A 版画面提示词
subject_definitions:
<Subject 1> is the actual final frame of the previous unit, supplied as image 1. Use it only for the existing characters' opening state and prop positions.
<Subject 2> is Hou Longtao, a twenty-three-year-old Beijing man seated in the left-side window seat of the first-class cabin; follow the appearance and wardrobe given in <Subject 2>.
<Subject 3> is the pair of black-framed glasses he wears; follow the appearance given in <Subject 3>.
<Subject 4> is the seat armrest in the first-class cabin; follow the appearance given in <Subject 4>.
<Subject 5> is the first-class cabin of flight CA984; follow the layout and lighting given in <Subject 5>.
summary:
[reference generation] Inside the daylight cruise of <Subject 5>, <Subject 2> stays seated at the window in complete silence, privately turning something over in his mind. Across three shots the light in his eyes arrives before the smile deepens, his fingertip taps the armrest of <Subject 4> twice, and the hand goes still as his gaze returns to the window.
retention_analysis:
<Subject 1> (appears in [Shot 1]): fully_preserved - the opening frame keeps the seated posture, the face angled toward the camera side, the closed-lipped faint smile, both hand positions on the armrests, the glasses with daylight on the lenses, the empty neighbouring seat, and the window light behind his right shoulder.
<Subject 2> (appears in [Shot 1], [Shot 2], [Shot 3]): fully_preserved - his appearance and wardrobe follow the reference without alteration.
<Subject 3> (appears in [Shot 1], [Shot 3]): fully_preserved - the glasses stay on his face throughout, with window daylight reflecting on the lenses.
<Subject 4> (appears in [Shot 1], [Shot 2], [Shot 3]): fully_preserved - the aisle-side armrest remains under his right hand and becomes the sole subject of the insert shot.
<Subject 5> (appears in [Shot 1], [Shot 2], [Shot 3]): fully_preserved - the cabin layout, seat arrangement, cool white toplight, and bright high-altitude window daylight follow the reference.
detailed_description:
The target video is live-action and cinematic, with cool white cabin toplight, bright high-altitude daylight entering from the window, and a clean, slightly desaturated palette. No subtitles, captions, titles, speech bubbles, or readable on-screen text appear at any point.
[Shot 1] An eye-level medium close shot with a 50mm feel, taken from the aisle side of <Subject 5>. <Subject 2> sits in the left-side window seat on screen right, filling about a third of the frame with a little headroom, while the lenses of <Subject 3> catch the window daylight. The frame opens exactly on the state carried over from <Subject 1>: unchanged seated posture, face angled toward the camera side, lips closed over the faintest trace of a smile, the right hand motionless on the aisle-side armrest of <Subject 4>, the left hand on the far armrest, the lit window behind his right shoulder, and the empty neighbouring seat softly out of focus in the lower left foreground. Focus holds on his eyes and lenses at shallow depth of field, while his right hand and the armrest stay readable along the lower frame edge. For the first 1.3 seconds he keeps this position, draws his gaze back from the daylight to the empty space ahead, and takes one shallow silent breath. From 00:01.300 the camera pushes in with small amplitude at slow speed by a few centimetres and stops at 00:03.900. From 00:01.300 to 00:05.700 he stays silent, his lips completely closed and forming no speech shapes. Between 00:02.200 and 00:03.700 his upper eyelids lift slightly, the light arrives in his eyes first, and only then does the smile deepen behind still-closed lips. At 00:03.400 and again at 00:03.800 his right index fingertip taps the armrest twice, the second tap slightly slower. At 00:05.700 he takes one short shallow breath and his chin settles back.
[Shot 2] At 00:05.900, the shot cuts on that breath to a low close insert of the aisle-side armrest of <Subject 4> beneath his right hand, framed slightly left of centre from the aisle, with the thumb and index finger sharply in focus and the cabin and cool toplight blurred into a bright wash behind. The camera holds a static shot. The hand continues from the two taps of the previous shot, fingertips already at rest and the back of the hand slightly loose, and he remains silent. Between 00:07.400 and 00:09.600 his thumb pad drags very slowly forward across the armrest surface for a short distance and then stops, as if weighing an amount that has been cut away, while the middle and ring fingers stay motionless. By 00:10.300 the thumb has settled and the whole hand is completely still.
[Shot 3] At 00:10.400, the shot cuts to a slightly closer eye-level medium close shot on the same axis as Shot 1, with <Subject 2> again on screen right, his right hand and the armrest kept along the lower frame edge so the stillness stays readable, focus on his eyes, and the window daylight still behind his right shoulder. The camera holds a static shot. Between 00:10.700 and 00:12.600 his breath loosens a notch, the smile stops spreading outward and closes into a pressed line, and the corners of his eyes soften while his lips stay shut. At 00:12.600 his expression settles completely and only the cabin drone remains. Between 00:12.600 and 00:13.000 he turns his head a few degrees toward the window, his gaze returns to the daylight outside, and his upper body leans slightly back into the seat. From 00:13.000 to the end he holds this stable state without further movement: lips closed, the smile held inside a pressed mouth, both hands on the armrests, right fingertips motionless, gaze resting on the light beyond the window, the neighbouring seat still empty under the cool white toplight.
overall_soundscape:
A low ventilation drone and the muffled engine rumble carried through the fuselage continue beneath a very faint high-altitude pressure hiss throughout. One shallow silent breath is audible early, followed by two soft leather taps of a fingertip on the armrest and a fine friction as a thumb pad drags across its surface, then a light brush of fabric against the seat back as he leans back. No voices of any kind, no announcements, and no service sounds are present.
non_diegetic_music:
N/A
B 版独白提示词
subject_definitions:
<Subject 1> is the actual final frame of the previous unit, supplied as image 1. Use it only for the existing characters' opening state and prop positions.
<Subject 2> is Hou Longtao, a twenty-three-year-old Beijing man seated in the left-side window seat of the first-class cabin; follow the appearance and wardrobe given in <Subject 2>.
<Subject 3> is the pair of black-framed glasses he wears; follow the appearance given in <Subject 3>.
<Subject 4> is the seat armrest in the first-class cabin; follow the appearance given in <Subject 4>.
<Subject 5> is the first-class cabin of flight CA984; follow the layout and lighting given in <Subject 5>.
summary:
[reference generation] Inside the daylight cruise of <Subject 5>, <Subject 2> (S1) stays seated at the window while his own off-screen inner monologue accounts for the lottery money. Across three shots the light in his eyes arrives before the smile deepens, his fingertip taps the armrest of <Subject 4> twice, and the hand goes still as his gaze returns to the window.
retention_analysis:
<Subject 1> (appears in [Shot 1]): fully_preserved - the opening frame keeps the seated posture, the face angled toward the camera side, the closed-lipped faint smile, both hand positions on the armrests, the glasses with daylight on the lenses, the empty neighbouring seat, and the window light behind his right shoulder.
<Subject 2> (appears in [Shot 1], [Shot 2], [Shot 3]): fully_preserved - his appearance and wardrobe follow the reference without alteration.
<Subject 3> (appears in [Shot 1], [Shot 3]): fully_preserved - the glasses stay on his face throughout, with window daylight reflecting on the lenses.
<Subject 4> (appears in [Shot 1], [Shot 2], [Shot 3]): fully_preserved - the aisle-side armrest remains under his right hand and becomes the sole subject of the insert shot.
<Subject 5> (appears in [Shot 1], [Shot 2], [Shot 3]): fully_preserved - the cabin layout, seat arrangement, cool white toplight, and bright high-altitude window daylight follow the reference.
detailed_description:
The target video is live-action and cinematic, with cool white cabin toplight, bright high-altitude daylight entering from the window, and a clean, slightly desaturated palette. No subtitles, captions, titles, speech bubbles, or readable on-screen text appear at any point.
[Shot 1] An eye-level medium close shot with a 50mm feel, taken from the aisle side of <Subject 5>. <Subject 2> (S1) sits in the left-side window seat on screen right, filling about a third of the frame with a little headroom, while the lenses of <Subject 3> catch the window daylight. The frame opens exactly on the state carried over from <Subject 1>: unchanged seated posture, face angled toward the camera side, lips closed over the faintest trace of a smile, the right hand motionless on the aisle-side armrest of <Subject 4>, the left hand on the far armrest, the lit window behind his right shoulder, and the empty neighbouring seat softly out of focus in the lower left foreground. Focus holds on his eyes and lenses at shallow depth of field, while his right hand and the armrest stay readable along the lower frame edge. For the first 1.3 seconds he keeps this position, draws his gaze back from the daylight to the empty space ahead, and takes one shallow silent breath. From 00:01.300 the camera pushes in with small amplitude at slow speed by a few centimetres and stops at 00:03.900. At 00:01.300 the twenty-three-year-old Beijing man (S1), in a low-mid register with solid but unforced chest resonance, a clean timbre carrying a thin layer of huskiness, relaxed unhurried tails, a standard Beijing accent with light natural erhua, clear consonants that land crisply without strain, a habitual volume slightly below ordinary conversation, a moderately slow pace with real breath pauses and shallow silent inhales, and sentence endings that settle downward with an offhand finish, never a broadcast or commercial narration voice, never a vocal-fry leading-man register, and never a slick drawl, speaks as a man privately re-checking an account rather than announcing anything, easing his breath open before the first word, weighting the sum slightly and slowing a fraction on those figures as if the fact is still not quite real, and taking only one very short breath at the comma, says in an off-screen voiceover: <d>[Chinese] 在赢了三千五百万美金的六合彩后，虽然在交税之后只剩下九百来万，但也很不错了，</d> while his lips remain completely closed. The monologue continues uninterrupted across both cuts of this unit until 00:12.600. Between 00:02.200 and 00:03.700, as the monologue reaches the prize figure, his upper eyelids lift slightly, the light arrives in his eyes first, and only then does the smile deepen behind still-closed lips. At 00:03.400 and again at 00:03.800 his right index fingertip taps the armrest twice, the second tap slightly slower. At 00:05.700 he takes the short breath at the comma and his chin settles back.
[Shot 2] At 00:05.900, the shot cuts on that breath to a low close insert of the aisle-side armrest of <Subject 4> beneath his right hand, framed slightly left of centre from the aisle, with the thumb and index finger sharply in focus and the cabin and cool toplight blurred into a bright wash behind. The camera holds a static shot. The hand continues from the two taps of the previous shot, fingertips already at rest and the back of the hand slightly loose; the voiceover carries over from the previous shot without interruption. As the monologue reaches the after-tax figure between 00:07.400 and 00:09.600, his thumb pad drags very slowly forward across the armrest surface for a short distance and then stops, as if weighing the amount that has been cut away, while the middle and ring fingers stay motionless. By 00:10.300 the line of the monologue has landed and the whole hand is completely still.
[Shot 3] At 00:10.400, the shot cuts to a slightly closer eye-level medium close shot on the same axis as Shot 1, with <Subject 2> (S1) again on screen right, his right hand and the armrest kept along the lower frame edge so the stillness stays readable, focus on his eyes, and the window daylight still behind his right shoulder. The camera holds a static shot. Between 00:10.700 and 00:12.600 the monologue reaches its final concession and acceptance; his breath loosens a notch, the smile stops spreading outward and closes into a pressed line, and the corners of his eyes soften while his lips stay shut. At 00:12.600 the monologue ends and only the cabin drone remains. Between 00:12.600 and 00:13.000 he turns his head a few degrees toward the window, his gaze returns to the daylight outside, and his upper body leans slightly back into the seat. From 00:13.000 to the end he holds this stable state without further movement: lips closed, the smile held inside a pressed mouth, both hands on the armrests, right fingertips motionless, gaze resting on the light beyond the window, the neighbouring seat still empty under the cool white toplight.
overall_soundscape:
A low ventilation drone and the muffled engine rumble carried through the fuselage continue beneath a very faint high-altitude pressure hiss throughout. One shallow silent breath is audible early, followed by two soft leather taps of a fingertip on the armrest and a fine friction as a thumb pad drags across its surface, then a light brush of fabric against the seat back as he leans back. No other passengers, announcements, or service sounds are present.
non_diegetic_music:
N/A
