東京マドギワランド

プロンプト

採用5カット実送信本文とCM · revision 2

# 東京マドギワランド 修正版
採用順1→2→5→4→3→CM。39.033秒。初回生成はシーン4のみ採用。以下は実送信本文。

## prompt_wan3_submitted_v1.txt
Create exactly one 30-second 16:9 live-action Japanese selfie travel vlog with synchronized Japanese dialogue. Tokyo Madogiwa Park is a Tokyo Disneyland parody themed around the Madogiwa characters. Cheerful sunny premium theme park, absurd dry office humor. Fukuchan remains delighted while making cutting remarks. This is one multi-shot generation, not separate videos.

References, ordered:
Image 1: character_fukuchan_basic_sheet.png — Fukuchan's exact real face, slim build, black hair and clothing. An adult MAN with bright friendly voice; feminine Japanese sentence endings do not change gender or identity. Holds one selfie stick in one hand throughout.
Image 2: character_takosan_park_mascot_sheet.png — full-size black hooded plush mascot, white face, black eyes, two human-shaped arms with round white fingerless hands, six lower-body tentacles. Entire costume levitates, no feet.
Image 3: character_yumemin_park_mascot_sheet.png — round blue plush tapir mascot with white rear, tiny ears, black eyes, flexible trunk. No human legs. Levitates.
Image 4: character_yametaro_park_mascot_sheet.png — ground-standing plush mascot, purple shirt, black pants, canonical face, white spectacle lenses.
Image 5: character_sobaya_basic_sheet.png — design for multiple adult theme park cast members costumed as Sobaya: white mask, gray exposed skin, muscular build. Multiple cast members are intentional, not one teleporting person.
Image 6: character_okayaman_basic_sheet.png — exact real face of Okayaman, on theater screen.
Image 7: environment_madogiwa_castle.png — castle and parade plaza.
Image 8: environment_shops_theater_sheet.png — top right shop interior, bottom right theater interior; other panels establish architecture. Never reproduce panel layout.
Image 9: prop_madogiwazoku_tshirt.webp — use only the white shirt and exact printed character artwork, 窓際族 and WINDOW-SIDE CREW. Ignore the promotional layout, SUZURI logo and background marks; these are not printed on the shirt.
Audio 1: voice_fukuchan.wav — voice-timbre reference only, no copied words or timing.
Audio 2: voice_okayaman.wav — voice-timbre reference only, no copied words or timing.

Character sheets are design references, never duplicated poses, labels or studio backgrounds.
Yametaro identity lock — Image 4 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.
Yametaro is silent in this episode; no speaking mouth animation is needed. Match low side-view nose and continuous skin-colored jaw. Costume fleece is the intentional medium change, not a new face.

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 or ambience. Speak only the dialogue written below. Adjust internal cut boundaries to complete natural utterances while preserving scene order and total duration of 30 seconds.

Shot 1 [0-4 s]: Selfie-stick smartphone perspective, Fukuchan chest-up with castle behind. Several Sobaya cast members wave independently in background. Fukuchan happily says 「今日は東京マドギワパークに来たわよー！」.

Shot 2 [4-12 s]: Same selfie framing with Takosan and Yumemin floating visibly clear of the ground behind his shoulders. Their fleece bodies sway softly; detached shadows show real elevation. Yametaro stands on pavement and waves. Fukuchan says 「この着ぐるみ、窓際族がオフィスで浮いてるエネルギーで飛んでるらしいわよ！」. Deliver this as a cheerful mock-factual explanation: the energy of the Madogiwa crew being socially out of place at the office powers these two costumes' literal levitation. This is a verbal pun; remain in the park, with no office cutaway, energy beams or added visual effects. Both floating mascots slowly turn toward him. No flying wires or exposed performers.

Shot 3 [12-18 s]: Cut to shop interior, same selfie viewpoint. Free hand holds a hanger with ONE white T-shirt at chest height beside face; other hand holds selfie stick. Show actual reference print facing camera, unfolded and sufficiently large, without covering his mouth. A Sobaya clerk stands behind. Fukuchan says 「このTシャツ、公式サイトでも買えるんですって！」. No invented website URL or price.

Shot 4 [18-23 s]: Selfie camera frames Fukuchan at foreground edge and parade action across the background. A decorated moving float carries Yametaro waving; Takosan and Yumemin hover beside it. Several Sobaya cast members perform entirely independent actions: one skilled dancer doing windmills, another dancing slow bon-odori, another awkwardly missing the beat, and one standing absolutely still drinking beer from a large mug through a straw at the mask mouth opening. Maintain correctly worn opaque white masks, no lifting or removing, uniformly gray exposed skin. No synchronized routine, no shared final freeze, no coordinated formation. Acrobatic skill is impressive only for the individual; others clearly slacked off rehearsal. Keep drinker plainly visible in same frame. Fukuchan brightly remarks 「練習サボってそうな人がいっぱいいるわねー！」.

Shot 5 [23-30 s]: Theater selfie view with Fukuchan and large screen showing Okayaman's face clearly. Fukuchan says 「仕事抜け出して来ちゃった！」 then stops. Okayaman responds smiling, visibly very surprised with a slight lean forward: 「おかやまん。大変驚いております」. Hold amused selfie ending, no extra spoken tag.

Speech and sound: Generate dialogue, facial motion, light cheerful original theme park parade music, crowd ambience and dance sounds together. Music below speech. Fukuchan uses his canonical bright adult male voice with the exact requested playful feminine phrasing; Okayaman has his canonical calm warm adult male voice. Pronunciation: 東京マドギワパーク=とうきょうマドギワパーク; 着ぐるみ=きぐるみ; 窓際族=まどぎわぞく; Tシャツ=ティーシャツ; 公式=こうしき. No other intelligible speech.
Lip sync: mouth, jaw and cheeks synchronized to generated phonemes. Keep visible speaking mouths clear and reasonably large, retain each speaker through final phoneme and closed mouth before cutting; no fast pans during speech. Mascots and Sobaya cast have no dialogue.
Hard constraints: 30 seconds; one selfie stick; one Fukuchan; one of each mascot; multiple Sobaya cast intentionally; only Takosan and Yumemin levitate; grounded Yametaro; no mascot costume removal; no coordinated Sobaya dance; at least one motionless beer drinker; exact provided dialogue once and in order; no subtitles, captions, watermark, invented brand signage or URL. Do not reproduce reference boards as split screens.



## prompt_wan3_scene01_v2.txt
Photoreal Japanese selfie-stick theme park vlog. 16:9, 480P, natural synchronized Japanese speech. Same costume and real identity throughout. Facial motion must match generated speech. Keep mouths visible while speaking; close mouth after final syllable. Never stretch vowels or rush speech to fill duration. No subtitles, captions, watermark or additional intelligible dialogue. Bright cheerful adult MALE Fukuchan voice despite feminine Japanese phrasing. All audio references are VOICE TIMBRE ONLY; do not copy their words or timing. Reference sheets provide identity and wardrobe only, never their panels or duplicate bodies. Do not add headless figures from wardrobe panels.
Create one 5-second replacement opening shot. The correct place name is 東京マドギワランド (とうきょう・まどぎわ・らんど), NEVER 東京マドギワパーク.
Image 1: Fukuchan canonical real face, black hair, slim build, black jacket, white printed shirt and SPONSOR lanyard.
Image 2: Sobaya cast costume, muscular gray exposed skin, white mask with red markings, white T-shirt.
Image 3: castle environment, ivory stone blue turrets and huge office windows.
Audio 1: Fukuchan voice.
Smartphone selfie perspective. Exactly one Fukuchan fills foreground chest-up, holding the selfie stick with one hand. Ivory castle behind him in sunny daylight. Three Sobaya-costumed cast members stand far behind and wave independently, masks continuously worn. No other named characters.
Fukuchan smiles at lens and says exactly once: 「今日は東京マドギワランドに来たわよー！」. Pronounce ランド clearly. Finish naturally, then hold closed-mouth smile for a short editing handle. No pan or cut while speaking. Subtle bright park ambience and low cheerful original music.


## prompt_wan3_scene02_v3.txt
CRITICAL CAST COUNT: exactly FOUR physical characters total: one Fukuchan, one Takosan, one Yumemin, one Yametaro. Reference sheets show multiple views of ONE mascot; never instantiate multiple views as separate characters. At opening, place ONE Yumemin far left, ONE Takosan center-left, Fukuchan center-right, ONE Yametaro far right. Track the same single body of each character continuously. When Yametaro reacts, move the existing far-right Yametaro toward Takosan; never spawn a second Yametaro. When Takosan rises, his old position becomes EMPTY. No double exposures or duplicate bodies.

Create one 10-second replacement mascot scene for 東京マドギワランド. Photoreal Japanese selfie vlog, 16:9, 480P. Fukuchan foreground holding one selfie stick, blue-turret ivory castle plaza behind him. Preserve exact reference identities, no reference-board layout.
Image 1: character_fukuchan_basic_sheet.png — Fukuchan, clean-shaven adult man, black jacket, slim build, exact real face.
Image 2: character_takosan_park_mascot_sheet.png — black hooded plush Takosan, white face, two round white hands and six lower tentacles.
Image 3: character_yumemin_park_mascot_sheet.png — blue round tapir plush Yumemin with white rear and trunk, no legs.
Image 4: character_yametaro_park_mascot_sheet.png — Yametaro plush mascot, purple shirt, white spectacle lenses, standing on pavement.
Image 5: environment_madogiwa_castle.png — park environment.
Audio 1: voice_fukuchan.wav — voice timbre only, not source dialogue or timing.
Yametaro identity lock — Image 4 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.
Yametaro is silent in this episode; no speaking mouth animation is needed. Match low side-view nose and continuous skin-colored jaw. Costume fleece is the intentional medium change, not a new face.


At first Takosan and Yumemin gently hover at low height behind Fukuchan, their detached shadows visible. Yametaro remains grounded. Fukuchan cheerfully says exactly once: 「この着ぐるみ、窓際族がオフィスで浮いてるエネルギーで飛んでるらしいわよ！」. Natural conversational speech with synchronized lips, jaw and cheeks; do not rush or stretch syllables. Keep his speaking face clear.
Near the end of the explanation, Takosan alone begins drifting helplessly upward like a helium balloon just released from someone's hand. His entire soft costume bobs, gently tilts and rotates in the breeze, tentacles trailing beneath him. Continuous smooth buoyant rise, diagonally upward and farther away, becoming smaller before disappearing beyond the top of frame. No wing flapping, swimming strokes, rocket propulsion, energy beam, sudden teleportation or visible wire. He is excessively socially adrift at the office, so his buoyancy carries him away; convey this visually without an added explanation.
Yametaro and Yumemin notice and visibly panic within the limits of their bulky plush costumes. Grounded Yametaro looks up, hurriedly waves his short arms and takes flustered little steps beneath Takosan. Yumemin stays at low hovering height, frantically bobs and turns his whole round body, lifting his trunk toward the departing Takosan. No invented arms or legs on Yumemin. Neither follows Takosan into the sky. Keep their reactions visible in the same shot. After the line, Fukuchan closes his mouth and glances upward. Hold enough silent reaction to see Takosan leave frame, without adding dialogue.
Generate native speech, gentle original park ambience and low music. Exactly one of each named character, stable identities and costume structures. No subtitles, watermark, costume removal or additional intelligible speech.


## prompt_wan3_scene05_v4.txt
Photoreal Japanese selfie-stick theme park vlog. 16:9, 480P, natural synchronized Japanese speech. Same costume and real identity throughout. Facial motion must match generated speech. Keep mouths visible while speaking; close mouth after final syllable. Never stretch vowels or rush speech to fill duration. No subtitles, captions, watermark or additional intelligible dialogue. Bright cheerful adult MALE Fukuchan voice despite feminine Japanese phrasing. All audio references are VOICE TIMBRE ONLY; do not copy their words or timing. Reference sheets provide identity and wardrobe only, never their panels or duplicate bodies. Do not add headless figures from wardrobe panels.
Create one 9-second replacement theater shot. Fukuchan is the sole named physical host. An audience of adult male and female fans occupies the theater seats behind him. OKAYAMAN EXISTS ONLY AS A VIDEO IMAGE ON ONE WALL DISPLAY. There must never be a physical Okayaman, no duplicate Okayaman, and Fukuchan must never acquire facial hair or change into Okayaman.
Image 1: FUKUCHAN, the ONLY foreground real person and camera holder. Preserve his clean-shaven canonical face, slim build, black hair, black blazer, white printed shirt and SPONSOR lanyard.
Image 2: OKAYAMAN, a DIFFERENT man, with mustache and beard and black hooded jacket. Use his face ONLY INSIDE the wall display's rectangular bezel. Never apply Image 2 to the foreground host.
Image 3: theater environment from bottom-right reference panel only: gold and ivory theater, blue seats and large wall display.
Audio 1: Fukuchan voice. Audio 2: Okayaman voice.
One continuous stable selfie composition filmed FROM THE REAR OF THE AUDITORIUM. Fukuchan stands in the rear aisle at foreground LEFT, holding the selfie stick; the camera looks over the audience toward the stage and wall screen at the FRONT, visible at upper RIGHT with its bezel. Audience members sit BETWEEN Fukuchan and the distant screen. Show the BACKS of their heads and shoulders, with occasional side profiles, never rows of audience faces looking into the camera. The wall display shows only one chest-up Okayaman. Foreground Fukuchan and onscreen Okayaman both remain visible throughout, never identity swap. Show several rows of adult male and female audience members in blue seats, all seated facing FORWARD toward Okayaman on the stage screen, away from Fukuchan and away from the selfie camera. The seat backs face the camera. Fukuchan is an incidental visitor at the back, NOT the performer receiving attention. No mirrors, portraits or additional screens. Audience faces must not duplicate Fukuchan or Okayaman.
First Fukuchan cheerfully says 「仕事抜け出して来ちゃった！」. ONLY foreground Fukuchan moves his mouth for this sentence; onscreen Okayaman silently listens with mouth closed. Then Fukuchan closes his mouth and looks toward the screen.
Then ONLY the man on the wall display, Okayaman, smiles with surprised reaction and says 「おかやまん。大変驚いております」. Foreground Fukuchan listens silently with mouth closed and remains recognizably Fukuchan. Keep both faces unchanged during the speaking turn. After Okayaman finishes the final syllable of 「大変驚いております」 and closes his mouth, the seated audience bursts into enthusiastic fan cheers and applause. Clearly audible layered adult MALE AND FEMALE cheers, lower male whoops mixed with higher female cheers; neither an all-female scream nor an all-male chant. Show fans raising their hands toward the SCREEN, clapping and leaning toward the STAGE with delight. Their gaze, torsos and cheers remain directed at Okayaman on the screen for the ENTIRE shot. Never turn toward Fukuchan or the camera, even during cheers. Preserve back-of-head views so attention direction is unambiguous. Hold this reaction for about two seconds at the end. Do not cover either spoken line with cheering. No additional scripted words or names. Keep Fukuchan physically in foreground and Okayaman exclusively within the screen. Stable selfie framing includes the audience reaction; no full-screen Okayaman shot. Generate the cheers and applause natively together with the video. Low theater ambience during speech, then lively fan reaction.


## prompt_wan3_scene03_v4.txt
Create one 6-second replacement scene for a live-action Japanese selfie vlog at 東京マドギワランド. 480P, 16:9, realistic sunny theme-park gift shop, polished but natural smartphone footage. Generate speech and lip motion TOGETHER; the generated embedded audio is the final speech source, not a later voice-over.

Image 1: Fukuchan, exact real face, slim adult Japanese MAN, clean-shaven, black hair, black blazer, white graphic shirt and SPONSOR lanyard. Preserve face throughout. Feminine Japanese phrasing does not change his gender or voice identity.
Image 2: Sobaya shop clerk identity ONLY: white mask with red markings, black hair, muscular build, uniformly neutral gray exposed skin. OVERRIDE HIS PLAIN WHITE SHIRT: HE WEARS THE PRINTED MADOGIWAZOKU T-SHIRT FROM IMAGE 4 for this entire scene.
Image 3: shop interior from the top-right panel only, warm wood fixtures, gold accents and large gridded windows. Never reproduce board layout.
Image 4: exact merchandise design reference. The RIGHT half shows the actual white T-shirt; the LEFT half is a large view of its printed artwork. Use the colorful character group and 窓際族 / WINDOW-SIDE CREW print. SUZURI header, pale background marks and advertising layout are outside the shirt and must not appear on clothing. Do not turn print figures into real people.
Audio 1: Fukuchan canonical voice-timbre reference ONLY. Do not copy reference words or reference timing.

One continuous stable medium-close smartphone SELFIE shot. Exactly two physical people: Fukuchan foreground left and one Sobaya clerk slightly behind on the right. Fukuchan holds one selfie stick in one hand and holds a hanger with ONE merchandise T-shirt in his other hand, chest-high beside his face. Keep Fukuchan's mouth unobstructed and large enough to read. Display shirt front clearly. Sobaya wears a SECOND physical shirt of the SAME design, stretched naturally over his muscular torso. Show his chest print unobstructed at sufficient size; he stands relaxed, silent, friendly. No beer mug obscuring his chest, no extra mascots or other named people. His mask stays correctly worn, opaque and unchanged throughout.

Fukuchan smiles and says exactly ONCE, at a natural conversational pace: 「このTシャツ、こうしきサイトでも買えるんですって！」
The intended words are このTシャツ、公式サイトでも買えるんですって！. The phonetic spelling こうしきサイト is deliberate to prevent the previous mispronunciation. Pronounce it as KOO-SHI-KI SA-I-TO (こう・し・き・さ・い・と), especially the しき followed by サイト. Do not say こうきしたい, こうきさい, or 好奇次第. These pronunciation guides are NOT additional spoken dialogue. Say the line once, never repeat either version. Bright friendly adult male voice with playful feminine phrasing.
Start speaking promptly after a brief smile, finish the short line naturally, then hold a closed-mouth smile while continuing to show BOTH printed T-shirts. Do not extend words or pauses to fill six seconds. No late second sentence. The extra tail is a silent editing handle. Animate lips, jaw and cheeks in sync with actual generated phonemes; close mouth after final syllable. Sobaya never speaks.
Subtle indoor shop ambience and very low cheerful original music. No intelligible background voices. No subtitles, labels, price, URL, watermark, extra dialogue, identity swaps, duplicated host, wrong shirt, blank Sobaya shirt, or shirt hiding either face. Keep the shot and identities stable without cuts or camera pans.


## CM
そば屋「窓際族Tシャツ、公式サイトで好評発売中！」。Irodori v4.1-Small、seed42、正典Sobaya_voice.wav。sobaya_monsterize.sh適用。Remotion商品カード4.8秒。