マドギワ族の秘密

プロンプト

現代・過去v2の実送信プロンプト（順に収録） · revision 1

Create exactly one 24-second 16:9 comedy video with synchronized Japanese audio. This request contains ONLY PRESENT-DAY office introduction and ending, in that order. A separate flashback will be inserted in editing; do not generate or visualize it here.

References (separate numbering by media type):
Image 1: Fukuchan PRESENT wardrobe sheet: exact face, swept-back black hair, orange-red mirrored wraparound sunglasses, navy Madogiwa T-shirt, black pants, white sneakers, wristwatch. Wear this outfit throughout.
Image 2: Yametaro exact stylized identity, purple shirt and compact body, speech mouth reference.
Image 3: Yotan face, blond hair, sunglasses, black leather rock clothing. No guitar or guitar strap in this scene.
Image 4: Actual navy T-shirt photo, authoritative chest print, color and placement. Preserve the printed character montage, 窓際族 and WINDOW-SIDE CREW. Printed characters remain flat ink, never become extra cast members. Do not copy the floor or hanging clothing tags.
Audio 1: Fukuchan voice-timbre reference only, bright and personable adult male voice. Feminine sentence endings are acting, not a different speaker.
Audio 2: Yametaro voice-timbre reference only, relaxed Kansai cadence.
Audio 3: Yotan voice-timbre reference only, same voice in flashback.

Continuity: exactly THREE physically present characters, Fukuchan, Yametaro, Yotan. Clean ordinary modern office with desks, rolling chairs and city-facing windows; daylight, normal ceiling, no jungle vegetation, roots, vines, ropes or weapons. No other people, even in background. Stage Fukuchan left, Yametaro center, Yotan right; keep layout, clothes, lighting and camera axis unchanged from introduction to ending. Start immediately with Fukuchan and his shirt readable. Eye-level medium shots for conversation and individual chest-up coverage. Use proportionate low camera for Yametaro; do not stretch his body.

Yametaro identity lock — Image 2 is the exact design authority, not loose inspiration. Preserve its head width-to-height ratio, hair silhouette and hairline, cheek and ear placement, short neck and compact body proportions in every frame. Keep the round black spectacle frames and the sheet's simple white lens/eye areas; do not invent visible pupils, irises, realistic eyeballs, eye sockets, eyelids or eyebrows. Keep the tiny black forehead mark separate from the low, subtle skin-colored nose; no black nose dot, outlined nostrils or projecting nose. Preserve the round cheeks and flat pink blush marks, smooth skin-colored chin, ears and nape, with no black contour bands or facial creases. Do not lengthen the forehead/head, inflate the hair, enlarge ears or add realistic lips. Animate the small inset mouth using the sheet's rest and A/I/U/E/O views as articulation guidance, with visible mouth interior when speaking; do not freeze the smile, protrude the lips or stretch the mouth into a new facial design. Emotion and cinematic lighting must not redesign the face. Wardrobe follows the scene instructions independently of identity. A previous generated take is not an identity reference.

Fukuchan's skin quality and apparent age must match the close-up portraits in Image 1. Keep his well-cared-for, clean, even complexion and natural fine skin texture. Preserve the sheet's subtle forehead lines, eye-area wrinkles, smile lines and nasolabial folds, including their natural changes with expression. Do not add or exaggerate dark spots, freckles, blemishes, rough pores or mottled skin beyond the sheet. Do not erase all wrinkles, de-age him, reshape his face, or give him plastic, waxy or porcelain skin. Preserve the scene's lighting and natural facial shading without turning shadows into blemishes.

Natural speech takes priority over approximate shot timestamps. Shot time ranges are editorial guides, not speech-duration targets or mandatory utterance start/end times. Deliver every line at a natural conversational pace appropriate to its emotion. Do not stretch syllables, vowels, final words or pauses to fill a time window, and do not rush speech to hit a timestamp. Let each utterance finish naturally, then close the mouth. Use remaining time for silent reactions, glances, small gestures and ambience. Speak only the dialogue written here. Adjust internal cuts to complete natural utterances, preserving scene order and total duration.
Use character sheets only for identity, body and wardrobe. Never reproduce sheet panels, labels, white backgrounds, headless wardrobe mannequins or multiple views as multiple people. For headless wardrobe views, use the face close-up to attach the SAME person's head naturally. Photorealistic environments and live-action humans coexist with the exact stylized nonhuman designs. Do not turn stylized characters into humans. Generate Japanese dialogue and facial animation together. Visible speakers: stable chest-up or close framing with mouth visible, natural lips/jaw/cheeks synchronized to generated phonemes; close mouth in silence and hold each speaker until the final phoneme and closure. Sobaya's fixed mask, when present, never deforms for speech. No subtitles, captions, title cards, watermarks, narration or background music. No invented or repeated dialogue. Reference audio is voice-timbre only: never copy its original words or timings.

Shot 1 [0-4 s] — Fukuchan cheerfully shows the chest print by lightly spreading the shirt sides, no covering the design. Says exactly once: 「窓際族Tシャツ買ったわよー」. Read 窓際族 as まどぎわぞく, Tシャツ as ティーシャツ. Natural cheerful final よー, not a sustained timing filler.
Shot 2 [4-12 s] — Yametaro looks at the shirt then at Yotan, mouth visible, says in one natural conversational turn: 「おー、かっこええな！そういえば、なんでよーたんたちまで窓際族なん？」. Others listen silently. よーたん is よーたん, 窓際族 is まどぎわぞく.
Shot 3 [12-16 s] — Yotan close-medium, serious storyteller delivery: 「実はだな」. Read じつはだな. Finish and close mouth, then hold a clean silent reaction for roughly one second. This is an EDIT HANDLE, not an invitation to add a line. All three remain in ordinary office.
Shot 4 [16-20 s] — After the silent edit handle, Yotan, same placement and clothes: 「ということがあったんだ」. This is the return-from-flashback line; no flashback images in this request. Calm and matter-of-fact.
Shot 5 [20-24 s] — Fukuchan cheerfully delivers: 「ギュンギュンどころじゃなかったわね！」. Read ぎゅんぎゅんどころじゃなかったわね. End with natural silent reaction in same office.

Sound: understated office room tone and soft air conditioning, cloth movement. No birds, monkeys, insects or jungle sound. No background conversation. No music. Exactly five spoken turns in written order. Keep edit handle between 実はだな and ということがあったんだ free of overlapping dialogue.
Create exactly one 28-second 16:9 surreal jungle-flashback comedy with synchronized Japanese dialogue and dense jungle ambience. ONLY the past, no present-day office conversation or transition back within this request.

References:
Image 1: Yotan character sheet. The FACE CLOSE-UPS are the exact identity and apparent-age authority. Match cheek volume, jawline, smoothness, nose, lips, face proportions, blond hair and tinted glasses precisely. No added moustache, beard, stubble, beard shadow, age spots, deep wrinkles, sunken cheeks or older-looking skin. Do not interpret photorealism as aging. Keep black leather wardrobe, remove guitar and strap. He is slim, 170cm tall.
Image 2: Okayaman full-body sheet, physical person, 180 cm tall slim build, black zipped hooded jacket, black slim chinos, black sneakers, exact face, medium black hair and beard. He is taller than Yotan (170 cm) and Tokun (165 cm), not muscular. Do not depict him on a monitor.
Image 3: Tokun character sheet. FACE CLOSE-UPS are exact identity and apparent-age authority: preserve rounded cheeks, short chin, face width, nose, lips, black hair, straw hat and tinted glasses. Clean-shaven appearance as in sheet: no added moustache, beard, stubble or dark beard shadow; no extra wrinkles, sagging or age spots. Preserve Hawaiian shirt, pink lei, dark loose trousers, slightly plump 165cm body. No ukulele.
Image 4: Sobaya, muscular 180 cm build, neutral gray exposed skin, white mask with exact black and red features, white T-shirt, dark trousers and white shoes. Use the UPDATED sheet: preserve broad chest-muscle shadows and natural shirt folds, but no nipple-shaped protrusions, pinpoint nipple shadows or corresponding tiny highlights through fabric. Do not flatten chest muscles. Maintain this in every view. Normal outfit, no weapon. One ordinary beer mug at his side, never raised to remove mask.
Image 5: Takosan exact compact toy-diorama 3D design, black hooded robe, white face, solid black eyes, two humanoid arms ending in white rounded undivided hands and SIX lower-body tentacles arranged radially. Keep arms distinct from lower tentacles. Use small speaking mouth as in sheet.
Image 6: Fukuchan ordinary outfit and hair from basic sheet, black jacket over white graphic shirt, dark trousers, white shoes; skin and face reference in close-up panels. No navy Madogiwa T-shirt, no orange mirrored sunglasses in this past scene.
Image 7: Cyan sasumata prop sheet. Exactly TWO actual props: Fukuchan uses the LONG-handled left design, Takosan the SHORT-handled right design. Same cyan chipped steel, same U-shaped head and sharp inner spikes; small Yumemin logo on straight shaft BELOW joining collar, NOT on the collar. The enlarged head is detail only, not a third prop. No pink version.
Audio 1: Yotan voice-timbre reference only, same person as current scene.
Audio 2: Sobaya voice-timbre reference only, retain its deep rough monstrous resonance, speak the new clipped Japanese line, do not copy source wording.
Audio 3: Okayaman voice-timbre reference only. Preserve speaker identity while intensifying into a terrified scream.
Audio 4: Takosan voice-timbre reference only, the user-adopted A voice from episode 80. Match this exact vocal identity, pitch character, resonance and texture; do not invent a new researcher voice or deepen it toward Sobaya. Only change acting to the clipped ultimatum written below. Do not copy the reference words. Tokun has no canonical voice entry in VOICE_CAST: generate a natural mid-register Japanese adult male voice with apprehensive polite speech. Fukuchan is completely silent in this request.

Environment is described ONLY in text. No background image or previous video is supplied. A credible real modern city office has a dense tropical jungle growing INSIDE its window-side area. Real photographic depth, tangible bark and foliage, ordinary office ceiling and floor-to-ceiling windows with city buildings outside, occasional desk and chair half obscured by leaves. Dappled natural window light, restrained natural contrast, no artificial illustration gloss, no theatrical backdrop or miniature diorama environment. Compose a new coherent space rather than copying earlier staging.


Continuity: exactly SIX characters overall. At opening ONLY Yotan, Okayaman and Tokun are visible, all seated side by side ON THE FLOOR against the SAME front surface of ONE enormous broad tree trunk. One trunk is wide enough for all three. Backs contact the bark, knees bent naturally, feet on floor, several horizontal turns of coarse rope bind their torsos and upper arms around the same trunk. No separate trees, no standing, no floating, no ropes around necks or faces. Maintain seated posture and visible chest ropes in close-ups. Left Yotan, center Okayaman, right Tokun. Maintain ropes and relative stature throughout. Sobaya, Takosan and Fukuchan enter together later through the central foliage-lined passage, exactly one of each. Sobaya center leader, Fukuchan with long cyan sasumata, Takosan with short cyan sasumata. Takosan holds shaft using rounded hands without fingers; lower tentacles remain lower limbs. No costume or body mutation. Props never duplicate or switch owners. The threat is acted through looks and posture. No stabbing action is scripted.
Sobaya keeps the exact rigid white mask fully correctly fitted throughout all shots, never lifts, shifts, removes, becomes transparent or transforms into exposed human face. Speech comes from behind mask, use head and body acting, not a moving mask mouth.

Fukuchan's skin quality and apparent age must match the close-up portraits in Image 6. Keep his well-cared-for, clean, even complexion and natural fine skin texture. Preserve the sheet's subtle forehead lines, eye-area wrinkles, smile lines and nasolabial folds, including their natural changes with expression. Do not add or exaggerate dark spots, freckles, blemishes, rough pores or mottled skin beyond the sheet. Do not erase all wrinkles, de-age him, reshape his face, or give him plastic, waxy or porcelain skin. Preserve the scene's lighting and natural facial shading without turning shadows into blemishes.

Natural speech takes priority over approximate shot timestamps. Shot time ranges are editorial guides, not speech-duration targets or mandatory utterance start/end times. Deliver every line at a natural conversational pace appropriate to its emotion. Do not stretch syllables, vowels, final words or pauses to fill a time window, and do not rush speech to hit a timestamp. Let each utterance finish naturally, then close the mouth. Use remaining time for silent reactions, glances, small gestures and ambience. Speak only the dialogue written here. Adjust internal cuts to complete natural utterances, preserving scene order and total duration.
Use character sheets only for identity, body and wardrobe. Never reproduce sheet panels, labels, white backgrounds, headless wardrobe mannequins or multiple views as multiple people. For headless wardrobe views, use the face close-up to attach the SAME person's head naturally. Photorealistic environments and live-action humans coexist with the exact stylized nonhuman designs. Do not turn stylized characters into humans. Generate Japanese dialogue and facial animation together. Visible speakers: stable chest-up or close framing with mouth visible, natural lips/jaw/cheeks synchronized to generated phonemes; close mouth in silence and hold each speaker until the final phoneme and closure. Sobaya's fixed mask, when present, never deforms for speech. No subtitles, captions, title cards, watermarks, narration or background music. No invented or repeated dialogue. Reference audio is voice-timbre only: never copy its original words or timings.

Shot 1 [0-3 s] — Wide establishing shot: all three men SEATED and tied side by side to ONE shared enormous trunk, all seated bodies readable, Okayaman noticeably tall and slim; foreground jungle foliage, office ceiling and city windows readable in same image. Immediate birds and distant monkeys. Brief orienting glances, no dialogue yet.
Shot 2 [3-7 s] — Yotan chest-up, SEATED with rope crossing his chest visible, face and age exactly as sheet, no added stubble, restrained by rope, confused: 「どこだここ」. Hold until finished, then listen.
Shot 3 [7-12 s] — Tokun chest-up, SEATED with rope crossing chest visible, clean-shaven sheet face and age, worried realization: 「ここってマドギワ族の縄張りじゃないですか？」. マドギワ族 is まどぎわぞく; 縄張り is なわばり. No ad-libbed greeting.
Shot 4 [12-18 s] — The three captors emerge through leaves, with office window mullions and ceiling glimpsed beyond foliage so this remains an indoor office jungle. Sobaya in center, both cyan props recognizable at their owners' sides. Cut closer to masked Sobaya, calm authoritative clipped speech: 「オマエラ、マドギワ、ハイッタ」. Pronounce おまえら、まどぎわ、はいった. Short deliberate breaks between phrases, no invented words. Mask stays fitted and rigid.
Shot 5 [18-23 s] — Takosan close-medium, short pole visible, small mouth articulating: 「仲間ナル、ギュンされる、エラべ」. Pronounce なかまなる、ぎゅんされる、えらべ. Dry ultimatum with clipped phrasing. Briefly incline fork toward captives from a distance. Fukuchan remains silent.
Shot 6 [23-25 s] — Brief close-up of Fukuchan holding LONG cyan sasumata upright. He gives an unsettling friendly smile, entirely silently. A distant monkey call punctuates the gap; no human chuckle or spoken ギュン.
Shot 7 [25-28 s] — Okayaman seated against the SAME trunk, face close-up with ropes visible below, visibly recoils against restraint, opens mouth in a terrified expression (not a friendly conversational smile), suddenly screams at full terrified intensity but with precise polite diction: 「大変驚いております！」. Pronounce たいへんおどろいております. This is the only screamed line, no extra scream syllables or added self-introduction. Preserve identity through expression. Hold through the final syllable and tiny reaction tail. End in jungle, ready for hard editorial cut to present.

Sound: completely immersive outdoor tropical JUNGLE ambience despite office visuals. Tropical birds echo through foliage, occasional distant monkey cries, continuous low insect bed, wind and leaf rustle, quiet rope/cloth and footsteps. Dialogue clearly foremost; let animal calls rise BETWEEN utterances, especially silent smile. No air conditioning, phones, keyboards, traffic, city or office room-tone sounds. No music or narrator. Do not show birds or monkeys as extra characters. Exactly five spoken turns, Fukuchan silent. No Yametaro here. No clothing swap with present. No need to make scenes match current office lighting: consistent dappled jungle daylight throughout this request.