窓際スーパーつらい — 爆走タンクローリー篇

プロンプト

Wan実送信本文：既存本編15秒＋福ちゃん追加10秒 · revision 2

# Adopted main film — 15 seconds, inserted at final 6–21 seconds

Create exactly one 15-second 16:9 spectacular desert-highway action commercial with Japanese speech and full cinematic sound. Escalate the scale and energy: roaring motorcycles, flying dust, a huge airborne boarding leap, a violent metal impact and an explosive fountain of golden beer. Fast, forceful, legible action rather than slow motion. Preserve Sobaya's canonical masked identity. An existing five-second product endcard is appended later; generate only this 15-second main film.

References (Image1 = Image 1, Image2 = Image 2, Audio1 = Audio 1):
Image 1: Sobaya character identity. Massive muscular body, short black hair, neutral gray skin, rigid opaque white mask with exact red markings and dark openings. White T-shirt, black pants, white shoes. One lead Sobaya and five identical biker followers, six riders total. All masks remain attached and unchanged throughout, even during shouting and drinking. No exposed face, shifting mask, flexible mask mouth, or animated mask lips. Ignore sheet layout, labels and studio background. Each rider has one empty beer mug in a secure hip holder at the start.
Image 2: the single silver beer tanker and its exact physical brand artwork. Preserve its cab, cylindrical tank, recognizable silver finish and label. Use brushed silver with broad diffuse highlights; avoid clear mirror images of bodies or hands on the tank. Ignore the studio background.
Audio 1: canonical Sobaya voice-timbre reference ONLY. Generate the new Japanese lines below, never copy reference words or timing. The leader is the only intelligible speaker; followers supply motorcycle noise, not additional dialogue. His opening cry is a full-throated excited bellow. There is NO final spoken line: the film ends while he drinks.

DIRECTION AND CAST LOCK: The truck's engine cab and windshield are at SCREEN RIGHT, the rear end of its tank at SCREEN LEFT. It drives FORWARD nose-first toward screen right in every moving shot. Motorcycles point and move in the SAME direction. Keep cameras on the truck's left flank, never cross the road axis or mirror the vehicle. Road markings and distant terrain sweep RIGHT TO LEFT relative to the tracking camera; tires roll forward; dust trails behind toward screen left. The cab leads the tank, never the reverse. No pedestrians anywhere at any time. Before the leap, every Sobaya is physically seated on a visible motorcycle; after the leap, the leader is attached to the truck and the five followers stay seated on their bikes. Never spawn a standing, walking or running Sobaya beside moving traffic. Remove the old walking queue entirely.

Shot 1 [0.0-2.7 s] — Low, dynamic LEFT-SIDE tracking wide shot. The truck's cab stays visibly at SCREEN RIGHT, pulling the long silver tank nose-first toward the RIGHT. Six black motorcycles surge up from the LEFT rear, stretched out in a moving formation, huge dust plumes trailing LEFT. Rapid forward travel is obvious from road markings streaking backward. The leader is nearest camera, unmistakably astride a moving motorcycle, both feet on pegs. From 0.8 to 2.2 seconds he bellows exactly once: "ビール！" Pronunciation: bii-ru! A huge excited chest-powered battle cry with a sustained vowel. Mask rigid. A brief closer tracking angle shows his face AND handlebars. No standing people or riders already attached to the truck.

Shot 2 [2.7-5.5 s] — Escalating high-speed CAR CHASE, two clean forceful cuts within the same screen direction. First a wheel-height shot alongside the tanker: its front cab is at right, nose-first forward motion to right, giant wheels pounding asphalt; motorcycle front wheels enter from left and rush alongside, riders lean forward under acceleration. Then a wider side tracking shot reveals the whole formation, front cab right and rear tank left. The leader accelerates from the rear axle toward the tank's middle, while the five other riders fan out and close in behind him. The truck makes one small evasive swerve without reversing its travel; bikers bank smoothly and maintain pursuit. Dust, road grit, vibrating suspension, cloth whipping in the wind, screaming engines and percussion. Keep each rider attached to a visible bike. No pedestrians, stationary clones or walking on the road. No camera-axis reversal, frontal view that hides travel direction, or reverse-driving shot.

Shot 3 [5.5-8.0 s] — ONE LEGIBLE SIDE-ON FULL-BODY BOARDING SHOT. The leader rises into a crouch on his still-moving bike, forcefully pushes off with BOTH legs and LEAPS diagonally up across a visible 1.5-meter gap toward the tanker. Show his entire muscular body suspended in open air with BOTH feet off the motorcycle, bent knees and two arms reaching toward the truck rail. Hold the actual airborne interval long enough to read it, at normal dramatic speed, not a prolonged slow-motion sequence. Both hands catch the side rail with a forceful body slam, then both feet find the lower step. The abandoned bike falls away BEHIND toward screen left without its rider. Five followers remain on bikes behind the truck. Cab remains right, forward travel remains right. Huge air rush, rail slam, wind and trailing dust. Do not replace the jump with a short step, running, teleporting or a rider already on the truck.

Shot 4 [8.0-10.0 s] — Same leader on the tanker step, left hand gripping the rail and both feet braced. He punches the silver panel once with his RIGHT fist. Metal buckles at contact; as the fist pulls out, a spectacular golden beer fountain bursts from ONE hole behind the label. Impact camera jolt, metal boom, bright beer spray. The truck brakes hard into a dust cloud, then show its wheels stopped. Five followers brake to a halt still astride their bikes. The leader remains on the step; no one is on foot. He takes his EMPTY mug from its hip holder with his right hand as the vehicle stops. The left hand remains attached to the rail. No extra arms and no duplicate mugs.

Shot 5 [10.0-12.0 s] — Stopped truck, dust settling. One continuous medium-close view of the leader catching the same beer jet in his right-hand mug. Frame from his right side with shoulder, upper arm, elbow, forearm and hand all visibly connected; left supporting hand visible farther away on the rail. Exactly TWO arms. Place the glass against sky and distant desert, away from mirror-like tank reflections. No helper, reflected human arm or detached hand enters. Keep the pouring stream and glass visible together as the golden level rises from empty to a thick white foam head. A refreshing commercial music hit replaces the action drums. Followers remain out of this close shot, seated on parked motorcycles elsewhere. The leader never leaves the truck step.

Shot 6 [12.0-15.0 s] — Heroic chest-up view of the SAME leader, still on the stopped tanker step, left hand on the rail. He raises the full mug in his RIGHT hand to his rigid mask's existing mouth opening and drinks in one deep triumphant sustained gulp. Glass covers the mouth opening, mask stays opaque and correctly attached; no exposed face or animated mask lips. Beer level visibly falls, throat and shoulders move with the swallow, foam and sunlight catch the rim. Tight framing with the distant desert and soft golden spray, no pedestrians. END THE FILM WHILE HE IS STILL DRINKING, mug still tilted at his mask. Do not lower the mug, look at camera, raise a toast, speak or add a reaction pose after the drink. No final slogan. Hard cut at exactly 15 seconds to the separately edited product endcard.

Sound: loud accelerating engines, aggressive action percussion, explosive opening shout, air rush during the leap, rail grab and one huge metal impact, braking tires, powerful beer spray, glass filling and gulp. Duck effects and score under the opening Japanese shout. This is the ONLY spoken line. No speech during or after drinking, no final slogan and no product-name call; the product announcement comes in post. End music with an energetic sting that resolves by 15 seconds; do not fade the image.

Hard constraints: forward motion always cab-first toward screen right; every approaching Sobaya rides a motorcycle; the lead performs a visible airborne leap before grabbing the tanker; no pedestrians in any shot. One truck, six riders initially, five rider-occupied bikes after the leader jumps. One fixed leak begins only after impact. Leader has exactly two anatomically connected arms; one left rail-gripping hand and one right mug-holding hand during pouring and drinking. No human reflections or phantom limbs near the pour. Mug is empty, then full, then its level drops during the final uninterrupted drink. The last frame still shows drinking. Preserve all masks, gray skin and identity. No generated captions, titles, split screens, watermark, product can, endcard or black fade. Only the physical tanker artwork contains lettering. Keep action big and readable, do not add unrelated stunts that obscure the airborne boarding or beer pour.

# Additional Fukuchan reel — 10 seconds; 0–6 intro, 6–10 outro

Create exactly one 10-second 16:9 photoreal cinematic Japanese beer-commercial insert reel. It contains two scenes in the SAME silver tanker cab with the SAME driver: six seconds BEFORE a motorcycle attack, then a hard cut to four seconds AFTER that attack. The already completed chase footage will be inserted between these two scenes by the editor. Do not regenerate the chase. Warm hard afternoon sun, pale blue sky, dry rocky desert and gray asphalt match the tanker reference. 480P, synchronized Japanese speech and subtle natural cab sound, NO MUSIC.

References (Image1 = Image 1, Image2 = Image 2, Image3 = Image 3, Audio1 = Audio 1):
Image 1: Fukuchan character sheet. Preserve the EXACT living man's face from the facial close-ups: facial proportions, eyes, nose, mouth, black hair, skin and slim build. Black blazer, printed white T-shirt, black trousers, sponsor lanyard. Clothing-only panels are diagrams, not headless people. One complete man only, not sheet panels or multiple views. Reference is identity, NOT a frozen facial expression. Animate his real lips, jaw and cheeks naturally for every phoneme.
Image 2: silver tanker truck. This is the vehicle Fukuchan drives; its silver cab, windshield geometry and industrial styling inform the interior. Use a realistic Japanese right-hand-drive truck cab, driver's seat on the vehicle's right, steering wheel directly in front of the driver, a visible shoulder seat belt. Do not copy the reference studio background or place a miniature truck inside the cab.
Image 3: still frame from the existing highway chase. Match the terrain, daylight, dust color, atmosphere and vehicle finish ONLY. Do not reproduce its motorcycle riders inside the cab or its action in this reel.
Audio 1: canonical Fukuchan voice-timbre reference ONLY. Preserve his bright energetic Japanese adult male voice. Do not copy sample words or timing. Generate the two short lines below in Japanese with natural clear diction.

Continuity: same face, black blazer, white printed shirt, seat belt, steering wheel, truck interior and camera side in both scenes. Camera is mounted near the front passenger/dashboard area, stable three-quarter chest-up view of the driver; face and mouth are well lit and unblocked. Truck travels nose-first along the straight desert road before the attack; only landscape motion is visible through the side window, never backwards driving. No other person inside the cab. Keep dashboard markings unreadable and do not generate subtitles or road-sign text: the editor adds a brief accurate warning-sign insert.

Shot 1 [0.0-2.5 s] — Calm cab interior before the attack. Fukuchan drives with both hands on the steering wheel, shoulder belt on, relaxed smile. Desert landscape slides naturally backward outside. At 0.8 seconds his eyes glance briefly toward a roadside warning sign offscreen, then back toward the road. He stays silent, mouth closed. No sign needs to be visible in this shot. The editor will cut away to the warning sign around 1.0-2.5 seconds, using this scene's engine sound. No music, just quiet diesel hum and tire noise.

Shot 2 [2.5-6.0 s] — Same cab, stable slightly closer chest-up portrait, forward travel continues. He gives a small amused smile and says from 2.8 to 5.2 seconds exactly once: "変な看板ギュン。" Pronunciation: hen-na kan-ban gyun. Lightly dismissive, cheerful and natural; he finds the warning sign ridiculous. The final word is a single clear GYUN, not jun or gun. Keep eyes mostly toward the road, hands on wheel, no large head turn during speech. Lips, jaw and cheeks synchronize to the spoken phonemes. Hold his face until the final sound ends and mouth closes. At 5.4-6.0 his eyes briefly flick toward the exterior mirror with mild curiosity, as if something approaches; no additional speech. End this scene cleanly at exactly 6.0 seconds. Do not include the attack or beer in this scene.

Shot 3 [6.0-10.0 s] — HARD CUT to the same driver and cab AFTER the attack. Truck is now stopped; desert outside is stationary and a little dusty. Fukuchan remains belted in his seat, one hand lightly on the steering wheel. From 6.0 to 6.5 he glances toward the exterior side mirror, registering that the unseen raider is just drinking beer on the side of the truck. Do NOT show a reflection, screen, inset or duplicate person; the preceding existing film supplies what he sees. Then his gaze settles near the road ahead in a readable three-quarter face shot. From 6.6 to 9.3 seconds he says exactly once: "飲みに来ただけギュン。" Pronunciation: no-mi-ni ki-ta da-ke gyun. A gentle incredulous, relieved, dryly comic realization, same bright Fukuchan timbre, slight breath at the end. Lips, jaw and cheeks synchronize precisely, mouth unblocked. Keep camera and head stable, hold through the final phoneme, then closed-mouth bemused smile until 10.0. No beer mug in his hands, no drinking by the driver, no additional line.

Sound: generate only Fukuchan's two exact Japanese lines, quiet diesel/cabin ambience and faint tire sound in the first six seconds, soft idle and settling dust ambience after the hard cut. Speech prominent and intelligible. No musical score, commercial jingle, narrator, motorcycle shout, brand call, extra words, laugh track or other speaker. Never mix the reference recording into the new scene.

Hard constraints: exactly 10 seconds; first scene ends at 6.0, after-attack scene occupies 6.0-10.0. Same recognizable Fukuchan face throughout. One person only. No headless bodies, sheet backgrounds, generated captions, warning text, watermarks, product endcard or black transition. Driver remains seated and belted. Both spoken lines are visible synchronized speech with no cut before the final phoneme. Preserve normal facial anatomy and exact identity. The editor will reuse an existing 15-second action film between these scenes; do not spend any of these ten seconds recreating it.