The Master Guide / 07 Video with Seedance · Speaking and singing recipes
12. Arabic audio-referenced close-up
A close-up Arabic hero line: a laugh breaking into the line, a sob, a shout, a punchline's timing. Here, late in the evening in a Cairo living room, a tired wife tells her husband, who sits off screen, that she is not angry, only worn out.
When to use it: a close-up Arabic hero line: a laugh breaking into the line, a sob, a shout, a punchline's timing. Here, late in the evening in a Cairo living room, a tired wife tells her husband, who sits off screen, that she is not angry, only worn out.
Methods:
Route 1, the audio-driven plate. Option: use it when a close-up hero line needs the body and breath to ride the voice [single strong source]. Seedance hears the approved recording and acts to it, then the exact take ships through lip sync. Used here because the line turns on a tired breath between its two phrases; this film's plan names Route 1 for this line and prices it (see Speaking shots and lip sync → Route 1, for close-up hero lines).
Otherwise the Arabic default is the sound-off plate, as in Video with Seedance → 8. Dialogue close-up.
Shared rules: the rules every recipe keeps, and the five rules for speaking and singing shots (template and route, the camera during speech, the angle, the native reviewer, the placed voice): Video with Seedance → How to read a recipe.
References: 1 Salma's sheet; 2 her identity close-up; 3 the Cairo living room's master plate; audio 1 the approved take, voice only, 2.6 s of speech, padded with room tone to the plate's length.
Settings: omni-reference · audio on · 5 s (start 1.0 s, speech 2.6 s long, ending at 3.6 s, tail 1.4 s) · 16:9.
Cost: 480p draft 15 cr (5 s × 3 cr) · 1080p finish 80 cr (5 s × 16 cr).
If Draft Mode refuses the audio reference: draft at 720p, for 35 cr.
After the finish: the returned audio is discarded, then the lip-sync pass at the measured onset.
Budget: high risk (a hero speaking shot, the lip-sync pass included): 310 cr expected (5 s × 62), 615 cr ceiling (5 s × 123).
What to change, what to keep: Template E (Speaking shots and lip sync → Templates C, E and F).
Yours to rewrite:
the Arabic line, in «…» on its own line, exactly as approved;
the times in VOICE AND TIMING and ACTION / TIMELINE, measured from your take;
the voice, in a few words;
the listener's one visible change.
Keep word for word: the shared wording (Video with Seedance → How to read a recipe), plus:
the audio role line ("it gives the voice, accent, pace and pauses only, and nothing in the picture");
"Only words between quotation marks are said aloud; everything else is direction.";
in AUDIO, "close, dry dialogue" and "No music, no subtitles and no on-screen text."
The line, in Arabic script, on its own line in the prompt:
«أنا مش زعلانة… أنا تعبت.»
SCENE CONTEXT: Late evening in a Cairo living room. A tired wife tells her husband, who sits just off screen across the low table, that she is not angry, only worn out.
ACTIVE REFERENCES: @image1 is SALMA: a woman in her early thirties, slight, dark hair loosely tied back, a faded grey cardigan over a cream top; both hands out of frame. This sheet gives identity and wardrobe and nothing else: the face and hair come from the close-up panel alone, the wardrobe and body from the full-body panels alone; the front panel's cropped head is never repeated; leave out its grey ground, layout and flat light. Match it exactly. @image2 is SALMA'S FACE: the identity close-up, used for the face alone. @image3 is THE LIVING ROOM: the sofa, the low table with a cold glass of tea, and the curtained window, as architecture, materials and light only; it does not decide where the camera stands. @audio1 is SALMA'S VOICE: the approved recorded take; it gives the voice, accent, pace and pauses only, and nothing in the picture.
VOICE AND TIMING: SALMA, voice: tired, low, Egyptian Arabic, as @audio1. In @audio1 the speech runs 2.6 s; her line starts at 1.0 s and ends at 3.6 s.
FIRST FRAME AND BLOCKING (in motion from frame one): Medium close-up. SALMA sits on the sofa at frame centre, turned a quarter toward frame right, lips closed, her eyes already on the cold tea, a small swallow in her throat; hands out of frame.
FORMAT MODE: A single unbroken shot with no cuts of its own. Real time.
OPTICS: Medium close-up on a 29° diagonal field of view; her mouth clear and evenly lit; the room soft behind her; the lens holds steady.
CAMERA: Handheld on the shoulder, 2 m from her at her eye height, breathing with the operator. One very slow step in toward her across the 5 s, about 35 cm, travelling, seen as her face growing by about a fifth, so that the cold tea slides out past the lower edge as her eyes lift; her face stays within about 30° of the lens; no fast reframing.
ACTION / TIMELINE: 0–1.0 s her lips are closed; she stares at the cold tea and swallows; 1.0–3.6 s she speaks: her eyes stay on the tea for the first phrase and lift to him only on the second, with one slow breath out after the pause; 3.6–5.0 s her lips are closed, her eyes stay on him, and she blinks once, slowly.
PERFORMANCE — SALMA: She wants him to hear it without a fight; the obstacle is her own pride; her tactic is to say it quietly and wait. Her gaze stays engaged, her pupils move, her blinks are uneven. A tightness in the jaw after the first phrase. Tired, never tearful. SALMA says it quietly, almost to herself:
"أنا مش زعلانة… أنا تعبت."
Only SALMA speaks; her husband, off screen, is silent through this beat: the spoon in his cup goes still at 3.8 s. Only words between quotation marks are said aloud; everything else is direction.
TEXTURE: Pilling on the grey cardigan; a water ring on the low table beside the cold tea; the room falling into darkness past 2 m.
LIGHTING: One soft window key from camera left, 5600 K, 2:1; a warm 3200 K lamp low behind her; exposed for her face, the mouth evenly lit.
AUDIO: Her voice, accent, pace and pauses follow @audio1; close, dry dialogue; 4.0 s a spoon put down on a saucer, off screen, low. No music, no subtitles and no on-screen text.
STYLE: A quiet, naturalistic drama look: soft contrast, clean blacks, warm lamp against cool window, real skin texture; light seen only where it lands, as highlights and reflections on real surfaces.
POSITIVE LOCKS: Exactly one speaker and one person in frame. Her hands stay out of frame. No other readable text.
Check the take
Seedance re-times an audio reference, by about 20% on measured takes, so measure the draft's real onset and span against the take before judging it. The timing gate, working figures: the overall span within ±5% of the take's, every internal pause and the emphatic beat within ±3 frames, measured on the draft and again on the finish (word timestamps from a transcription tool, or markers in Premiere).
Common failure and fix: the gate fails because Seedance re-timed the line. Change one thing and retry within the shot's budget; if it still fails, the shot goes back to the default sound-off plate.