The Master Guide / 10 Voice and lip sync · Speaking shots and lip sync
English lines and Seedance's own voice
English principal lines have their own decided route. Seedance's own voice is a named option for English and other languages it speaks well, never for Arabic.
English principal lines have their own decided route. Seedance's own voice is a named option for English and other languages it speaks well, never for Arabic.
The English route
The route for every English principal line (decided):
Record and approve the line with the approved ElevenLabs voice.
Give the recording to Seedance as the audio reference, audio on, the line quoted in English, the speech length and window stated outside the quotes (Template E).
Lay the approved recording's placed copy back on the DIALOGUE track, with the dry take beneath it on DIALOGUE ALT 1, switched off (see Sound design and mix → The placed voice), Seedance's returned audio muted, and check the lips at full size.
A shot whose lips don't hold gets one Sync-3 pass to the recording (the plate-length file built from the dry take, cut_off). Price every English principal-line shot as high risk with one Sync pass inside its budget (about 27 Higgsfield credits at 7 s; prices and their dates in Speaking shots and lip sync → Attempts and budget per line): Seedance re-times an audio reference, by about a fifth in takes made on this stack.
Seedance's own English voice serves one-line extras with no fixed voice, and principal lines only through the option below.
The lay-back is English only. It is the one route that lays a take back without a Sync pass first, and only where the full-size lip check passes. Arabic, and the Franco quote, are never laid back without Sync.
The evidence for English on Seedance 2.5, strongest first: Seedance's own voice [proven]; the audio-referenced take, which is the English route (decided); lip sync afterwards, onto a separate recording [best available, not yet validated].
Seedance's own voice
Option: use it when a principal line is in English or another language Seedance speaks well (creators ship it in Japanese, Korean and Italian), and you choose it for the project [proven]. It is never a default, and never used for Arabic.
One fixed voice description per character (decided), written once in plain words: age, timbre, pitch, pace, accent, texture. No real person's or performer's name.
The client approves the voice from a sample: one short audio-on clip of the character saying a line of the film, made with that description and priced in the anchors estimate. It sits in the anchors presentation's voices section like a recorded voice (see Client presentations → The anchors presentation, in its twelve sections).
The same description goes word for word into every shot where the character speaks. The line's emotion goes on its own line, never inside the description. The line itself is in quotes exactly as spoken, said once, with about 0.5–1 s of quiet before and after, and the lips still when not speaking. The prompt is Template E's own-voice form.
Check every take: the words against the canonical line, the lips at full size, a video-analysis report on the line, and your ear on the voice against the approved sample. A voice that drifts is regenerated with one diagnosed change. No Sync pass where the lips hold.
Write the choice in your decisions for the project, and the description in your shot plan.
Why: nine independent festival creators let Seedance voice their main lines. Soulless and The Elevator repeat one fixed voice description, unchanged in every prompt, with the emotion on a separate line.