▮▮▮ Seedance 2.5 · Scene 2 · Coverage
**Style:** Digital cinema capture, shot on an ARRI Alexa 35, Zeiss Supreme Prime, clean with smooth falloff, normal lenses — 35mm, 40mm, 50mm — the human eye, T2.8, shallow but controlled, no filtration, clean and modern. PERIOD — this film is set in Present day: keep EVERYTHING in frame period-accurate to that time (wardrobe, hair, vehicles, technology, phones/screens, signage, packaging, logos, décor); do NOT show anything from a later era — no anachronistic devices, cars, clothing or branding that would not exist then.
**Lighting:** Motivated lighting — every source justified in frame, natural contrast.
Time and weather, the same in every shot: DAY, overcast, dark storm clouds threatening rain. Seen OUTSIDE through the windows — keep it consistent and matching the time & weather: @Image 4 — Rural State Road — An empty rural road seen as the car tears down it. The world seen through the windows is the plate @Image 4 (attach it as a reference).
Lights are independent: turning one light source on or off (a headlamp, phone screen, lamp, dome/cabin light, headlights) changes ONLY that one light; it NEVER dims, brightens, switches or tints any other light, the ambient/cabin lighting, or the overall exposure. Every other light stays exactly as it was.
Light holds still: a face brightening as it lifts into the key, a jawline sharpening, a shadow sliding across a shoulder as someone turns - those are RIGHT, and they come from the SUBJECT moving through light that has not changed. What never happens is the light following the performance: it does not swell on a line, dip in a pause, warm on a close moment, or re-key to favour whoever is speaking.
**Color:** Clean digital capture, fine grain, natural filmic grade, true colour, natural saturation, filmic blacks with shadow detail. Palette #1a1a1a #2b3a45 #8a6a4a #d9c9a8.
**Camera (STRICT):** CUT 1 — Wide, eye level, 35mm. CUT 2 — Medium close-up, eye level, 50mm. CUT 3 — Close-up, low angle, 85mm. CUT 4 — Medium close-up, eye level, 50mm. CUT 5 — Close-up, high angle, 65mm. CUT 6 — Close-up, low angle, 50mm. CUT 7 — Wide, low angle, 35mm. CUT 8 — Wide, eye level, 35mm. Each cut is its own setup, cut together in order at the pace of the action - never a move from one setup into the next. Lens range: 35–50mm for the wide setups, longer for the closer cuts. Movement within each cut: slow deliberate dolly moves. Hold the line: screen direction and eyelines match across the cuts, and the 180-degree line holds.
**Skin:** Faces are not lit separately: skin carries only the light already in the room and its own soft shadows — no added face fill, no glow, no sourceless rim or kicker from behind or below, and a face is NEVER lit brighter than the space around it.
Living faces: they blink naturally, the eyes are wet and alive with real catch-lights, breath moves the chest — behaviour, not displayed emotion. Never waxy, glassy, frozen or dead-eyed, and never an unblinking stare; small and truthful, never mugging.
The involuntary layer: under the words and through every silence, a real face carries constant low-level activity that is NOT an expression and NOT a reaction — the eyes flicker in tiny shifts while they hold on someone, blinking is irregular (a quick one, a long gap, a slow heavy one, never metronomic), the lower lids sit fractionally tense, breath is uneven and visibly moves the chest and shoulders, the nostrils widen on an intake, the mouth is dry-pressed or left slightly parted between lines, and the throat swallows. Stillness here means RESTRAINED, never motionless: a body in a chair keeps correcting its balance, weight shifts once, a shoulder lifts and drops, and the head drifts by fractions of a centimetre with the breathing. A person who is listening, waiting or saying nothing is doing all of this too — nobody in frame is ever a held pose.
**Acting:** Perform to a naturalistic, restrained standard — professional screen acting, NO overacting or exaggeration; play the subtext and react small. No mugging, no theatrical double-takes, no telegraphed or announced emotion, no cartoon faces, no over-gesturing, no eyebrow "reaction" acting, no everyone-reacts-at-once. Nothing louder than life.
@Image 2 (Julian) — balance a rigid, high-society preppiness with a transformation into a wild-eyed, joyful romantic rebel.
**Physics:** Solid objects: every prop is opaque and has real weight — paper, glass, plastic, cloth and banknotes are never transparent, translucent or see-through, and never ghost or double.
**Composition:** Classical coverage — master, singles, over-the-shoulders.
**Continuity:** No look-alike copying: characters in similar clothing (e.g. two camouflage coats, two uniforms) each keep their OWN distinct items — a tag, label, badge, patch, accessory, light or piece of gear on one person is NEVER copied, mirrored or duplicated onto another just because their outfits look alike.
Placement & continuity: every person and object keeps the SAME position, seat, pose and placement it already has, and every prop stays WITH its owner and WHERE it was last seen. Nothing moves to a new person or place, and nothing appears, disappears, or changes — its state (on/off, open/closed), what someone holds, wardrobe or colour — UNLESS the described action explicitly does so. Everything the action does not name stays EXACTLY as it was in the moment before.
Hold constant: time of day, sky, weather, light direction/intensity, colour grade and exposure; every face, hair, wardrobe and prop; the setting and the world seen outside any windows — all stay IDENTICAL from the first frame to the last. Nothing brightens, darkens, warms, cools or shifts; no item appears, disappears or changes mid-shot UNLESS the described action explicitly does so.
**Wardrobe:** @Image 2 — A suffocating, bespoke black Tom Ford tuxedo with satin lapels, a crisp white pleated tuxedo shirt, a perfectly tied black silk bowtie, and highly polished black patent leather dress shoes; sweating, highly tense, and visibly anxious.
**Technical:** Aspect ratio 21:9. About 15 seconds, 8 cuts. Frame rate: 24 fps, natural motion blur.
**Audio:** the dialogue in English, word for word as written. Julian speaks in the voice of @Audio 1. Under it, wind howling softly outside the car chassis, low distant rumble of thunder, with the sounds of the action as written. No music. No subtitles.
**People count (STRICT):** CUT 2 — @Image 2 (one person). CUT 3 — @Image 2 (one person). CUT 5 — @Image 2 (one person).
Never clone, mirror, duplicate or multiply anyone.
Cast count: the named cast above, PLUS any background figures the blocks above explicitly describe, and NOBODY ELSE — no extra people, no invented crowd, no duplicates of anyone.
**Object rules (STRICT):** @Image 5 (Ignition Keys): one of it, never duplicated.
@julians-phone (Julian's Phone): one of it, never duplicated.
@microfiber-cloth (Microfiber Cloth): one of it, never duplicated.
@Image 6 (1965 Ford Mustang): one of it, never duplicated.
One-owner rule: every item of wardrobe, gear, grooming and every prop belongs to ONE character only — it never appears on, transfers to, or is duplicated on anyone else. If an item is not listed for a character above, that character does NOT have it.
No invented writing: no readable lettering, signage, labels, logos or numbers anywhere unless the action or the blocks above state the exact words; where they do - a named sign, a screen, a message, a card - reproduce them verbatim and invent nothing, with no gloss, sheen, glow or white edging on the strokes.
**Reference-use rule (STRICT):** From @Image 2:pristine-groom take the character's face, hair, body and wardrobe. From @Image 5, @julians-phone, @microfiber-cloth and @Image 6 take each one's design.
Locked to reference: any attached reference images are the exact and only source for each character's identity, face and wardrobe and each asset's design — reproduce them precisely; invent nothing, substitute nothing, add no new person or object.
Full reference look: give each character their COMPLETE look from their reference image — face, hair, body, skin tone and physical CONDITION/state (injuries, scars, paleness, bruises, dirt, sweat, grooming) AND every worn item: hat/cap/headgear, glasses, jacket, gloves, badge, patch. Do NOT drop, simplify, clean up, heal or bare-head a character just because the action text doesn't mention it, and hold it even when they move or lean into a new pose — the reference is exactly how that character looks and is dressed.
**Assets:**
@Image 2:pristine-groom — CHARACTER, Julian, exactly as shown in the reference.
@Image 5 — PROP, Ignition Keys, exactly as shown in the reference.
@julians-phone — PROP, Julian's Phone, exactly as shown in the reference.
@microfiber-cloth — PROP, Microfiber Cloth, exactly as shown in the reference.
@Image 6 — VEHICLE, 1965 Ford Mustang, exactly as shown in the reference.
**Avoid:** Restaging, relighting, grade shifts between cuts, invented people or objects, new dialogue or action, on-screen text, captions, watermarks, subtitles, and any brand mark or lettering NOT named above — signage and logos the blocks above DO describe belong in frame and are reproduced exactly as written there. Nothing that is not already in the master + coverage.
The look avoids: heavy grain, diffusion glow, bloom, skin smoothing.
Never: identity morphing or face drift, swapping who is who, extra or duplicate people, malformed hands or fingers, floating or teleporting objects, relighting or grade shifts mid-shot, wardrobe or prop changes, captions, subtitles, watermarks or any text laid over the picture, and no logo the blocks above do not name.
**Action:**
As the take opens, already true — do not re-stage it, just honour it: @Image 2 sweating, highly tense, and visibly anxious.
OPENING FRAME: the clip OPENS on @Image 1 — reproduce that frame EXACTLY as the first frame of the clip (same framing, lens, staging, wardrobe, light and grade), then play the action forward from it. It is the starting instant, not a separate shot — the SAME instant the action starts from: anyone or anything the action names at its start that is not visible in that frame is out of view at that instant, and comes into the frame only as the action describes.
0:00–0:03 — CUT 1 — Wide, eye level, 35mm. Inside the pristine, cherry-red cabin of the @Image 6, the atmosphere is thick with a suffocating tension. The flat, cold daylight of a late autumn afternoon filters through the clean safety glass windows, casting a muted glow over the deep-red textured vinyl of the symmetrical dual-cowl dashboard. Outside, the bleak, dead-yellow fields of the Rural State Road stretch out under a heavy, steel-grey sky, where dark storm clouds hang low, threatening an imminent downpour. CUT.
0:03–0:06 — CUT 2 — Medium close-up, eye level, 50mm. @Image 2 sits gripped by panic in the driver's seat, his athletic frame stiff and tense inside his pristine, bespoke black Tom Ford tuxedo. His face is pale, a bead of sweat tracing his sharp jawline under his neatly styled dark brown hair. He stares straight ahead through the windshield at the empty asphalt road, his blue eyes wide with high-strung anxiety as the weight of his mother's demands presses down on him. CUT.
0:06–0:07 — CUT 3 — Close-up, low angle, 85mm. With a trembling, pale hand, @Image 2 reaches forward and grips the heavy metal head of the @Image 5 already resting in the ignition slot. He twists his wrist with a sharp, desperate force. CUT.
0:07–0:10 — CUT 4 — Medium close-up, eye level, 50mm. The vintage starter motor grinds for a fraction of a second before the massive V8 engine roars to life with a deafening, throaty rumble that vibrates through the floorboards and shakes the entire frame of the car. The five-pod instrument cluster on the dashboard instantly springs to life, its warm amber backlighting illuminating the dials. CUT.
0:10–0:11 — CUT 5 — Close-up, high angle, 65mm. Without a moment's hesitation, @Image 2 drops his hand to the polished chrome T-bar shifter on the floor console. He yanks the lever backward, slotting it firmly into drive with a heavy, mechanical click. CUT.
0:11–0:12 — CUT 6 — Close-up, low angle, 50mm. His highly polished black patent leather dress shoe stomps hard onto the accelerator pedal. CUT.
0:12–0:14 — CUT 7 — Wide, low angle, 35mm. The rear tires of the cherry-red @Image 6 screech violently against the coarse gravel shoulder, kicking up a small cloud of grey dust and dry dirt. The classic muscle car surges forward with aggressive power, nose lifting slightly as it tears down the empty, undivided rural highway, accelerating rapidly into the desolate landscape under the gathering storm. CUT.
0:14–0:15 — CUT 8 — Wide, eye level, 35mm. An exterior wide shot holding on the empty stretch of @Image 4 as the dust settles, with the @Image 6 disappearing into a speck on the horizon under the dark, looming storm clouds. CUT.
CLOSING BEAT — do NOT end the take on the last spoken word. After the final line of dialogue, hold about one to two seconds of reaction and stillness before the clip ends: the characters settle, a breath, a held eyeline or a small closing gesture, the moment coming to rest. This gives the cut a tail to edit on, so the scene does not jump-cut the instant the last word is spoken. It is a HELD ending, not a new action — nobody leaves, no new event begins, and the camera does not cut, re-angle, or push in.
ALSO TRUE IN THIS PART — from the Bible and the scene; keep each exactly:
- Time and weather, the same in every shot: DAY, overcast, dark storm clouds threatening rain. Seen OUTSIDE through the windows — keep it consistent and matching the time & weather: @rural-state-road — Rural State Road — An empty rural road seen as the car tears down it. The world seen through the windows is the plate @rural-state-road (attach it as a reference).
6 pictures