プロフェッショナル 窓際の流儀

Wan 3.0 30秒・Image 3 + Audio 1プロンプト

Create exactly one 30.0-second, 16:9, photorealistic Japanese television work-profile documentary parody with synchronized Japanese speech and sound. This is an explicit close visual parody homage to the restrained observational language of NHK's “Professional: Shigoto no Ryugi”: dignified public-broadcast seriousness, patient human observation, natural window light, understated handheld 35–50 mm photography, subtle shallow depth of field, muted neutral colors, gentle film grain, confident hard cuts, contemplative pauses, and quiet emotional escalation. The comic subject is a window-side employee who drinks beer and repeatedly opens and closes a spreadsheet. Do not reproduce any original program footage, exact copyrighted music melody, or original logo. Generate an original serious documentary score with sparse piano, restrained atmospheric guitar, and low strings.

References:
- Image 1: episode-local identity sheet for Sobaya. Preserve exactly one Sobaya with the same white mask, black circular eye holes, red vertical mask marks, short black hair, neutral-gray exposed skin, broad muscular 180 cm / 100 kg silhouette, fitted white short-sleeve T-shirt, dark jeans, and white sneakers. The mask is mandatory and remains correctly worn in every frame. Use Image 1 only for identity, anatomy, clothing, mask, and proportions. Do not reproduce its panel layout, labels, white background, multiple views, reaction boxes, or duplicate bodies. Do not treat the sheet's beer mug as an additional prop.
- Image 2: adopted production environment and composition reference, the candid documentary side-angle B. Preserve the same exact high-rise Akasaka window-side workstation throughout: floor-to-ceiling window geometry, Tokyo Tower skyline, concrete wall column, dark desk, black chair, exactly one laptop used by itself through its built-in keyboard and trackpad, plain papers, one clear half-full golden beer mug, soft overcast morning light, restrained gray palette, and lived-in 45-degree observational camera language. There is no external keyboard and no separate mouse. Image 1 overrides Image 2 for Sobaya's identity. Image 3 overrides the small poster inside Image 2 for Yametaro identity, exact poster artwork, and exact reward text. Use Image 2 as one real scene, not as a still to copy rigidly.
- Image 3: exact standalone Yametaro wanted-poster prop. Place exactly one copy of this same poster on the pale office wall or column beside the window throughout every shot where that wall is visible. Preserve its exact beige paper, ornamental border, canonical deformed Yametaro portrait, and the exact text hierarchy “WANTED / YAMETARO / REWARD / ¥200,000,000”. Do not redesign Yametaro as a generic human, change the reward, duplicate the poster, or reproduce the poster as a full-screen title.
- Audio 1: Sobaya's canonical voice-timbre reference only. Preserve this same deep, rough, monster-like adult male speaker identity for all three Japanese answers below. Do not copy, repeat, or paraphrase the words or timing contained in Audio 1. Generate the new scripted lines below exactly once each.

Overall continuity: exactly one Sobaya occupies the same workstation in one continuous morning. Exactly one clear beer mug remains on the desk, initially half full and later slightly lower after one careful sip. Exactly one laptop remains on the same desk and is the only computer equipment; Sobaya uses only its built-in keyboard and trackpad. Exactly one Image 3 wanted poster stays fixed to the same wall position whenever that wall is visible, with unchanged Yametaro identity, design, and reward. No external keyboard, separate mouse, monitor, phone, additional device, second wanted poster, or poster variation appears. The laptop may show a generic spreadsheet-like grid or a blank desktop glow, but no legible words, numbers, icons, brand marks, or fake captions. Sobaya's mandatory white mask never comes off, lifts, shifts, becomes transparent, deforms, or reveals a human face. The mask conceals his mouth, so visible lip motion is not required; synchronize the generated Audio 1-timbre answers to his breathing, head movement, pauses, and hand gestures instead. No interviewer is seen or heard. No other employee enters the foreground. The same desk, chair, window, skyline, props, wardrobe, lighting direction, voice identity, and subdued color grade persist across every shot.

Shot 1 [0.0-4.0 s] — Cold open observational montage with two clean hard cuts. Begin on an exterior-feeling telephoto insert through office glass: Tokyo Tower softly out of focus while Sobaya enters the same workstation background and sits. Cut to a close insert of his large gray hand placing exactly one half-full beer mug beside the laptop at 1.5 seconds. Cut at 2.7 seconds to a medium three-quarter profile as he wakes the laptop with its built-in trackpad and places both hands naturally over the laptop. Leave clean upper-left and lower-left negative space for later title typography. Serious, composed, never slapstick.

Shot 2 [4.0-12.0 s] — Hard cut to the same medium three-quarter interview-style profile at the same desk, matching Image 2. Sobaya raises the one beer mug, takes exactly one calm controlled sip through the mask's small mouth opening without moving or removing the mask, sets the mug back in its original place, exhales gently, flexes his fingers once, and places both hands steadily on the laptop's built-in keyboard and trackpad. In the Audio 1 timbre he says exactly once in natural Japanese: “出社したら真っ先にビールを飲んでます。手の震えが止まって、エクセルの操作が安定するんですよね。” Pronunciation: 出社=しゅっしゃ, 真っ先=まっさき, 手の震え=てのふるえ, エクセル=えくせる, 操作=そうさ, 安定=あんてい. Low, calm, matter-of-fact documentary interview delivery, with a small pause after the first sentence. The beer level falls only slightly and never spills. The mask conceals his mouth; synchronize head nods, breath and finger settling to the speech rhythm. Do not cut before the final phrase ends. Camera makes a nearly imperceptible slow push-in. Keep lower-left clear for name and job-title graphics.

Shot 3 [12.0-18.0 s] — Hard cut to a shoulder-height side close-up of Sobaya's two hands using only the laptop's built-in keyboard and trackpad within the same desk geometry. He opens a generic spreadsheet-like window, closes it, pauses, opens it again, closes it again, then leaves it open. In the same Audio 1 timbre he says exactly once in natural Japanese: “開いたり閉じたりしてます。こうすると働いてるって思われますから。” Pronunciation: 開いたり=ひらいたり, 閉じたり=とじたり, 働いてる=はたらいてる, 思われます=おもわれます. Calm deadpan delivery with a faintly practical tone. The change between grid and plain neutral desktop must be visually obvious even though all text remains unreadable. Exactly two complete open-close cycles, then a final open state. Time the two cycles to the spoken verbs. No external keyboard, separate mouse, impossible cursor movement, extra windows, extra hands, or changing laptop shape. Do not cut before the final phrase ends.

Shot 4 [18.0-22.0 s] — Hard cut to quiet observational B-roll in the same location. Close insert of the now-steady hand resting on the built-in trackpad, the half-full-to-slightly-lower beer mug, reflected Tokyo Tower glass, and Sobaya's still masked profile. One subtle rack focus from the spreadsheet glow to the beer mug, then to his mask. No dialogue and no text.

Shot 5 [22.0-29.0 s] — Hard cut to a chest-up frontal portrait beside the same floor-to-ceiling window. Sobaya is standing, mask still correctly worn, shoulders broad and calm, performing a slightly intellectual but natural interview pose while speaking. In the same Audio 1 timbre he says exactly once in natural Japanese: “窓際に追い込まれてるうちは二流。一流の窓際族は、窓際を作り出す。” Pronunciation: 窓際=まどぎわ, 追い込まれてる=おいこまれてる, 二流=にりゅう, 一流=いちりゅう, 窓際族=まどぎわぞく, 作り出す=つくりだす. Deliver the first sentence as a quiet diagnosis, pause briefly, then deliver the second sentence with controlled conviction. From 22.0 to 25.2 seconds, he folds his left arm across his torso and gently rests his right hand against the chin area of the mask, with a small thoughtful head tilt and steady off-camera gaze. At the pause between sentences, he releases the chin. From 25.2 to 29.0 seconds, he keeps the left arm folded and calmly raises his right index finger near chest height as he states the second sentence, then holds the pose until the final sound ends. The mask conceals his mouth; synchronize breath, head motion and the gesture change to the generated speech. Very slow 50 mm push-in, composed public-broadcast portraiture, clean lower-left and upper-left negative space. Do not cut before the final phrase ends. Intelligent, credible, restrained, not a superhero pose and not exaggerated comedy mugging.

Shot 6 [29.0-30.0 s] — Hold the final window portrait almost still. The original score reaches one restrained emotional chord and decays. Leave clean space for a title added in post.

Audio: Generate exactly the three scripted Sobaya answers above in the same Audio 1 voice identity, once each and only in Shots 2, 3 and 5. No interviewer voice, no narration, no extra dialogue, no paraphrase, no repeated phrase, no whisper, no humming, and no other vocalized words. Also generate subtle office room tone, distant HVAC, one soft beer-glass placement, one quiet sip, built-in trackpad clicks synchronized to the two open-close cycles, restrained laptop-keyboard taps, and a fully original premium documentary underscore using sparse piano, restrained atmospheric guitar texture, and low strings. The music must be serious and emotionally building but must not copy or quote any existing television theme or recognizable melody. Keep ambience and music below the dialogue. Wan embedded audio is the initial synchronization master. If a word is mispronounced, omitted, repeated, paraphrased, or later performance changes are requested, only the affected phrase or shot may be repaired in Remotion using the episode-local canonical Irodori masters, followed by re-audit.

Editing-safe constraints: no generated title, subtitle, caption, lower third, broadcaster logo, clock, readable office signage, readable spreadsheet text, watermark, or interface labels. Keep upper-left and lower-left safe areas free of face, hands, beer mug, laptop, and Tokyo Tower whenever specified. Use clean hard cuts at the exact shot boundaries; no dissolves, wipes, split screens, montage frames, or black question cards inside this Wan source. The black question cards and all exact typography will be inserted later in Remotion.

Hard constraints: exact duration 30.0 seconds; one Wan task; exactly one Sobaya; exactly one laptop used alone through its built-in keyboard and trackpad; zero external keyboards; zero separate mice; exactly one beer mug; exactly one Image 3 wanted poster; same office and desk throughout; exact chronological prop continuity; mask present and correctly worn in every Sobaya frame; exactly three scripted Sobaya answers and no other speech; no extra people near the subject; no duplicate body, face, limbs, poster, or props; no sheet borders or labels; no fake broadcaster branding; no readable generated text except the one Image 3 poster; no subtitles; no watermark.