看见细节，也听见细节。

subject_definitions:
<Picture 1> is a composition and pacing reference for [Shot 1], showing a dark graphite field, three ordered processing stages, restrained lime accents, wide margins, and a left-to-right connection.
<Subject 1> is the fountain pen and cream paper in <Picture 2>, with an engraved gold nib, an amber-brown barrel, visible paper fibers, and a dark blue ink stroke.

summary:
[reference generation] Generate an 8.000-second two-shot technical product film. Transfer the ordered progression of <Picture 1> into a clean, physical light-and-glass visualization, then cut to <Subject 1> as the concrete material-detail payoff. The final edit will add the exact technical typography separately.

retention_analysis:
<Picture 1> ([Shot 1] composition reference): partially_preserved - preserve three ordered stages, left-to-right flow, graphite background, lime accents, and clear spacing; replace the typography and flat diagram with three unlabeled glass planes.
<Subject 1> (appears in [Shot 2]): fully_preserved - preserve the pen's engraved gold nib, amber-brown barrel, blue ink, cream paper, and their coherent material appearance.

detailed_description:
The target video combines restrained photoreal studio visualization with tactile macro photography. Use black graphite, thin lime highlights, warm gold, and cream paper; readable words and logos will be placed during editing rather than generated inside the image.
[Shot 1] A wide frontal view establishes a dark studio with a matte graphite floor and three thin rectangular glass planes arranged from left to right. The arrangement follows the ordered spacing of <Picture 1>. Each pane stands upright, and all three remain separately identifiable throughout the shot. A narrow lime reflection outlines their upper edges. The floor gives only a faint reflection beneath them. Leave generous negative space above and below the objects. The camera trucks right with small amplitude at slow speed, giving the glass a slight parallax while preserving the left-to-right order. The lighting remains soft and directional, without a bright flare or a surrounding particle cloud.
At the opening, a stable low-detail contour appears inside the left pane: the diagonal outline of a pen touching paper. Between 00:00.500 and 00:01.600, this contour resolves into an orderly group of fine light lines that flows directly toward the middle pane. A short, dry electronic click coincides with the transfer. Between 00:01.600 and 00:03.000, the middle pane gently changes the spacing of these lines while preserving their organization, then passes them toward the right pane in one continuous movement. There is no intermediate full image appearing in the middle pane. The final plane receives the same composition, with a more defined nib edge and a fine dark ink curve. A few selected lime paths illuminate briefly within this final plane as its detail settles. Treat this as a conceptual visualization, with unchanged object placement and no numeric performance counters.
At 00:00.700, an adult female narrator with a low, clear English voice and measured delivery (S1) says in an off-screen voiceover: <d>[English] A smaller draft. A richer world.</d> while no speaking face appears on screen. She finishes by 00:03.600. The camera slows as the final glass plane resolves, and one soft, rounded tone marks the completion of the light path.
[Shot 2] At 00:04.000, the shot cuts to a close macro view of <Subject 1>, matching the diagonal pen orientation inside the final pane. The engraved gold nib rests against cream paper, held by steady fingertips from the upper right. The camera tracks left with small amplitude at slow speed as the nib draws a single shallow blue curve. Wet ink catches the window light, then darkens as it enters the paper fibers. A close paper scratch replaces the electronic clicks. At 00:07.300, the nib lifts a few millimeters and the scratch stops. Hold the completed curve, engraved metal, and gentle shadow through 00:08.000.

overall_soundscape:
A very soft studio air tone underlies sparse electronic clicks and one short rounded transition tone. The second shot exposes the pen's delicate contact scratch and a tiny fingertip movement, ending on quiet room ambience.

non_diegetic_music:
N/A

subject_definitions:
<Subject 1> is the fountain pen in <Picture 1>, with an engraved gold nib, a narrow central slit, a dark breather hole, and a polished translucent amber-brown barrel.
<Subject 2> is the writing surface and hand in <Picture 1>, with textured cream paper, a dark blue curved ink stroke, and fingertips holding the pen from the upper right.

summary:
[reference generation] Generate a 5.1667-second macro product film using <Subject 1> and <Subject 2>. Show one legible physical action, the nib depositing ink into paper fibers, with synchronized fine contact sound and one restrained off-screen spoken line.

retention_analysis:
<Subject 1> (appears in [Shot 1]): fully_preserved - preserve the nib silhouette, slit, engraving, amber-brown barrel, and consistent pen proportions.
<Subject 2> (appears in [Shot 1]): fully_preserved - preserve the paper texture, blue ink, fingertip placement, and upper-right to lower-left pen orientation.

detailed_description:
The target video is a photoreal macro film with warm side lighting, neutral cream highlights, and deep amber reflections. The palette and restrained composition come from the supplied pen reference; the entire 5.1667-second shot remains a single physically continuous take.
[Shot 1] An extreme close view establishes <Subject 1>, the engraved gold nib and amber-brown fountain pen, entering diagonally from the upper right. <Subject 2>, the cream paper and writer's fingertips, fills the remaining frame. The nib point rests just right of center, with enough sharp paper around it to reveal individual raised fibers. The writer's thumb and index finger stay on the barrel above the nib; the grip remains fixed. Soft window light from the upper left creates a narrow gold reflection along one tine and a broad, gentle shadow below the hand. The camera tracks left with small amplitude at slow speed, holding the contact point at a consistent scale rather than orbiting the pen.
During the first half second, the hand settles and the nib touches the paper with a tiny audible tap. From 00:00.500 to 00:02.500, the hand draws one short, shallow curve toward the lower left. The two gold tines flex only slightly under the pressure. A fine meniscus of dark blue ink joins the slit to the paper, and a continuous line forms directly behind the moving tip. The newest ink catches a small wet highlight; ink deposited earlier gradually becomes matte. The nib's movement produces a delicate, close, grainy scratch synchronized to the visible contact, with subtle fluctuations as it crosses the paper fibers. Neither the ink stroke nor the engraved nib changes identity between moments.
At 00:00.900, a calm adult female narrator with a low, clear English voice and an unhurried delivery (S1) says in an off-screen voiceover: <d>[English] Every detail. Every sound.</d> while no speaking face appears on screen. She finishes by 00:03.200, allowing the physical scratch to remain exposed afterward. The narration sits slightly above the contact sound without masking it.
From 00:03.200 to 00:04.500, the curve finishes, pressure releases, and the same nib lifts only a few millimeters. A tiny wet bead remains at the end of the line. The camera slows to a stop and holds the engraved metal, blue ink edge, and soft hand reflection through 00:05.1667. The final pause lets the viewer inspect the material response. Keep the image free of captions, labels, logos, and added interface graphics.

overall_soundscape:
A quiet indoor room tone supports the nib's initial tap, continuous paper scratch, and faint finger movement against the barrel. The scratch stops precisely as the nib lifts, leaving the final moment with soft room ambience.

non_diegetic_music:
N/A

subject_definitions:
<Subject 1> is the adult astronaut in <Picture 1>, with a white helmet, dark inner padding, a clear curved visor, a pale flight suit, short brown hair, and the same recognizable face.
<Subject 2> is the cockpit in <Picture 1>, with a rounded window to the astronaut's right, a dark window surround, and warm reflected exterior light across the visor.

summary:
[reference generation] Generate a 5.1667-second live-action cockpit scene using <Subject 1> and <Subject 2>. A tense reentry moment settles into control while the astronaut delivers one short English line with visible lip synchronization and detailed mechanical ambience.

retention_analysis:
<Subject 1> (appears in [Shot 1]): fully_preserved - preserve the face, helmet proportions, visor shape, padding, and suit design.
<Subject 2> (appears in [Shot 1]): fully_preserved - preserve the window position, cabin geometry, dark framing, and warm exterior illumination.

detailed_description:
The target video uses a realistic close-range camera, natural skin texture, controlled orange illumination, and deep neutral cabin shadows. The full 5.1667-second film stays in one continuous shot with a stable spatial relationship between the astronaut and window.
[Shot 1] A tight medium close-up frames <Subject 1>, the astronaut with the white helmet and transparent visor from the reference, slightly left of center. His shoulders and upper flight suit are visible at the lower edge. <Subject 2>, the cockpit, surrounds him, with its rounded exterior window occupying the right third. The window frame remains a solid, fixed part of the cabin. Warm orange light sweeps slowly across the right side of his face and the visor as the spacecraft moves through the atmosphere. The visor carries a curved reflection of the window, offset from the face without hiding the mouth. The camera pushes in with small amplitude at slow speed and shakes slightly with the cabin vibration. It never changes sides or cuts away.
During the opening second, the astronaut watches an instrument just below the lens. His jaw is set and his brow tightens slightly. A small bead of sweat near the right temple catches the orange light and moves downward a few millimeters. Low structural vibration and close suit breathing establish the physical space; two brief electronic beeps come from the unseen instrument panel ahead of him. The exterior glow brightens gradually rather than flashing across the entire screen. Helmet edges, visor thickness, and reflected window position stay coherent as his head makes one small downward motion.
At 00:01.000, <Subject 1> (S1), speaking in a steady adult male English baritone with a close helmet-microphone texture, looks back toward the instrument and says: <d>[English] Reentry stable. We're coming home.</d> His visible lips articulate the entire line naturally and finish by 00:03.700. Speech is clear over the cabin rumble, with only a restrained communication-radio texture. His voice is calm rather than shouted. The helmet, face, and chest remain in the same framing during the spoken sentence, allowing the synchronization to be assessed.
After speaking, his lips close and he releases one measured breath. The warning beeps stop while the cabin vibration softens slightly. From 00:04.000 onward, his eyes lift toward the warm window and his brow relaxes without turning into a broad smile. Hold that small expression and the slowly moving visor reflection until 00:05.1667. Keep the cockpit free of added captions, readable instrument text, branding, or floating interface elements.

overall_soundscape:
Deep cabin vibration, intermittent metallic rattles, a low exterior roar, and close helmet breathing remain spatially distinct. Two short panel beeps precede the spoken line; vibration decreases slightly after the line finishes without an abrupt silent gap.

non_diegetic_music:
N/A

subject_definitions:
<Subject 1> is the elderly woman in <Picture 1>, with short silver curls, a blue linen shirt, and the same softly lined face, seated on the left.
<Subject 2> is the elderly man in <Picture 1>, with a neat white beard, swept-back white hair, and a cream cardigan, seated on the right.
<Subject 3> is the seaside cafe in <Picture 1>, with a striped awning, a small table, ceramic coffee cups, and a sunlit turquoise harbor behind the couple.

summary:
[reference generation] Generate a 7.2917-second naturalistic two-person conversation using <Subject 1>, <Subject 2>, and <Subject 3>. Preserve the reference couple and cafe while introducing two clear English lines, accurate speaker turns, small facial reactions, and audible harbor ambience.

retention_analysis:
<Subject 1> (appears in [Shot 1]): fully_preserved - preserve the woman's face, silver curls, blue shirt, and left-hand screen position.
<Subject 2> (appears in [Shot 1]): fully_preserved - preserve the man's face, white beard, cream cardigan, and right-hand screen position.
<Subject 3> (appears in [Shot 1]): fully_preserved - preserve the awning, table, coffee cups, harbor, and warm morning light.

detailed_description:
The target video is a warm, naturalistic live-action scene with lifelike skin texture and soft reflected daylight. A single 7.2917-second take gives each speaker room to finish and leaves a final moment for an unforced reaction.
[Shot 1] A close medium two-shot establishes <Subject 1>, the elderly woman with silver curls and a blue linen shirt, on the left, and <Subject 2>, the white-bearded elderly man in a cream cardigan, on the right. They sit at <Subject 3>, the small seaside cafe table, beneath the striped awning. Two ceramic cups rest securely on their saucers near the bottom edge of the frame, with neither cup obscuring a face. The harbor and distant pale buildings remain softly focused behind them. Warm sunlight touches the woman's curls and the man's cardigan, while light reflected from the tabletop softly illuminates their mouths and eyes. A gentle breeze lifts only the loose edge of the awning and a few strands of hair.
The camera holds their eye level and pushes in with small amplitude at slow speed. The initial composition contains both faces, both shoulders, and enough tabletop to maintain their spatial relationship. During the opening half second, the woman turns her eyes toward the man and rests her fingertips beside her saucer. The man notices her look and turns his head a few degrees toward her. Their bodies stay relaxed and seated. A faint cup-on-saucer click is heard as her hand settles, followed by quiet harbor ambience.
At 00:00.700, <Subject 1> (S1), with a warm, lightly textured adult female English voice and a gentle questioning cadence, says: <d>[English] Same table tomorrow?</d> Her visible lips articulate the question and close by 00:02.300. <Subject 2> listens with his lips closed, maintaining eye contact. Leave a natural pause of about half a second, during which a distant gull and a soft wave against the harbor wall remain audible.
At 00:02.900, <Subject 2> (S2), with a low, softly textured adult male English voice and an affectionate, conversational delivery, replies: <d>[English] I wouldn't miss it.</d> He finishes by 00:04.500. The woman listens with her mouth closed and a small smile forming at one corner. Their voices remain distinct and never overlap.
From 00:04.700 onward, <Subject 1> (S1) releases a short, quiet laugh, and <Subject 2> (S2) answers with a soft chuckle. Their smiles settle rather than becoming exaggerated. Hold the shared look, coherent hands, still cups, and gently moving awning through 00:07.2917. The image contains no subtitles, logos, new props, or additional foreground people.

overall_soundscape:
Soft waves touch the harbor wall beneath distant gulls, light cafe room tone, one ceramic click, and a breeze moving the awning. The final brief laughter stays close to the couple, while background sound remains quiet enough to hear each word.

non_diegetic_music:
N/A