窓際スーパーつらい — そば屋&たこさん初めての東京篇

Wan 3.0 — 採用前半20秒＋後半20秒の実送信本文

Create one 20-second 16:9 Hollywood-scale live-action alien invasion film segment with synchronized predominantly English dialogue. Okayaman’s catchphrase and the final product announcer remain Japanese. Photoreal humans and city, premium cinematic alien-machine VFX. Part of a 40-second beer commercial; earnest danger, no comic music until the final beer pour.
References for THIS segment only:
Image 1: Fukuchan face and clothing, lean 170cm. Image 2: Yotan face, hair and clothing, lean 170cm, no guitar. Use face closeups for identity; never reproduce sheet panels as people.
Image 3: Okayaman episode-only physical appearance, 180cm slim, charcoal suit and white open collar, exact face and beard. Exactly one of each human, no other named characters.
Image 4: GRAY CLAY PREVIS ONLY for relative scale, approximate spacing and three-machine count in Shot 2. NOT final appearance, NOT exact camera or architecture, NOT a first frame. Rebuild a cinematic photoreal rainy neon Akasaka using the world description, NEVER render gray clay, white models, daytime or simplified toy materials. Keep broad near/middle/far distribution with several blocks between machines; camera and detailed street arrangement may vary. Alien machine silhouette: broad curved crescent hull, three long articulated metal legs, dark graphite plating, tiny dark cockpit window, armored belly reservoir with narrow inspection slit, flexible extraction hoses. Do not copy the simplified faceted hull as finished design. Heads above nearby ten-story roofs; cars smaller than machine feet. About 50m nominal size; building comparison controls visible scale.
Image 5: Takosan PILOT identity only in final cockpit shot: small black hooded character, ivory face, black round eyes, two white fingerless round hands on controls, six organic lower tentacles curled around seat. No speech. Organic tentacles never become machine legs or extraction pipes. Cockpit internal shot fits this small pilot, not a huge pilot visible outside.
Image 6: standalone intact beer server PROP ONLY. Real commercial brushed stainless cabinet, TWO black tap handles, perforated drip tray, realistic applied silver brand STICKERS on side and front using the can label design: black double border, black italic Madogiwa, red SUPER “TRY”, black 生, red 窓際スーパーつらい. Stickers are separate adhesive badges, not oversized words printed directly on steel. Preserve appliance shape and stickers; place naturally in the izakaya with scene lighting. Do not copy plain studio background, product camera or lighting. Server begins intact, taps closed, NO hose or puncture. The plug action is defined by Shot 4, not by this image. Keep broad exposed steel below/left of front sticker as puncture target, leave label intact.
No Sobaya in this segment, no white-mask humans, no beer cans, no final product card. Logo on the SERVER is intentional. Exactly three distant machines but only nearest machine's single extraction hose enters this izakaya.
World: deliberately exaggerated Hollywood TOKYO, Akasaka nightlife. Wet neon street, red lantern izakaya, dense vertical Japanese signs and overhead wires, steam, monumental close-looking red Tokyo Tower behind buildings, temple-style roof silhouette among skyscrapers. No London landmarks. Japanese street signs may read 居酒屋, 赤坂, 焼鳥, 生ビール. Rainy night, red/cyan reflections, distant searchlights, fleeing extras and stopped taxis. Not a geographically literal street survey. Preserve Tokyo Tower as recognizable three-dimensional distant landmark, not a sticker. Enormous mothership above clouds. Do not render subtitles: exact Japanese subtitles will be added in post. Keep the bottom 12 percent free of mouths, hands and critical contact details. No floating explanatory text. No green screens.
Audio and performance: Generate only specified lines, each exactly once; no introductions or added catchphrases. Voice references provide timbre only, never their original words, language or timing. Preserve each speaker’s recognizable timbre while delivering the specified English in natural Hollywood-film cadence. Do not translate English dialogue into Japanese. Okayaman alone speaks his specified Japanese catchphrase; nobody reacts to the language switch. All speech, including English, is generated by Wan. Generate visible human lips, jaw and cheeks in sync with speech, keep face readable and hold through final phoneme. Other people do not mouth the line. Takosan silent. Dialogue intelligible above effects. Scale comes from layered occlusion, atmospheric depth and heavy acceleration, not fast camera spins or constant explosions. City and machine use cinematic full-scale materials and lighting, never a toy-city or miniature diorama; only the small pilot retains his reference character design. Original epic orchestral dread, low alien mechanical resonance, rain and distant panic; no quoted copyrighted score. No humorous sting at Okayaman's line.

Audio 1 Fukuchan voice-timbre only. Audio 2 Yotan voice-timbre only. Audio 3 Okayaman voice-timbre only.
Shot 1 [0-4 s] Inside warm Akasaka izakaya: Fukuchan, Yotan, Okayaman at one table beside street window and entrance. Begin macro on a partly drunk beer mug RESTING motionless on the table. Mug already contains golden beer at a fixed level, no liquid stream above it, no bottle, pitcher or faucet overhead, no hands touching it. Initially liquid is still; low ground rumble then creates fine ripples that grow without changing liquid level. Nothing is poured into or removed from this mug. Hanging light sways and crockery rattles. Cut within shot to Fukuchan chest-up at 0.9s, he looks up and says 1.0-2.5s “Was that an earthquake?”. Yotan visible chest-up turns toward window, says 2.6-3.9s “Guys… look outside.”. No creature or beer theft explained yet.
Shot 2 [4-8 s] Hard cut to a sustained elevated wide panorama over Akasaka, from a rooftop overlooking the neon streets, not a fast drone orbit. Exactly THREE identical 50m tripods visible SIMULTANEOUSLY throughout this shot, widely separated across several city blocks, never in a row. Nearest machine at left midground looms above izakaya roofs, plants one mechanical foot with weight and delayed dust/water displacement. Second machine farther back across a broad avenue faces another direction, lowering a collection hose behind buildings. Third is distant toward red Tokyo Tower, partially veiled in rain and smoke but its crescent hull remains readable. Buildings and atmospheric perspective separate all three; each has three coherent mechanical legs, partly occluded by architecture. Tokyo Tower remains physically larger and farther away. Foreground rooftop edge, fleeing people and taxis on streets below establish scale; lightning briefly silhouettes the fleet. Hold the composition long enough to discover all three. No closeup beer contents, no visible pilots yet. Deep alien horn reverberates across separate districts, distant impacts and epic dread. No named dialogue.
Shot 3 [8-11.2 s] Hard cut doorway medium chest-up: Okayaman has approached threshold, sees towering machine outside, recoils backward toward the interior. Face and mouth readable despite frightened breathing; shouts exactly 「大変驚いております！」 (たいへん おどろいて おります) during 8.3-10.9s. Polite wording delivered in genuine terror, not a news report. Other two remain behind him inside. He does not run away down the street.
Shot 4 [11.2-16.5 s] Same interior, branded rectangular brushed stainless beer server secured on counter. Front has the silver Madogiwa / SUPER “TRY” / 生 / 窓際スーパーつらい sticker matching Image 6, with visible sticker border. Initially undamaged front metal, tap CLOSED, no attached hose. Mechanical hose approaches window, Fukuchan chest-up warns 11.3-12.1s “Get down!”, all three duck. At 12.5s probe breaks window ABOVE heads. Closeup at 13s: central pointed HOLLOW PLUG advances perpendicular to flat steel front BELOW/LEFT of the front logo sticker, not into tap or existing fitting. At 13.5s steel dents inward, plug penetrates and tip disappears INSIDE server. Outer sealing collar presses flush around puncture and locking claws grip damaged metal at 14s. Contact point then stays FIXED through all remaining Earth shots; hose may flex farther away but plug never withdraws or slides. Only AFTER mechanical lock at 14.5s pump starts, golden beer flows from server OUT through enclosed transparent tube AWAY from server toward machine, never into server. Keep contact readable for final two seconds; logo stays intact above connection. No spurting liquid, no free jet, no collecting from mug, no flesh injury. Image 6 defines only the intact appliance; generate the puncture, sealing collar and locking claws from these action instructions. Do not introduce a hose before its arrival.
Shot 5 [16.5-20 s] Sealed plug remains embedded, closed normal tap untouched; transparent hose carries beer away from server; Yotan raises face slightly from cover, says 17.0-18.0s “…Beer?”. At 18.2s cut to close cockpit: normal small Takosan calmly operates controls with round white hands, six tentacles around seat. Exterior rain and moving reflected city light. A small cockpit gauge rises to indicate filling of the remote armored belly reservoir; do not relocate the belly reservoir beside the pilot or turn it into a giant beer mug. End on harvesting machinery with continuing pump sound. No Sobaya on Earth. No commercial music yet.
Hard constraints: exact 20s, all five named lines complete, no extra spoken words, no beer shown flowing before puncture, people remain inside/at threshold, only the nearest machine attacks this izakaya; its pilot is revealed only after theft. Other two machines stay in separate districts. End before mother ship command; no product end card in this segment.
Create ONE 20-second 16:9 Hollywood-scale mothership sequence, second half of a 40-second beer commercial. Sobaya speaks English, final announcer Japanese. Generate synchronized speech, effects and music. No Earth street scenes or recap.
References, scoped by shot:
Image 1: Sobaya identity ONLY for Shots 1 and 4. Rigid white mask, red markings, black openings, neutral gray exposed skin, muscular body, white T-shirt, black trousers. Mask remains opaque, properly worn and absolutely rigid during speech and sip; no face reveal, lifting or lip animation on mask. Do not introduce his reference beer mug before Shot 4. One Sobaya only, never in hangar.
Image 2: Takosan identity ONLY for disembarked pilots in Shots 2 and 3. Black hood and robe, ivory face, black round eyes, two white round fingerless hands and six lower-body organic tentacles. Exactly three pilots, silent, small versus machines. No seated cockpit pilot and no human legs. No reference-panel duplication.
Image 3: GRAY CLAY hangar previs ONLY for adopted angular machine geometry, relative scale, three dock positions and ceiling pipe route. Adopt the broad shallow angular triangular/kite hull with raised edge, THREE articulated mechanical legs with round joints, and small cylindrical empty cockpit under hull, matching the first-half v3 design. NOT a curved crescent saucer. Legs partly bent and locked in cradles. Full beer reservoir is a separate enclosed compartment, NOT pilot compartment. Gray pipes in previs must become transparent inspection pipelines as specified below. Previs is NOT final material, lighting, exact framing or animation. Rebuild in cinematic full-scale graphite metal and dramatic cold industrial lighting; no gray clay, toy diorama, frozen pilots, incidental human scale figures or exact camera lock. Hangar is kilometer-scale; machines approximately 50m. Three normal-size Takosan pilots swim in open space, all cockpits empty.
Image 4: silver Super TRY can identity ONLY in Shot 5, with black Madogiwa wordmark, red SUPER “TRY” and Japanese brand label. Do not display a can in command or hangar shots. Do not copy reference background or sheet layout.
Audio 1: Sobaya canonical voice-TIMBRE only, not its Japanese words or timing. Generate the two specified English sentences in this recognizable low voice. Narrator is a separate deep authoritative Japanese advertising voice generated by Wan. No other dialogue, no extra catchphrase. Preserve source-generated audio for later editing.
Continuity: hangar units are OTHER returning machines, reservoirs already full, no need to match the Earth machine's fill state. Machines and equipment mechanically fixed. Hangar ZERO GRAVITY: only pilots drift freely; closed pipes carry beer under pump pressure. Command deck has local artificial gravity, beer pours down into mug normally. Earth curved horizon seen through command window, never rain inside ship. Scale from receding structure and occlusion, not fast camera spins.
Extraction readability: each machine has a thick sealed pipe rising from its separate reservoir to the ceiling manifold. Majority of each VISIBLE vertical run is transparent reinforced glass/polymer, metal rings limited to joins; transparent pipes reveal moving amber-gold beer with fine microbubbles and a distinct advancing liquid front. Initially transparent riser empty, pump then drives a clear golden front UPWARD FROM machine TO ceiling, followed by full flow. Keep this action in focus against dark structure, warm backlight through liquid. It is liquid, not glowing energy. First nearest line, then middle, then far line. No opaque pipe hiding the transfer, no flow down toward machine, no free jets or floating beer. Pressure storage behind a viewing window with moving separator/diaphragm gives a visible received-volume change without an impossible gravity waterline in zero-G. No need to empty whole fleet before serving.
Audio/graphics: original epic low orchestral tension and immense hangar resonance, sequential pump hum and liquid sounds under clear speech. At local 11s switch to uplifting beer-ad music. No copyrighted tune. No generated subtitles, Japanese captions added later; keep bottom 12 percent free of critical action. No green screens. Generate monitor schematics but not readable interface words. Exact lines each once, hold shot through last phoneme. Takosan stays silent.
All following times are LOCAL seconds (add 20 for complete film).
Shot 1 [0-4 s] Vast dark mothership command deck: multiple receding structural ribs and distant light rows establish immense depth behind Sobaya seated as superior commander. Earth’s blue curved horizon and white storm systems fill the distant panoramic window, no toy globe. Frame him and large tactical monitor, then chest-up for line 0.8-2.9s “Not a drop left behind.”, low and coldly commanding. Rigid mask unchanged. Monitor shows an unlabeled image of the tripod, city collection nodes, tank icon and progress bar reaching full. At 3-4s use nearly front-facing steady monitor insert, clear bottom area for later Japanese UI overlay, no generated readable UI text. No green screen. Monitor displays machine harvesting data, never a human portrait.
Shot 2 [4-7 s] Sustained wide shot of kilometer-scale ZERO-G mothership hangar, deep perspective with repeated structural ribs and distant lights. Exactly THREE returning tripods already parked in separate docking cradles receding diagonally into the distance, legs partly bent and mechanically clamped as in Image 3, each belly reservoir full of golden beer. Machine cockpits are EMPTY: pilots have ALREADY disembarked before this shot. Exactly THREE normal small Takosan pilots swim slowly between the giant machines in zero gravity, black robes drifting, six lower-body tentacles gently undulating; round white hands push from handrails, momentum carries them onward. Show one nearer pilot and two farther away, clearly separated, no synchronized motions. No jetpacks, no water, no bubbles. Aircraft-sized equipment dwarfs pilots; hangar dwarfs machines. All machine hardware remains fixed despite zero gravity. Ceiling extraction hoses are already securely connected to each reservoir, with long clear inspection sections initially empty. Show golden fronts climbing visibly through them when pumps engage. At 5s the nearest pump engages, at 6s the middle pump, at 6.8s the farthest pump; each starts golden flow and an indicator light in sequence, not all at once. Full draining is not completed in this short shot.
Shot 3 [7-10.5 s] Cut to nearest machine’s sealed extraction connection, beer visibly driven UP through enclosed transparent pipe by pump pressure. Hold connection and rising liquid front in the same readable close shot for at least 1.2 seconds, not an opaque pipe insert. A reservoir inspection gauge falls as fluid leaves. In depth behind it other docked machines continue pumping; empty cockpits stay empty. A small disembarked Takosan drifts past behind the pipe, not operating cockpit controls. Cut along the ceiling manifold into a huge sealed mother ship receiving tank, liquid confined inside piping and bladder-separated storage, no free floating beer or spills. Match cut follows the shared outlet pipe toward command-deck beer tap. Sequential pump hums layer into one industrial rhythm, orchestral tension resolves. The entire fleet’s reservoirs need not finish emptying before Sobaya can serve the collected beer.
Shot 4 [10.5-16 s] Sobaya alone at command-deck tap. Closeup empty glass mug held upright under nozzle, hand pulls lever, golden beer fills with creamy white foam, lever closes, then he lifts mug. At 11s switch from invasion orchestra into crisp uplifting premium beer-ad music. Macro liquid and condensation, warm gold against cold blue Earth. Brief sip at existing mask mouth opening, no lifting/removing mask and no glass passing through solid mask; lower mug fully before line. Finish the pour and brief sip by 13.0s; use concise cinematic inserts without extending the drink. At 13.1-15.9s he says “Now that’s a planet worth visiting.” in relaxed, satisfied English. Stay on masked head and upper body through sentence. Relaxed satisfaction, not evil laughter.
Shot 5 [16-20 s] Premium fixed product beauty shot, silver can from Image 4 and freshly poured glass on left, Earth through mothership window behind. Right 45 percent quiet dark negative space for exact later typography. Condensation, golden beer, clean foam. Separate deep Japanese male announcer 16.6-19s says exactly 「窓際、スーパーツライ。」 (まどぎわ、すーぱーつらい), crisp punchy advertising cadence. No other dialogue, no generated captions, no generated slogans. Finish music decisively at end.
Hard constraints: exactly 20s; three spoken lines only including narrator; Sobaya never on Earth; exactly three parked machines and three disembarked normal-size Takosan pilots in the hangar wide; pilots swim silently and do not drink. Machines already docked and hoses connected before pumps start; first collected beer reaches ship storage before serving; source-to-destination flow unambiguous. Mask maintained all frames. Commercial reveal only at pouring. No Earth cast, no old crescent machine design, no opaque clay pipes, no seated pilots.