窓際スーパーつらい — 地下労働篇

プロンプト

Wan 3.0 v4 — 前半・後半の実送信プロンプト · revision 4

## Part 1 — 15秒

Create one 15-second video, part 1 of a 45-second commercial.
Japanese live-action psychological suspense film homage, grounded underground-labor oppression and intimate dorm temptation, 16:9. No environment references: invent locations from this text. Low stained concrete ceilings, damp walls, gray-brown muted palette, greenish practical fluorescent light and bare bulbs, restrained contrast, readable shadows. Eye-level 35-50mm perspective and patient closeups, modest camera movement. No glossy heroic rim light, teal-orange spectacle or fantasy castle.
Direction includes ACTING, EDITING AND SOUND, not only color. Perform as a quiet Japanese dramatic scene: conversational volume, listeners visibly processing what was said, slight hesitation before answers, relaxed natural phrasing with clear vowels/consonants. Do not rush lines to hit shot boundaries or use ad-announcer cadence in character dialogue. Pauses marked by punctuation are short meaningful breaths, not drawn-out syllables. No shouted comic delivery, exaggerated gasps or laughter track. Keep voices intelligible above location sound. No background intelligible speech, no added lines.
Character sheets define identity only, not composition. The new Yametaro sheet defines continuous peach skin around cheeks/ears/jaw/nape, natural scalp hair boundary, no black drawn contour or dark plate behind face. Preserve EXACT sheet proportions and materials; neither make him cuter nor more human. Preserve black hair and actual black glasses. Sobaya masks remain correctly worn including while drinking, no lifting, removal, transparency or deformation.
Yametaro speaking animation: Image 1 lower row REST/A/I/U/E/O supplies articulated mouth shapes, not fixed expressions. During EVERY Yametaro spoken line continuously open and close the mouth and move lower jaw/cheeks with the generated Japanese phonemes. Show a warm dark mouth cavity during open vowels, change width and roundness naturally, close briefly for consonants and pauses. Never preserve a painted closed smile while speech is audible. Keep face proportions and identity stable; no exaggerated screaming. The mouth shapes are reference poses to interpolate, not panels to reproduce.
Speech quoted below is the sole spoken script; kana spellings are pronunciation canon, never substitute kanji guesses. Generate speech and visible mouth motion together. Finish each line before cutting away, mouths rest during listening. Reference audio is voice timbre only, not source dialogue. No subtitles, title overlays, labels, panels, unintended people or phantom props.

References (numbering applies to THIS part only):
Image 1: character_yametaro_speech_v4_sheet.png — Yametaro, adult stylized character in purple leaf-pattern shirt, black trousers, round glasses; all design and body proportions from sheet.
Image 2: character_sobaya_basic_sheet.png — muscular gray-skinned Sobaya with white/red mask, dirty white T-shirt, black pants and white shoes. Three distinct copies intentionally appear in labor.
Image 3: prop_yume_points_review_v1.png — wall poster design: daily work 200 points, release 100,000, beer 200. Physical poster only.
Image 4: product_super_try.png — silver 350ml beer can used ONLY in final dorm shot, not labor or payout.
Audio 1: voice_ref_yametaro.wav — Yametaro voice timbre only, adult low-mid register with natural Kansai intonation. No other speaking character in this part. No Yotan present.

Shot 1 [0-7 s]: Tight three-quarter SIDE tracking closeup on Yametaro's face, shoulders and both hands pressing the SIDE of a horizontal radial wooden bar. A heavy stone/iron capstan around a vertical axle is in background. Exactly FOUR separate radial bars with exactly ONE worker assigned to each: Yametaro on bar A, one Sobaya on bar B, one Sobaya on bar C, one Sobaya on bar D. Each Sobaya is strong enough to operate his own bar alone. No two workers share, queue at or touch the same bar; each keeps his own bar throughout. All workers WALK FORWARD in one rotational direction, force tangent to the circle, never backward or toward the central axle. Stable hand/shoulder contact, grounded push-off steps and clear outer lane. Camera travels alongside Yametaro, no backward-tracking illusion or wide machine demonstration. At 0.7-5.8 Yametaro says 「本番も……バックアップも……消すんやなかった……！」. Read 本番 as ほんばん. Tired regret under breath, small breaths from physical effort, every word clear, no cheerful grin; wait on his exhausted reaction until bell. Continuous wood creak, low gear grind, shoes on gritty concrete, ventilation rumble. No music.

Shot 2 [7-12 s]: Bell rings once, gears slow to a halt. Hard cut to a narrow payout window in the adjacent concrete corridor. Camera remains on Yametaro's face and purple-shirt upper torso, with poster beside him. One offscreen clerk's hand places a single 200-point voucher in Yametaro's small hand; clerk stays entirely offscreen. Yametaro is the ONLY visible person and ONLY speaker. He receives, looks at the 100,000 release line, then grips his voucher. At 7.7-11.5 says 「じゅうまんポイント……ためて、絶対ここ出たる……」. Pronunciation ためて = ta-me-te, three clear morae; 絶対 = ぜったい; do not pronounce とめて. Resolve spoken quietly to himself after one small breath, not a triumphant speech. Paper rustle and lingering distant motor hum; no score.

Shot 3 [12-15 s]: Hard cut into cramped dorm, no foreground conversation yet. Worn tatami, thin futons along both walls, low wooden table, small refrigerator and snack shelf at back, poster on rear wall, cold overhead fluorescent tube and weak amber doorway. Two Sobayas sit along left wall. Both have beer mugs and several empty cans, snack packets, open point envelopes with almost nothing left. They spend their savings on beer and ignore release points. One finishes a drink through existing mask mouth opening, the other opens another can; no spoken lines. Yametaro sits at the front table holding his only 200-point voucher and glances at their spending, then looks down at it. No Yotan, no seller visible, no other people. Fluorescent hum, refrigerator compressor, soft ventilation, cloth against futon, occasional quiet glass contact and can-tab hiss. Realistic dry sound with short room reflections, not noisy party ambience. Hold last half-second with room tone, ready for hard cut to approaching foreman in part 2.



## Part 2 — 30秒

Create one 30-second video, part 2 of a 45-second commercial. Start inside dorm, no labor recap.
Japanese live-action psychological suspense film homage, grounded underground-labor oppression and intimate dorm temptation, 16:9. No environment references: invent locations from this text. Low stained concrete ceilings, damp walls, gray-brown muted palette, greenish practical fluorescent light and bare bulbs, restrained contrast, readable shadows. Eye-level 35-50mm perspective and patient closeups, modest camera movement. No glossy heroic rim light, teal-orange spectacle or fantasy castle.
Direction includes ACTING, EDITING AND SOUND, not only color. Perform as a quiet Japanese dramatic scene: conversational volume, listeners visibly processing what was said, slight hesitation before answers, relaxed natural phrasing with clear vowels/consonants. Do not rush lines to hit shot boundaries or use ad-announcer cadence in character dialogue. Pauses marked by punctuation are short meaningful breaths, not drawn-out syllables. No shouted comic delivery, exaggerated gasps or laughter track. Keep voices intelligible above location sound. No background intelligible speech, no added lines.
Character sheets define identity only, not composition. The new Yametaro sheet defines continuous peach skin around cheeks/ears/jaw/nape, natural scalp hair boundary, no black drawn contour or dark plate behind face. Preserve EXACT sheet proportions and materials; neither make him cuter nor more human. Preserve black hair and actual black glasses. Sobaya masks remain correctly worn including while drinking, no lifting, removal, transparency or deformation.
Yametaro speaking animation: Image 1 lower row REST/A/I/U/E/O supplies articulated mouth shapes, not fixed expressions. During EVERY Yametaro spoken line continuously open and close the mouth and move lower jaw/cheeks with the generated Japanese phonemes. Show a warm dark mouth cavity during open vowels, change width and roundness naturally, close briefly for consonants and pauses. Never preserve a painted closed smile while speech is audible. Keep face proportions and identity stable; no exaggerated screaming. The mouth shapes are reference poses to interpolate, not panels to reproduce.
Speech quoted below is the sole spoken script; kana spellings are pronunciation canon, never substitute kanji guesses. Generate speech and visible mouth motion together. Finish each line before cutting away, mouths rest during listening. Reference audio is voice timbre only, not source dialogue. No subtitles, title overlays, labels, panels, unintended people or phantom props.

References (numbering applies to THIS part only):
Image 1: character_yametaro_speech_v4_sheet.png — Yametaro identity, exact new outline-free proportions and surface treatment, adult voice and role.
Image 2: character_yotan_basic_sheet.png — Yotan's face from closeups, blond hair, glasses. Beige work jacket and trousers in this story. No headless panel figures.
Image 3: character_sobaya_basic_sheet.png — exactly TWO background Sobayas, muscular gray skin, white/red masks, white T-shirts, black trousers.
Image 4: prop_yume_points_review_v1.png — physical reward poster, correct design and values; not a shot reference.
Image 5: product_super_try.png — physical silver 350ml can, exact black/red design; not scene or camera reference.
Audio 1 voice_ref_yametaro.wav: Yametaro canonical voice timbre, natural adult Kansai inflection.
Audio 2 voice_ref_yotan.wav: Yotan canonical voice timbre, friendly conversational adult voice. Speak clearly without changing identity.

Dorm continuity: cramped concrete room, low stained ceiling, worn tatami, futons along both walls, low wooden table, small refrigerator and snack shelf at rear, reward poster on rear wall, overhead fluorescent tube and weak amber doorway. Yametaro sits left foreground at table clutching one 200-point voucher. Yotan crouches right beside him. Two Sobayas along left/rear wall drink and open more beer, empty cans and snacks accumulating next to depleted point envelopes. They ignore the savings goal. No third Sobaya, extra clerk or background Yametaro.

Acting: Yotan is a personable foreman using ordinary kindness as temptation, never an obvious villain. Soft volume, small smile, casual offer, watches Yametaro's response. Yametaro is tired and guarded, glances from voucher to cold can, replies after a brief hesitation. After drinking, relief escapes involuntarily before he can compose himself; the famous line is an intimate astonished admission, not a meme scream. Do not distort his head or face for emotion.
Sound: no score until the final product shot. Maintain low refrigerator compressor, ventilation fan, faint fluorescent buzz, soft futon/clothing rustle, occasional offscreen can tab and mug set-down. Background sounds stay well below dialogue. Let room tone occupy pauses, do not fill every gap with music or silence. Subtly reduce background level for the first sip and relief but preserve believable room presence. Can hiss, sip and swallow come from visible action, not added narration.

Shot 1 [0-4 s]: Seated-eye-level two-shot then gentle attention to Yotan. Yotan shows one unopened cold can. After 0.3 seconds of eye contact says 0.3-3.7 「やめたろうくん。今日の働き、見てたよ。ほら」. 今日=きょう, 働き=はたらき. Quiet conversational kindness. Can ownership remains Yotan's; no simultaneous purchase action.

Shot 2 [4-8.5 s]: Closeup Yametaro, look at can, pause, tighten fingers around voucher. At 4.4-8.1 says 「いや……ポイントは使わん。ここ出るために、ためるんや」. Pronunciation: ためるんや = ta-me-ru-n-ya, no leading ha/za syllable; begin ためるんや cleanly after a short breath. Read 使わん as つかわん. Disciplined resolve with temptation in eye movements. Mouth remains visible.

Shot 3 [8.5-12.5 s]: Yotan chest-up, places his unopened FREE can gently on table near Yametaro, then leaves hand relaxed. At 8.8-12.1 says 「これは俺のおごり。頑張った初日くらい、いいだろ？」. 俺=おれ, 頑張った=がんばった, 初日=しょにち. Warm, low-pressure invitation, slight softening at end. Small pause after おごり, no sinister laugh.

Shot 4 [12.5-14.5 s]: Yametaro hesitates, looks down, at 13.0-14.1 says 「……いっぽんだけやで」. Resigned permission to himself. Place voucher safely on his own side of table and take the free can.

Shot 5 [14.5-20.5 s]: SIX FULL SECONDS of silent physical drinking. Camera close enough to see can and lips. First open the tab with a crisp hiss (14.5-15.5). Then lift can (15.5-16.2). Clearly place rim against lips and TILT CAN to drink one visible sustained sip (16.2-18.0). Can actually touches lips; throat gives a small natural swallowing motion. Then separate can from lips and lower fully to chest (18.0-19.0). Finish swallowing with mouth closed, exhale softly, hold silent relief (19.0-20.5). No words during any of this. The sip is mandatory and must NOT be skipped, implied offscreen, substituted by opening the tab, or moved under the following dialogue. Keep all six seconds; do not spend them repeating tab opening. Character design remains unchanged.

Shot 6 [20.5-24 s]: Hard cut AFTER drinking to Yametaro face closeup, can ALREADY below chin at chest. After a brief look of disbelief says 20.8-23.7 「キンキンにひえてやがる……！ うますぎるやろ……！」. ひえて = hi-e-te. Natural adult voice breaking into genuine relief, first phrase breathy then warmer, not announcer or shout. Can remains down for entire line, mouth unobstructed and synchronized; no second sip.

Shot 7 [24-27 s]: Yotan slight smile, second unopened can visible. At 24.2-26.0 says 「にほんめからは、ばいてんでね」. Pronunciation ばいてん = ba-i-te-n, four morae with distinct い, no extra ん after ば. Casual friendly matter-of-fact delivery, short knowing pause after line. Yametaro silently slides his 200 voucher to Yotan, who takes it and places second can near him by 26.8. Preserve clear ownership, no extra hands or teleporting objects.

Shot 8 [27-30 s]: Hard cut to polished beer advertisement. Silver can LEFT on ice, large golden beer with foam behind, dark navy empty space RIGHT for postproduction text. Exact Image 5 can design, no generated titles. This is the FIRST bright glamour lighting and music: a brief original refreshing beer-commercial sting. Deep resonant professional male narrator distinct from both characters says at 27.3-29.3 「まどぎわ、スーパーツライ。」. Clear Japanese advertising cadence, musical tail ends cleanly. No extra voice.