in frame, and then low under anything spoken (see Music → Suno always, hero tracks included).
Never regenerate a shot for its sound: mute the failed part; listen for bleed at cuts (see Sound design and mix → Keeping or muting in-shot sound).
Seedance voices one-line extras and walla outside Arabic only (decided). Seedance never voices Arabic: Arabic extras are recorded, and Arabic crowd murmur comes from real CC0 (public-domain) crowd recordings or voices recorded in ElevenLabs (see Speaking shots and lip sync → Other speaking routes).
Seedance's own voice. Option: use it when a principal line is in English or another language Seedance speaks well, and you choose it for the project [proven]. The client approves the voice from a sample. The sample is made with one fixed voice description (age, timbre, pitch, pace, accent, texture; never a person's name). Paste that description unchanged into every shot where the character speaks, with the emotion on its own line. Never for Arabic: every Arabic voice is recorded (see Speaking shots and lip sync → English lines and Seedance's own voice).
Audio off only for sound-off speaking plates, video edits and video-to-video, and shots with no usable sound. Audio-referenced speaking shots set audio as their route says (see Speaking shots and lip sync → Choosing the route per shot).