Late-Qing China, the Xianheng Tavern. Kong Yiji, a failed scholar in a filthy long gown, is the butt of every joke, until one winter day a weak voice calls from below the counter. Lu Xun's 1919 classic, retold as a hand-drawn-style anime short through the memories of the tavern's twelve-year-old waiter.
Pika H3 Flex fl2va (all-sparse), 768p; Pika H3 Flex ref2va (all-sparse), 768p
Sound
Mandarin, with Chinese and English subtitles
Final film
Subtitles: English and 中文. Switch tracks from the player's captions menu.
Reference pictures
Made first with Qwen Image 2.1 Turbo, then passed to every shot that needs them, so the same faces, costumes and places come back from shot to shot.
The young waiter
Character
Image prompt
Character reference sheet, single character, full body three-quarter view standing, neutral calm expression, hands visible, plain light grey-beige paper background, even soft light. the young waiter, a thin twelve-year-old boy drawn in 2D anime style: round boyish face, large dark attentive eyes, shaved Qing forehead with a short black queue, a plain faded indigo cotton short jacket and trousers, a stained off-white cloth apron tied at the waist, black cloth shoes. Chinese 2D hand-drawn anime film style, clean ink-like linework, soft cel shading, hand-painted watercolor texture, restrained palette of ink black, faded indigo, grey-white and weathered wood. One person only, no props on the ground, no text, no characters, no signature, no watermark.
Tavern owner
Character
Image prompt
Character reference sheet, single character, full body three-quarter view standing, neutral calm expression, hands visible, plain light grey-beige paper background, even soft light. the tavern owner, a stout man in his fifties drawn in 2D anime style: round fleshy face with heavy jowls, small sharp narrow eyes, a thin drooping black moustache, a black round skullcap, a long black queue, a dark brown cotton long gown under a black sleeveless waistcoat, often holding an abacus. Chinese 2D hand-drawn anime film style, clean ink-like linework, soft cel shading, hand-painted watercolor texture, restrained palette of ink black, faded indigo, grey-white and weathered wood. One person only, no props on the ground, no text, no characters, no signature, no watermark.
Kong Yiji
Character
Image prompt
Character reference sheet, single character, full body three-quarter view standing, neutral calm expression, hands visible, plain light grey-beige paper background, even soft light. Kong Yiji, a tall gaunt scholar of about fifty, drawn in 2D anime style: long pale bluish-white face lined with deep wrinkles and two small healed scars on the left cheek and forehead, a scruffy unkempt grey-white beard, tired drooping eyes, the shaved Qing forehead with a thin greasy grey queue down his back; he wears a long grey-white scholar's gown (changshan) that is filthy, stained and torn at the hem and sleeves, never washed or mended, with cloth shoes worn through. Chinese 2D hand-drawn anime film style, clean ink-like linework, soft cel shading, hand-painted watercolor texture, restrained palette of ink black, faded indigo, grey-white and weathered wood. One person only, no props on the ground, no text, no characters, no signature, no watermark.
Kong Yiji, after the beating
Character
Image prompt
Edit this character into his final appearance, same person, same face identity, same grey-white beard and grey queue, same 2D anime style and plain light grey-beige background. His face becomes darker, hollow-cheeked and much thinner. Replace the long grey-white gown with a ragged, patched, dirty dark-grey lined short jacket and worn trousers. He now sits on the ground on a flattened woven cattail-straw mat bag, his broken legs folded limply beneath him, the mat hung from his shoulders by a straw rope; both hands rest on the ground beside the mat, caked with mud. Full body visible, single person, no text, no watermark.
Xianheng Tavern, street side
Location
Image prompt
Empty environment background, wide establishing view from the grey flagstone street looking at the shopfront, afternoon in autumn. the empty Xianheng tavern in the water town of Luzhen, drawn as a 2D anime watercolor background: a large L-shaped wooden counter opens directly onto the grey flagstone street; behind the counter sit a low clay stove with a copper pot of hot water for warming wine, rows of large brown earthenware wine jars with red cloth-and-clay seals, and small tin wine pots; a dark lacquered wooden shopfront with removable plank doors, black-tiled eaves above, whitewashed walls; through a doorway at the back is a dim inner room with square tables; a small dark chalk slate hangs on a pillar by the counter covered only with faint smudges; a raised wooden threshold below the counter's street side. Warm amber glow and faint steam from the stove behind the counter, cool grey-blue daylight in the street. No people at all, no animals, no readable text, no signboard characters, no couplets, no watermark. Chinese 2D hand-drawn anime film style, clean ink-like linework, soft cel shading, hand-painted watercolor texture, restrained palette of ink black, faded indigo, grey-white and weathered wood. Anime film background art, 16:9.
Tavern interior
Location
Image prompt
Empty environment background, interior view from behind a tavern counter looking outward toward the street, late-Qing Shaoxing water town, autumn afternoon. Foreground: a long worn wooden countertop runs horizontally across the lower third of the frame; on the inner side at left a low whitewashed clay stove with a copper pot of steaming hot water and several small tin wine pots, at right three large dark-brown earthenware wine jars with red cloth-and-clay seals. Beyond the counter: the open shopfront with dark lacquered wooden door posts and black-tiled eaves edge at the top, a raised stone threshold, then the grey flagstone street and, across it, whitewashed walls with black-tiled roofs and a dark wooden door, in cool grey-blue daylight. Warm amber stove glow inside, cool light outside. Chinese 2D hand-drawn anime film background art, hand-painted watercolor texture, clean linework. No people, no animals, no readable text, no signboards, no watermark. 16:9.
Title card
Title card
Picture prompt
Cinematic anime film poster, 16:9, using the tavern in Picture 1 and the scholar Kong Yiji from Picture 2 (same face, grey-white beard, grey queue, filthy torn long grey-white scholar's gown). Late autumn dusk on the flagstone street. Kong Yiji stands alone at the street side of the tavern counter, tall and gaunt, seen from a low three-quarter angle on the left third of the frame, one long hand resting on the counter beside a small dish of aniseed beans and a neat row of copper coins, his face half lit by the warm amber glow and rising steam from the copper pot behind the counter, the rest of the street sinking into cold grey-blue mist with a few fallen leaves blowing past. Strong color dramaturgy: a single warm pool of amber light against cold blue-grey. On the right side of the frame, large elegant vertical Chinese brush calligraphy title "孔乙己" in black ink with a small red seal stamp beside it, and below it in small neat characters "鲁迅 原著". The title text is spelled exactly 孔乙己. Chinese 2D hand-drawn anime film style, clean ink linework, soft cel shading, hand-painted watercolor background with paper grain, dramatic, melancholy, high detail. No other text, no watermark.
Shot by shot
Each clip below is the raw generation that went into the cut, before music, mixing and captions. Its sound (effects and voice) came out of the same generation.
Title opening
5 s · First frame to video · Pika H3 Flex fl2va (all-sparse), 768p
<Picture 1> A ceremonial cinematic film opening title shot in Chinese 2D hand-drawn anime style, starting exactly from <Picture 1>: the late-Qing tavern at dusk, the tall gaunt scholar Kong Yiji in his filthy grey-white long gown standing at the counter on the left, warm amber steam rising from the copper pot, cold grey-blue mist in the street, and the black brush-calligraphy title with its red seal on the right.
[Shot 1] One continuous 5-second shot with no cut. The camera performs a very slow, smooth, stately push-in of a few percent, like the opening frame of a feature film. Thick steam rises and curls from the copper pot and drifts across the warm light; the amber glow breathes and flickers gently; a few dry autumn leaves tumble slowly across the street from left to right; thin mist rolls along the flagstones. Kong Yiji stays in place, his torn gown hem stirring in the breeze; in the second half he slowly lifts his eyes and gives a faint, bitter half-smile, his hand still resting beside the coins. The calligraphy title, the red seal and the small characters on the right remain perfectly still, sharp, unchanged and fully legible for the entire shot, never morphing, never moving with the scene. No other text appears. Clean ink linework and watercolor texture stay stable throughout.
overall_soundscape: A deep, resonant single temple-bell or gong strike at the very start, slowly decaying; soft wind, rustling leaves and a faint hiss of steam beneath it. No speech, no voices.
non_diegetic_music: A low sustained string drone rises under the bell, joined by a single solemn guqin phrase that resolves at the end, majestic and melancholy, like a feature film's opening fanfare in a Chinese idiom.
Shot 1
15 s · Reference to video · Pika H3 Flex ref2va (all-sparse), 768p
subject_definitions:
<Subject 1> is the empty Xianheng tavern in the water town of Luzhen, drawn as a 2D anime watercolor background: a large L-shaped wooden counter opens directly onto the grey flagstone street; behind the counter sit a low clay stove with a copper pot of hot water for warming wine, rows of large brown earthenware wine jars with red cloth-and-clay seals, and small tin wine pots; a dark lacquered wooden shopfront with removable plank doors, black-tiled eaves above, whitewashed walls; through a doorway at the back is a dim inner room with square tables; a small dark chalk slate hangs on a pillar by the counter covered only with faint smudges; a raised wooden threshold below the counter's street side, as shown in <Picture 1>. This image supplies the set layout, materials and colors, not a locked camera angle; it is empty and contains no people.
summary:
[reference generation] A continuous 15.0833-second 2D anime shot from an adaptation of Lu Xun's story set in a late-Qing village tavern. Several laborers in patched short jackets stand at the street side of the counter, drinking from coarse bowls; one tips his bowl back, another leans an elbow on the wood.
retention_analysis:
<Subject 1> (appears in [Shot 1]): fully_preserved - Keep the L-shaped counter facing the street, the stove and copper pot behind it, the earthenware jars, the chalk slate on the pillar, the dark shopfront and black-tiled eaves in the same spatial arrangement. Lighting follows the time of day described in this shot.
detailed_description:
Chinese 2D hand-drawn anime film style: clean confident ink-like linework, soft cel shading with two-tone shadows, hand-painted watercolor backgrounds with visible paper grain, late-Qing Shaoxing water-town atmosphere. Restrained palette of ink black, faded indigo, grey-white plaster and weathered wood, with warm amber firelight and wine steam inside the counter against cool grey-blue daylight in the street. Expressive but grounded anime acting, consistent character model sheets, stable line weight. No readable text, no Chinese characters, no subtitles, no captions, no logos, no watermarks anywhere in the frame.
Autumn afternoon: warm amber stove glow and wine steam behind the counter, cool grey-blue daylight on the street flagstones.
[Shot 1] One continuous 15.0833-second shot with no internal cut. Wide establishing view from the flagstone street, matching the reference: the counter opening onto the street, black-tiled eaves above. A slow, steady push in toward the counter over the whole shot.
The location is <Subject 1>; keep its counter, stove, jars and shopfront in their established places. Referenced characters on screen: none. Each referenced character appears exactly once, never duplicated; extras, if any, are simpler background figures who do not resemble the referenced characters.
Several laborers in patched short jackets stand at the street side of the counter, drinking from coarse bowls; one tips his bowl back, another leans an elbow on the wood. Steam curls from the copper pot behind the counter. Nobody faces the camera and nobody speaks; their lips stay closed. From second one the unseen narrator speaks over the scene. In the last seconds the push settles on the warm steam above the counter.
The unseen narrator, in the calm, low, slightly weary voice of a man in his forties, speaking Mandarin Chinese with measured, reflective pacing, remembering his childhood (S1), says in an off-screen voiceover, slowly and evenly, a man remembering, from second one to second thirteen: <d>[Chinese] 鲁镇的酒店，格局和别处不同：当街一个曲尺形的大柜台，柜里预备着热水，可以随时温酒。</d> All visible characters keep their lips closed during the narration.
Animation stays smooth and on-model: stable faces, consistent line weight, cloth folds and queues moving with the body, steam and firelight drifting gently. Keep anatomy correct, with no extra limbs or duplicated props. End on the steam rising over the counter; the next shot cuts behind the counter to the boy at the stove.
Keep a short breath of silence after the final spoken words; no unfinished syllable at the end. Mouths move only for the written Chinese lines; no other intelligible speech, no ad-libs. Absolutely no subtitles, captions, on-screen text, Chinese characters, signboards with writing, logos or watermarks.
overall_soundscape:
Low street murmur, bowls set down on wood, a ladle clinking in the hot-water pot, a distant wooden boat oar. Natural acoustic space of a small street-front tavern; no words beyond the written lines.
non_diegetic_music:
A sparse, soft solo guqin plucking a slow melancholy pentatonic phrase under the narration, very quiet.
Shot 2
12 s · Reference to video · Pika H3 Flex ref2va (all-sparse), 768p
subject_definitions:
<Subject 1> is the young waiter, a thin twelve-year-old boy drawn in 2D anime style: round boyish face, large dark attentive eyes, shaved Qing forehead with a short black queue, a plain faded indigo cotton short jacket and trousers, a stained off-white cloth apron tied at the waist, black cloth shoes, as shown in <Picture 1>. This image supplies identity and costume only; its plain portrait backdrop is not the story location.
<Subject 2> is the tavern owner, a stout man in his fifties drawn in 2D anime style: round fleshy face with heavy jowls, small sharp narrow eyes, a thin drooping black moustache, a black round skullcap, a long black queue, a dark brown cotton long gown under a black sleeveless waistcoat, often holding an abacus, as shown in <Picture 2>. This image supplies identity and costume only; its plain portrait backdrop is not the story location.
<Subject 3> is the interior of the same tavern seen from behind the counter: the worn wooden countertop across the foreground, the clay stove with copper hot-water pot and tin wine pots, dark earthenware wine jars with red seals, the open shopfront and the grey flagstone street with whitewashed walls and black-tiled roofs beyond, as shown in <Picture 3>. This image supplies the set layout, materials and colors, not a locked camera angle; it is empty and contains no people.
summary:
[reference generation] A continuous 12.2500-second 2D anime shot from an adaptation of Lu Xun's story set in a late-Qing village tavern. The boy lowers a small tin wine pot into the copper pot of hot water with both hands.
retention_analysis:
<Subject 1> (appears in [Shot 1]): fully_preserved - Keep his twelve-year-old face and size, short black queue, faded indigo short jacket and the off-white apron. Keep the 2D anime character design while performing the directed actions.
<Subject 2> (appears in [Shot 1]): fully_preserved - Keep his stout build, round fleshy face, small narrow eyes, thin moustache, black skullcap and the dark brown gown with black waistcoat. Keep the 2D anime character design while performing the directed actions.
<Subject 3> (appears in [Shot 1]): fully_preserved - Keep the L-shaped counter facing the street, the stove and copper pot behind it, the earthenware jars, the chalk slate on the pillar, the dark shopfront and black-tiled eaves in the same spatial arrangement. Lighting follows the time of day described in this shot.
detailed_description:
Chinese 2D hand-drawn anime film style: clean confident ink-like linework, soft cel shading with two-tone shadows, hand-painted watercolor backgrounds with visible paper grain, late-Qing Shaoxing water-town atmosphere. Restrained palette of ink black, faded indigo, grey-white plaster and weathered wood, with warm amber firelight and wine steam inside the counter against cool grey-blue daylight in the street. Expressive but grounded anime acting, consistent character model sheets, stable line weight. No readable text, no Chinese characters, no subtitles, no captions, no logos, no watermarks anywhere in the frame.
Autumn afternoon: warm amber stove glow and wine steam behind the counter, cool grey-blue daylight on the street flagstones.
[Shot 1] One continuous 12.2500-second shot with no internal cut. Medium shot from inside behind the counter, matching the interior reference: the boy on the left by the clay stove, the stout owner standing a step behind him on the right with folded arms, the bright street visible beyond the counter.
The location is <Subject 3>; keep its counter, stove, jars and shopfront in their established places. Referenced characters on screen: <Subject 1>, <Subject 2>. Each referenced character appears exactly once, never duplicated; extras, if any, are simpler background figures who do not resemble the referenced characters.
The boy lowers a small tin wine pot into the copper pot of hot water with both hands. The owner leans over his shoulder, narrow eyes fixed on the pot, suspicious. During the first six seconds the unseen narrator speaks while both characters keep their lips closed. At about seven seconds the owner points a thick finger at the wine pot and speaks. The boy flinches slightly and keeps his eyes down.
The unseen narrator, in the calm, low, slightly weary voice of a man in his forties, speaking Mandarin Chinese with measured, reflective pacing, remembering his childhood (S1), says in an off-screen voiceover, reflective, from second zero to second five: <d>[Chinese] 我从十二岁起，便在镇口的咸亨酒店里当伙计。</d> All visible characters keep their lips closed during the narration. <Subject 2>, in the thick, deep, gruff voice of a stout man in his fifties, speaking Mandarin Chinese bluntly with a mocking edge (S4), says gruffly and suspiciously, from second seven to second ten: <d>[Chinese] 酒里羼水了没有？看仔细了！</d> Only this character's lips move for this line; listeners stay silent.
Animation stays smooth and on-model: stable faces, consistent line weight, cloth folds and queues moving with the body, steam and firelight drifting gently. Keep anatomy correct, with no extra limbs or duplicated props. End on the boy bent over the pot, the owner still watching; next shot holds on the boy alone.
Keep a short breath of silence after the final spoken words; no unfinished syllable at the end. Mouths move only for the written Chinese lines; no other intelligible speech, no ad-libs. Absolutely no subtitles, captions, on-screen text, Chinese characters, signboards with writing, logos or watermarks.
overall_soundscape:
Hot water bubbling, a tin pot clinking against copper, fire crackling in the stove. Natural acoustic space of a small street-front tavern; no words beyond the written lines.
non_diegetic_music:
N/A
Shot 3
12 s · Reference to video · Pika H3 Flex ref2va (all-sparse), 768p
subject_definitions:
<Subject 1> is the young waiter, a thin twelve-year-old boy drawn in 2D anime style: round boyish face, large dark attentive eyes, shaved Qing forehead with a short black queue, a plain faded indigo cotton short jacket and trousers, a stained off-white cloth apron tied at the waist, black cloth shoes, as shown in <Picture 1>. This image supplies identity and costume only; its plain portrait backdrop is not the story location.
<Subject 2> is the interior of the same tavern seen from behind the counter: the worn wooden countertop across the foreground, the clay stove with copper hot-water pot and tin wine pots, dark earthenware wine jars with red seals, the open shopfront and the grey flagstone street with whitewashed walls and black-tiled roofs beyond, as shown in <Picture 2>. This image supplies the set layout, materials and colors, not a locked camera angle; it is empty and contains no people.
summary:
[reference generation] A continuous 12.2500-second 2D anime shot from an adaptation of Lu Xun's story set in a late-Qing village tavern. The boy stands idle, chin resting on one hand on the countertop, eyes drifting to the street, then sighs quietly through his nose.
retention_analysis:
<Subject 1> (appears in [Shot 1]): fully_preserved - Keep his twelve-year-old face and size, short black queue, faded indigo short jacket and the off-white apron. Keep the 2D anime character design while performing the directed actions.
<Subject 2> (appears in [Shot 1]): fully_preserved - Keep the L-shaped counter facing the street, the stove and copper pot behind it, the earthenware jars, the chalk slate on the pillar, the dark shopfront and black-tiled eaves in the same spatial arrangement. Lighting follows the time of day described in this shot.
detailed_description:
Chinese 2D hand-drawn anime film style: clean confident ink-like linework, soft cel shading with two-tone shadows, hand-painted watercolor backgrounds with visible paper grain, late-Qing Shaoxing water-town atmosphere. Restrained palette of ink black, faded indigo, grey-white plaster and weathered wood, with warm amber firelight and wine steam inside the counter against cool grey-blue daylight in the street. Expressive but grounded anime acting, consistent character model sheets, stable line weight. No readable text, no Chinese characters, no subtitles, no captions, no logos, no watermarks anywhere in the frame.
Autumn afternoon: warm amber stove glow and wine steam behind the counter, cool grey-blue daylight on the street flagstones.
[Shot 1] One continuous 12.2500-second shot with no internal cut. Close medium shot of the boy alone behind the counter beside the stove, warm firelight on his cheek, the street softly out of focus behind the counter.
The location is <Subject 2>; keep its counter, stove, jars and shopfront in their established places. Referenced characters on screen: <Subject 1>. Each referenced character appears exactly once, never duplicated; extras, if any, are simpler background figures who do not resemble the referenced characters.
The boy stands idle, chin resting on one hand on the countertop, eyes drifting to the street, then sighs quietly through his nose. He absently turns a tin wine pot a quarter turn in the hot water. His mouth stays closed during the narration.
The unseen narrator, in the calm, low, slightly weary voice of a man in his forties, speaking Mandarin Chinese with measured, reflective pacing, remembering his childhood (S1), says in an off-screen voiceover, quietly, a little weary, from second one to second ten: <d>[Chinese] 掌柜是一副凶脸孔，主顾也没有好声气，教人活泼不得。</d> All visible characters keep their lips closed during the narration.
Animation stays smooth and on-model: stable faces, consistent line weight, cloth folds and queues moving with the body, steam and firelight drifting gently. Keep anatomy correct, with no extra limbs or duplicated props. End with the boy lifting his head toward the street; next shot shows what he sees.
Keep a short breath of silence after the final spoken words; no unfinished syllable at the end. Mouths move only for the written Chinese lines; no other intelligible speech, no ad-libs. Absolutely no subtitles, captions, on-screen text, Chinese characters, signboards with writing, logos or watermarks.
overall_soundscape:
Fire crackle, bubbling water, muffled chatter of drinkers at the counter. Natural acoustic space of a small street-front tavern; no words beyond the written lines.
non_diegetic_music:
N/A
Shot 4
12 s · Reference to video · Pika H3 Flex ref2va (all-sparse), 768p
subject_definitions:
<Subject 1> is Kong Yiji, a tall gaunt scholar of about fifty, drawn in 2D anime style: long pale bluish-white face lined with deep wrinkles and two small healed scars on the left cheek and forehead, a scruffy unkempt grey-white beard, tired drooping eyes, the shaved Qing forehead with a thin greasy grey queue down his back; he wears a long grey-white scholar's gown (changshan) that is filthy, stained and torn at the hem and sleeves, never washed or mended, with cloth shoes worn through, as shown in <Picture 1>. This image supplies identity and costume only; its plain portrait backdrop is not the story location.
<Subject 2> is the empty Xianheng tavern in the water town of Luzhen, drawn as a 2D anime watercolor background: a large L-shaped wooden counter opens directly onto the grey flagstone street; behind the counter sit a low clay stove with a copper pot of hot water for warming wine, rows of large brown earthenware wine jars with red cloth-and-clay seals, and small tin wine pots; a dark lacquered wooden shopfront with removable plank doors, black-tiled eaves above, whitewashed walls; through a doorway at the back is a dim inner room with square tables; a small dark chalk slate hangs on a pillar by the counter covered only with faint smudges; a raised wooden threshold below the counter's street side, as shown in <Picture 2>. This image supplies the set layout, materials and colors, not a locked camera angle; it is empty and contains no people.
summary:
[reference generation] A continuous 12.2500-second 2D anime shot from an adaptation of Lu Xun's story set in a late-Qing village tavern. Far down the street, a tall thin figure in a long grey-white scholar's gown walks unhurriedly toward the tavern, the torn hem swaying, his gown standing out against the short jackets of the people around him.
retention_analysis:
<Subject 1> (appears in [Shot 1]): fully_preserved - Keep his tall thin build, long pale wrinkled face with the small scars, scruffy grey-white beard, grey queue and the filthy torn long grey-white scholar's gown. Keep the 2D anime character design while performing the directed actions.
<Subject 2> (appears in [Shot 1]): fully_preserved - Keep the L-shaped counter facing the street, the stove and copper pot behind it, the earthenware jars, the chalk slate on the pillar, the dark shopfront and black-tiled eaves in the same spatial arrangement. Lighting follows the time of day described in this shot.
detailed_description:
Chinese 2D hand-drawn anime film style: clean confident ink-like linework, soft cel shading with two-tone shadows, hand-painted watercolor backgrounds with visible paper grain, late-Qing Shaoxing water-town atmosphere. Restrained palette of ink black, faded indigo, grey-white plaster and weathered wood, with warm amber firelight and wine steam inside the counter against cool grey-blue daylight in the street. Expressive but grounded anime acting, consistent character model sheets, stable line weight. No readable text, no Chinese characters, no subtitles, no captions, no logos, no watermarks anywhere in the frame.
Autumn afternoon: warm amber stove glow and wine steam behind the counter, cool grey-blue daylight on the street flagstones.
[Shot 1] One continuous 12.2500-second shot with no internal cut. Wide view down the flagstone street from beside the counter, the shopfront at the right edge of the frame; a slow, slight push toward the approaching figure.
The location is <Subject 2>; keep its counter, stove, jars and shopfront in their established places. Referenced characters on screen: <Subject 1>. Each referenced character appears exactly once, never duplicated; extras, if any, are simpler background figures who do not resemble the referenced characters.
Far down the street, a tall thin figure in a long grey-white scholar's gown walks unhurriedly toward the tavern, the torn hem swaying, his gown standing out against the short jackets of the people around him. Halfway through he lifts his chin and looks toward the counter. His lips stay closed; only the unseen narrator speaks.
The unseen narrator, in the calm, low, slightly weary voice of a man in his forties, speaking Mandarin Chinese with measured, reflective pacing, remembering his childhood (S1), says in an off-screen voiceover, with a faint hint of fondness, from second two to second ten: <d>[Chinese] 只有孔乙己到店，才可以笑几声，所以至今还记得。</d> All visible characters keep their lips closed during the narration.
Animation stays smooth and on-model: stable faces, consistent line weight, cloth folds and queues moving with the body, steam and firelight drifting gently. Keep anatomy correct, with no extra limbs or duplicated props. End as he nears the counter; next shot he stands at it.
Keep a short breath of silence after the final spoken words; no unfinished syllable at the end. Mouths move only for the written Chinese lines; no other intelligible speech, no ad-libs. Absolutely no subtitles, captions, on-screen text, Chinese characters, signboards with writing, logos or watermarks.
overall_soundscape:
Footsteps on flagstones, distant hawker calls without words, a dog barking once far away. Natural acoustic space of a small street-front tavern; no words beyond the written lines.
non_diegetic_music:
N/A
Shot 5
15 s · Reference to video · Pika H3 Flex ref2va (all-sparse), 768p
subject_definitions:
<Subject 1> is Kong Yiji, a tall gaunt scholar of about fifty, drawn in 2D anime style: long pale bluish-white face lined with deep wrinkles and two small healed scars on the left cheek and forehead, a scruffy unkempt grey-white beard, tired drooping eyes, the shaved Qing forehead with a thin greasy grey queue down his back; he wears a long grey-white scholar's gown (changshan) that is filthy, stained and torn at the hem and sleeves, never washed or mended, with cloth shoes worn through, as shown in <Picture 1>. This image supplies identity and costume only; its plain portrait backdrop is not the story location.
<Subject 2> is the empty Xianheng tavern in the water town of Luzhen, drawn as a 2D anime watercolor background: a large L-shaped wooden counter opens directly onto the grey flagstone street; behind the counter sit a low clay stove with a copper pot of hot water for warming wine, rows of large brown earthenware wine jars with red cloth-and-clay seals, and small tin wine pots; a dark lacquered wooden shopfront with removable plank doors, black-tiled eaves above, whitewashed walls; through a doorway at the back is a dim inner room with square tables; a small dark chalk slate hangs on a pillar by the counter covered only with faint smudges; a raised wooden threshold below the counter's street side, as shown in <Picture 2>. This image supplies the set layout, materials and colors, not a locked camera angle; it is empty and contains no people.
summary:
[reference generation] A continuous 15.0833-second 2D anime shot from an adaptation of Lu Xun's story set in a late-Qing village tavern. Kong Yiji straightens his filthy gown, reaches into his sleeve and, one coin at a time, lays out nine large copper coins in a neat row on the counter, tapping each down with a long finger.
retention_analysis:
<Subject 1> (appears in [Shot 1]): fully_preserved - Keep his tall thin build, long pale wrinkled face with the small scars, scruffy grey-white beard, grey queue and the filthy torn long grey-white scholar's gown. Keep the 2D anime character design while performing the directed actions.
<Subject 2> (appears in [Shot 1]): fully_preserved - Keep the L-shaped counter facing the street, the stove and copper pot behind it, the earthenware jars, the chalk slate on the pillar, the dark shopfront and black-tiled eaves in the same spatial arrangement. Lighting follows the time of day described in this shot.
detailed_description:
Chinese 2D hand-drawn anime film style: clean confident ink-like linework, soft cel shading with two-tone shadows, hand-painted watercolor backgrounds with visible paper grain, late-Qing Shaoxing water-town atmosphere. Restrained palette of ink black, faded indigo, grey-white plaster and weathered wood, with warm amber firelight and wine steam inside the counter against cool grey-blue daylight in the street. Expressive but grounded anime acting, consistent character model sheets, stable line weight. No readable text, no Chinese characters, no subtitles, no captions, no logos, no watermarks anywhere in the frame.
Autumn afternoon: warm amber stove glow and wine steam behind the counter, cool grey-blue daylight on the street flagstones.
[Shot 1] One continuous 15.0833-second shot with no internal cut. Medium shot from the street side: Kong Yiji standing at the counter in the center, three or four other laborers in short jackets lean on the counter behind him, drawn as simpler background figures on the left, steam behind the counter.
The location is <Subject 2>; keep its counter, stove, jars and shopfront in their established places. Referenced characters on screen: <Subject 1>. Each referenced character appears exactly once, never duplicated; extras, if any, are simpler background figures who do not resemble the referenced characters.
Kong Yiji straightens his filthy gown, reaches into his sleeve and, one coin at a time, lays out nine large copper coins in a neat row on the counter, tapping each down with a long finger. He keeps his lips closed while the unseen narrator speaks from second zero to second six. At about eight seconds he lifts his chin and speaks toward the inside of the counter. The laborers glance at him and grin.
The unseen narrator, in the calm, low, slightly weary voice of a man in his forties, speaking Mandarin Chinese with measured, reflective pacing, remembering his childhood (S1), says in an off-screen voiceover, drily, from second zero to second six: <d>[Chinese] 孔乙己是站着喝酒而穿长衫的唯一的人。</d> All visible characters keep their lips closed during the narration. <Subject 1>, in the thin, slightly hoarse, nasal voice of a man around fifty, speaking Mandarin Chinese in a slow pedantic drawl with old-fashioned bookish cadence, rising and stammering when embarrassed (S2), says with grave scholarly dignity, slowly, from second eight to second twelve: <d>[Chinese] 温两碗酒，要一碟茴香豆。</d> Only this character's lips move for this line; listeners stay silent.
Animation stays smooth and on-model: stable faces, consistent line weight, cloth folds and queues moving with the body, steam and firelight drifting gently. Keep anatomy correct, with no extra limbs or duplicated props. End on the row of nine coins and his upturned chin; next shot the laborers start teasing.
Keep a short breath of silence after the final spoken words; no unfinished syllable at the end. Mouths move only for the written Chinese lines; no other intelligible speech, no ad-libs. Absolutely no subtitles, captions, on-screen text, Chinese characters, signboards with writing, logos or watermarks.
overall_soundscape:
Coins clicking one by one on wood, low chuckles, steam hissing. Natural acoustic space of a small street-front tavern; no words beyond the written lines.
non_diegetic_music:
N/A
Shot 6
12 s · Reference to video · Pika H3 Flex ref2va (all-sparse), 768p
subject_definitions:
<Subject 1> is Kong Yiji, a tall gaunt scholar of about fifty, drawn in 2D anime style: long pale bluish-white face lined with deep wrinkles and two small healed scars on the left cheek and forehead, a scruffy unkempt grey-white beard, tired drooping eyes, the shaved Qing forehead with a thin greasy grey queue down his back; he wears a long grey-white scholar's gown (changshan) that is filthy, stained and torn at the hem and sleeves, never washed or mended, with cloth shoes worn through, as shown in <Picture 1>. This image supplies identity and costume only; its plain portrait backdrop is not the story location.
<Subject 2> is the empty Xianheng tavern in the water town of Luzhen, drawn as a 2D anime watercolor background: a large L-shaped wooden counter opens directly onto the grey flagstone street; behind the counter sit a low clay stove with a copper pot of hot water for warming wine, rows of large brown earthenware wine jars with red cloth-and-clay seals, and small tin wine pots; a dark lacquered wooden shopfront with removable plank doors, black-tiled eaves above, whitewashed walls; through a doorway at the back is a dim inner room with square tables; a small dark chalk slate hangs on a pillar by the counter covered only with faint smudges; a raised wooden threshold below the counter's street side, as shown in <Picture 2>. This image supplies the set layout, materials and colors, not a locked camera angle; it is empty and contains no people.
summary:
[reference generation] A continuous 12.2500-second 2D anime shot from an adaptation of Lu Xun's story set in a late-Qing village tavern. The burly laborer points at a scar on Kong Yiji's forehead and calls out with a grin.
retention_analysis:
<Subject 1> (appears in [Shot 1]): fully_preserved - Keep his tall thin build, long pale wrinkled face with the small scars, scruffy grey-white beard, grey queue and the filthy torn long grey-white scholar's gown. Keep the 2D anime character design while performing the directed actions.
<Subject 2> (appears in [Shot 1]): fully_preserved - Keep the L-shaped counter facing the street, the stove and copper pot behind it, the earthenware jars, the chalk slate on the pillar, the dark shopfront and black-tiled eaves in the same spatial arrangement. Lighting follows the time of day described in this shot.
detailed_description:
Chinese 2D hand-drawn anime film style: clean confident ink-like linework, soft cel shading with two-tone shadows, hand-painted watercolor backgrounds with visible paper grain, late-Qing Shaoxing water-town atmosphere. Restrained palette of ink black, faded indigo, grey-white plaster and weathered wood, with warm amber firelight and wine steam inside the counter against cool grey-blue daylight in the street. Expressive but grounded anime acting, consistent character model sheets, stable line weight. No readable text, no Chinese characters, no subtitles, no captions, no logos, no watermarks anywhere in the frame.
Autumn afternoon: warm amber stove glow and wine steam behind the counter, cool grey-blue daylight on the street flagstones.
[Shot 1] One continuous 12.2500-second shot with no internal cut. Medium two-shot from the street: Kong Yiji at the counter on the right holding a bowl of wine, a burly laborer in his thirties with a sun-browned face, wearing a patched black cotton short jacket on the left leaning toward him, other drinkers behind.
The location is <Subject 2>; keep its counter, stove, jars and shopfront in their established places. Referenced characters on screen: <Subject 1>. Each referenced character appears exactly once, never duplicated; extras, if any, are simpler background figures who do not resemble the referenced characters.
The burly laborer points at a scar on Kong Yiji's forehead and calls out with a grin. Kong Yiji ignores him, eyes forward, and sips from his bowl. At about six seconds a second laborer, wiry with a thin moustache, leans in from behind the first and adds his jab. The others laugh.
A burly laborer in his thirties with a sun-browned face, wearing a patched black cotton short jacket, in the loud, rough, cheerful voice of a laborer in his thirties, speaking Mandarin Chinese with teasing relish (S5), says loudly and teasingly, from second one to second four: <d>[Chinese] 孔乙己，你脸上又添上新伤疤了！</d> Only this speaker's lips move for this line. A wiry laborer in his forties with a thin moustache, wearing a faded brown short jacket and a straw-rope belt, in the raspy, mid-pitched voice of a wiry laborer in his forties, speaking Mandarin Chinese flatly and offhandedly (S6), says slyly, from second six to second nine: <d>[Chinese] 你一定又偷了人家的东西了！</d> Only this speaker's lips move for this line.
Animation stays smooth and on-model: stable faces, consistent line weight, cloth folds and queues moving with the body, steam and firelight drifting gently. Keep anatomy correct, with no extra limbs or duplicated props. End on Kong Yiji lowering the bowl, eyes widening; next shot he protests.
Keep a short breath of silence after the final spoken words; no unfinished syllable at the end. Mouths move only for the written Chinese lines; no other intelligible speech, no ad-libs. Absolutely no subtitles, captions, on-screen text, Chinese characters, signboards with writing, logos or watermarks.
overall_soundscape:
Bowls clink, rough laughter rises after the second line, steam hiss. Natural acoustic space of a small street-front tavern; no words beyond the written lines.
non_diegetic_music:
N/A
Shot 7
15 s · Reference to video · Pika H3 Flex ref2va (all-sparse), 768p
subject_definitions:
<Subject 1> is Kong Yiji, a tall gaunt scholar of about fifty, drawn in 2D anime style: long pale bluish-white face lined with deep wrinkles and two small healed scars on the left cheek and forehead, a scruffy unkempt grey-white beard, tired drooping eyes, the shaved Qing forehead with a thin greasy grey queue down his back; he wears a long grey-white scholar's gown (changshan) that is filthy, stained and torn at the hem and sleeves, never washed or mended, with cloth shoes worn through, as shown in <Picture 1>. This image supplies identity and costume only; its plain portrait backdrop is not the story location.
<Subject 2> is the empty Xianheng tavern in the water town of Luzhen, drawn as a 2D anime watercolor background: a large L-shaped wooden counter opens directly onto the grey flagstone street; behind the counter sit a low clay stove with a copper pot of hot water for warming wine, rows of large brown earthenware wine jars with red cloth-and-clay seals, and small tin wine pots; a dark lacquered wooden shopfront with removable plank doors, black-tiled eaves above, whitewashed walls; through a doorway at the back is a dim inner room with square tables; a small dark chalk slate hangs on a pillar by the counter covered only with faint smudges; a raised wooden threshold below the counter's street side, as shown in <Picture 2>. This image supplies the set layout, materials and colors, not a locked camera angle; it is empty and contains no people.
summary:
[reference generation] A continuous 15.0833-second 2D anime shot from an adaptation of Lu Xun's story set in a late-Qing village tavern. Kong Yiji puts down the bowl and opens his eyes wide in offended dignity, raising a long finger.
retention_analysis:
<Subject 1> (appears in [Shot 1]): fully_preserved - Keep his tall thin build, long pale wrinkled face with the small scars, scruffy grey-white beard, grey queue and the filthy torn long grey-white scholar's gown. Keep the 2D anime character design while performing the directed actions.
<Subject 2> (appears in [Shot 1]): fully_preserved - Keep the L-shaped counter facing the street, the stove and copper pot behind it, the earthenware jars, the chalk slate on the pillar, the dark shopfront and black-tiled eaves in the same spatial arrangement. Lighting follows the time of day described in this shot.
detailed_description:
Chinese 2D hand-drawn anime film style: clean confident ink-like linework, soft cel shading with two-tone shadows, hand-painted watercolor backgrounds with visible paper grain, late-Qing Shaoxing water-town atmosphere. Restrained palette of ink black, faded indigo, grey-white plaster and weathered wood, with warm amber firelight and wine steam inside the counter against cool grey-blue daylight in the street. Expressive but grounded anime acting, consistent character model sheets, stable line weight. No readable text, no Chinese characters, no subtitles, no captions, no logos, no watermarks anywhere in the frame.
Autumn afternoon: warm amber stove glow and wine steam behind the counter, cool grey-blue daylight on the street flagstones.
[Shot 1] One continuous 15.0833-second shot with no internal cut. Medium close shot on Kong Yiji at the counter, a wiry laborer in his forties with a thin moustache, wearing a faded brown short jacket and a straw-rope belt in the left foreground slightly out of focus.
The location is <Subject 2>; keep its counter, stove, jars and shopfront in their established places. Referenced characters on screen: <Subject 1>. Each referenced character appears exactly once, never duplicated; extras, if any, are simpler background figures who do not resemble the referenced characters.
Kong Yiji puts down the bowl and opens his eyes wide in offended dignity, raising a long finger. Then the wiry laborer answers with relish, miming a man hung up and beaten. Kong Yiji's face begins to flush.
<Subject 1>, in the thin, slightly hoarse, nasal voice of a man around fifty, speaking Mandarin Chinese in a slow pedantic drawl with old-fashioned bookish cadence, rising and stammering when embarrassed (S2), says indignant, drawn-out, from second one to second five: <d>[Chinese] 你怎么这样凭空污人清白……</d> Only this character's lips move for this line; listeners stay silent. A wiry laborer in his forties with a thin moustache, wearing a faded brown short jacket and a straw-rope belt, in the raspy, mid-pitched voice of a wiry laborer in his forties, speaking Mandarin Chinese flatly and offhandedly (S6), says mocking, from second six to second thirteen: <d>[Chinese] 什么清白？我前天亲眼见你偷了何家的书，吊着打。</d> Only this speaker's lips move for this line.
Animation stays smooth and on-model: stable faces, consistent line weight, cloth folds and queues moving with the body, steam and firelight drifting gently. Keep anatomy correct, with no extra limbs or duplicated props. End on Kong Yiji's flushed face; next shot he argues back.
Keep a short breath of silence after the final spoken words; no unfinished syllable at the end. Mouths move only for the written Chinese lines; no other intelligible speech, no ad-libs. Absolutely no subtitles, captions, on-screen text, Chinese characters, signboards with writing, logos or watermarks.
overall_soundscape:
A few snorts of laughter, a bowl set down hard. Natural acoustic space of a small street-front tavern; no words beyond the written lines.
non_diegetic_music:
N/A
Shot 8
15 s · Reference to video · Pika H3 Flex ref2va (all-sparse), 768p
subject_definitions:
<Subject 1> is Kong Yiji, a tall gaunt scholar of about fifty, drawn in 2D anime style: long pale bluish-white face lined with deep wrinkles and two small healed scars on the left cheek and forehead, a scruffy unkempt grey-white beard, tired drooping eyes, the shaved Qing forehead with a thin greasy grey queue down his back; he wears a long grey-white scholar's gown (changshan) that is filthy, stained and torn at the hem and sleeves, never washed or mended, with cloth shoes worn through, as shown in <Picture 1>. This image supplies identity and costume only; its plain portrait backdrop is not the story location.
<Subject 2> is the empty Xianheng tavern in the water town of Luzhen, drawn as a 2D anime watercolor background: a large L-shaped wooden counter opens directly onto the grey flagstone street; behind the counter sit a low clay stove with a copper pot of hot water for warming wine, rows of large brown earthenware wine jars with red cloth-and-clay seals, and small tin wine pots; a dark lacquered wooden shopfront with removable plank doors, black-tiled eaves above, whitewashed walls; through a doorway at the back is a dim inner room with square tables; a small dark chalk slate hangs on a pillar by the counter covered only with faint smudges; a raised wooden threshold below the counter's street side, as shown in <Picture 2>. This image supplies the set layout, materials and colors, not a locked camera angle; it is empty and contains no people.
summary:
[reference generation] A continuous 15.0833-second 2D anime shot from an adaptation of Lu Xun's story set in a late-Qing village tavern. Kong Yiji's face turns red, the veins on his forehead stand out, and he argues, stammering, jabbing a finger on the counter on each phrase.
retention_analysis:
<Subject 1> (appears in [Shot 1]): fully_preserved - Keep his tall thin build, long pale wrinkled face with the small scars, scruffy grey-white beard, grey queue and the filthy torn long grey-white scholar's gown. Keep the 2D anime character design while performing the directed actions.
<Subject 2> (appears in [Shot 1]): fully_preserved - Keep the L-shaped counter facing the street, the stove and copper pot behind it, the earthenware jars, the chalk slate on the pillar, the dark shopfront and black-tiled eaves in the same spatial arrangement. Lighting follows the time of day described in this shot.
detailed_description:
Chinese 2D hand-drawn anime film style: clean confident ink-like linework, soft cel shading with two-tone shadows, hand-painted watercolor backgrounds with visible paper grain, late-Qing Shaoxing water-town atmosphere. Restrained palette of ink black, faded indigo, grey-white plaster and weathered wood, with warm amber firelight and wine steam inside the counter against cool grey-blue daylight in the street. Expressive but grounded anime acting, consistent character model sheets, stable line weight. No readable text, no Chinese characters, no subtitles, no captions, no logos, no watermarks anywhere in the frame.
Autumn afternoon: warm amber stove glow and wine steam behind the counter, cool grey-blue daylight on the street flagstones.
[Shot 1] One continuous 15.0833-second shot with no internal cut. Close shot of Kong Yiji, the counter edge in the foreground, laughing drinkers blurred behind him; very slow push in.
The location is <Subject 2>; keep its counter, stove, jars and shopfront in their established places. Referenced characters on screen: <Subject 1>. Each referenced character appears exactly once, never duplicated; extras, if any, are simpler background figures who do not resemble the referenced characters.
Kong Yiji's face turns red, the veins on his forehead stand out, and he argues, stammering, jabbing a finger on the counter on each phrase. After his line the whole counter bursts into laughter; he blinks, lips pressed shut, then turns his face away.
<Subject 1>, in the thin, slightly hoarse, nasal voice of a man around fifty, speaking Mandarin Chinese in a slow pedantic drawl with old-fashioned bookish cadence, rising and stammering when embarrassed (S2), says stammering and flustered, rising in pitch, from second one to second nine: <d>[Chinese] 窃书不能算偷……窃书！……读书人的事，能算偷么？</d> Only this character's lips move for this line; listeners stay silent.
Animation stays smooth and on-model: stable faces, consistent line weight, cloth folds and queues moving with the body, steam and firelight drifting gently. Keep anatomy correct, with no extra limbs or duplicated props. End on him turning away amid the laughter; next shot the narrator explains his past.
Keep a short breath of silence after the final spoken words; no unfinished syllable at the end. Mouths move only for the written Chinese lines; no other intelligible speech, no ad-libs. Absolutely no subtitles, captions, on-screen text, Chinese characters, signboards with writing, logos or watermarks.
overall_soundscape:
Loud hearty laughter of several men from second nine, bowls rattling on the counter. Natural acoustic space of a small street-front tavern; no words beyond the written lines.
non_diegetic_music:
N/A
Shot 9
15 s · Reference to video · Pika H3 Flex ref2va (all-sparse), 768p
subject_definitions:
<Subject 1> is Kong Yiji, a tall gaunt scholar of about fifty, drawn in 2D anime style: long pale bluish-white face lined with deep wrinkles and two small healed scars on the left cheek and forehead, a scruffy unkempt grey-white beard, tired drooping eyes, the shaved Qing forehead with a thin greasy grey queue down his back; he wears a long grey-white scholar's gown (changshan) that is filthy, stained and torn at the hem and sleeves, never washed or mended, with cloth shoes worn through, as shown in <Picture 1>. This image supplies identity and costume only; its plain portrait backdrop is not the story location.
<Subject 2> is the empty Xianheng tavern in the water town of Luzhen, drawn as a 2D anime watercolor background: a large L-shaped wooden counter opens directly onto the grey flagstone street; behind the counter sit a low clay stove with a copper pot of hot water for warming wine, rows of large brown earthenware wine jars with red cloth-and-clay seals, and small tin wine pots; a dark lacquered wooden shopfront with removable plank doors, black-tiled eaves above, whitewashed walls; through a doorway at the back is a dim inner room with square tables; a small dark chalk slate hangs on a pillar by the counter covered only with faint smudges; a raised wooden threshold below the counter's street side, as shown in <Picture 2>. This image supplies the set layout, materials and colors, not a locked camera angle; it is empty and contains no people.
summary:
[reference generation] A continuous 15.0833-second 2D anime shot from an adaptation of Lu Xun's story set in a late-Qing village tavern. Kong Yiji slowly picks up one aniseed bean, chews, stares into the wine bowl.
retention_analysis:
<Subject 1> (appears in [Shot 1]): fully_preserved - Keep his tall thin build, long pale wrinkled face with the small scars, scruffy grey-white beard, grey queue and the filthy torn long grey-white scholar's gown. Keep the 2D anime character design while performing the directed actions.
<Subject 2> (appears in [Shot 1]): fully_preserved - Keep the L-shaped counter facing the street, the stove and copper pot behind it, the earthenware jars, the chalk slate on the pillar, the dark shopfront and black-tiled eaves in the same spatial arrangement. Lighting follows the time of day described in this shot.
detailed_description:
Chinese 2D hand-drawn anime film style: clean confident ink-like linework, soft cel shading with two-tone shadows, hand-painted watercolor backgrounds with visible paper grain, late-Qing Shaoxing water-town atmosphere. Restrained palette of ink black, faded indigo, grey-white plaster and weathered wood, with warm amber firelight and wine steam inside the counter against cool grey-blue daylight in the street. Expressive but grounded anime acting, consistent character model sheets, stable line weight. No readable text, no Chinese characters, no subtitles, no captions, no logos, no watermarks anywhere in the frame.
Autumn afternoon: warm amber stove glow and wine steam behind the counter, cool grey-blue daylight on the street flagstones.
[Shot 1] One continuous 15.0833-second shot with no internal cut. Medium side view: Kong Yiji alone at the far end of the counter, hunched over his bowl, the laughing laborers small in the background on the left.
The location is <Subject 2>; keep its counter, stove, jars and shopfront in their established places. Referenced characters on screen: <Subject 1>. Each referenced character appears exactly once, never duplicated; extras, if any, are simpler background figures who do not resemble the referenced characters.
Kong Yiji slowly picks up one aniseed bean, chews, stares into the wine bowl. He smooths the frayed edge of his sleeve. His lips stay closed throughout the narration.
The unseen narrator, in the calm, low, slightly weary voice of a man in his forties, speaking Mandarin Chinese with measured, reflective pacing, remembering his childhood (S1), says in an off-screen voiceover, quietly and evenly, from second one to second thirteen: <d>[Chinese] 听人家背地里谈论，孔乙己原来也读过书，但终于没有进学，又不会营生。</d> All visible characters keep their lips closed during the narration.
Animation stays smooth and on-model: stable faces, consistent line weight, cloth folds and queues moving with the body, steam and firelight drifting gently. Keep anatomy correct, with no extra limbs or duplicated props. End with him draining the bowl; next shot shows what he does for a living.
Keep a short breath of silence after the final spoken words; no unfinished syllable at the end. Mouths move only for the written Chinese lines; no other intelligible speech, no ad-libs. Absolutely no subtitles, captions, on-screen text, Chinese characters, signboards with writing, logos or watermarks.
overall_soundscape:
Distant laughter dying down, wind in the eaves. Natural acoustic space of a small street-front tavern; no words beyond the written lines.
non_diegetic_music:
N/A
Shot 10
15 s · Reference to video · Pika H3 Flex ref2va (all-sparse), 768p
subject_definitions:
<Subject 1> is Kong Yiji, a tall gaunt scholar of about fifty, drawn in 2D anime style: long pale bluish-white face lined with deep wrinkles and two small healed scars on the left cheek and forehead, a scruffy unkempt grey-white beard, tired drooping eyes, the shaved Qing forehead with a thin greasy grey queue down his back; he wears a long grey-white scholar's gown (changshan) that is filthy, stained and torn at the hem and sleeves, never washed or mended, with cloth shoes worn through, as shown in <Picture 1>. This image supplies identity and costume only; its plain portrait backdrop is not the story location.
summary:
[reference generation] A continuous 15.0833-second 2D anime shot from an adaptation of Lu Xun's story set in a late-Qing village tavern. For the first half Kong Yiji copies a book with a fine brush, his writing small and neat (the strokes are tiny and unreadable at this distance).
retention_analysis:
<Subject 1> (appears in [Shot 1]): fully_preserved - Keep his tall thin build, long pale wrinkled face with the small scars, scruffy grey-white beard, grey queue and the filthy torn long grey-white scholar's gown. Keep the 2D anime character design while performing the directed actions.
detailed_description:
Chinese 2D hand-drawn anime film style: clean confident ink-like linework, soft cel shading with two-tone shadows, hand-painted watercolor backgrounds with visible paper grain, late-Qing Shaoxing water-town atmosphere. Restrained palette of ink black, faded indigo, grey-white plaster and weathered wood, with warm amber firelight and wine steam inside the counter against cool grey-blue daylight in the street. Expressive but grounded anime acting, consistent character model sheets, stable line weight. No readable text, no Chinese characters, no subtitles, no captions, no logos, no watermarks anywhere in the frame.
Autumn afternoon: warm amber stove glow and wine steam behind the counter, cool grey-blue daylight on the street flagstones.
[Shot 1] One continuous 15.0833-second shot with no internal cut. A cutaway: a cramped dim rented room with a small wooden desk by a paper window, a stack of thread-bound books and an ink stone; medium shot of Kong Yiji at the desk.
This brief cutaway takes place in Kong Yiji's cramped rented room as described, not in the tavern. Referenced characters on screen: <Subject 1>. Each referenced character appears exactly once, never duplicated; extras, if any, are simpler background figures who do not resemble the referenced characters.
For the first half Kong Yiji copies a book with a fine brush, his writing small and neat (the strokes are tiny and unreadable at this distance). Then he yawns, pushes the brush and paper aside, rests his cheek on his arm and dozes. The books slide slightly. His lips stay closed while the narrator speaks.
The unseen narrator, in the calm, low, slightly weary voice of a man in his forties, speaking Mandarin Chinese with measured, reflective pacing, remembering his childhood (S1), says in an off-screen voiceover, quietly, with dry irony, from second zero to second fourteen: <d>[Chinese] 幸而写得一笔好字，替人家抄抄书；可惜他好喝懒做，坐不到几天，便连人带书一齐失踪。</d> All visible characters keep their lips closed during the narration.
Animation stays smooth and on-model: stable faces, consistent line weight, cloth folds and queues moving with the body, steam and firelight drifting gently. Keep anatomy correct, with no extra limbs or duplicated props. End on him asleep; next shot returns to the tavern counter.
Keep a short breath of silence after the final spoken words; no unfinished syllable at the end. Mouths move only for the written Chinese lines; no other intelligible speech, no ad-libs. Absolutely no subtitles, captions, on-screen text, Chinese characters, signboards with writing, logos or watermarks.
overall_soundscape:
Brush scratching on paper, a yawn, a gentle snore. Natural acoustic space of a small street-front tavern; no words beyond the written lines.
non_diegetic_music:
N/A
Shot 11
12 s · Reference to video · Pika H3 Flex ref2va (all-sparse), 768p
subject_definitions:
<Subject 1> is Kong Yiji, a tall gaunt scholar of about fifty, drawn in 2D anime style: long pale bluish-white face lined with deep wrinkles and two small healed scars on the left cheek and forehead, a scruffy unkempt grey-white beard, tired drooping eyes, the shaved Qing forehead with a thin greasy grey queue down his back; he wears a long grey-white scholar's gown (changshan) that is filthy, stained and torn at the hem and sleeves, never washed or mended, with cloth shoes worn through, as shown in <Picture 1>. This image supplies identity and costume only; its plain portrait backdrop is not the story location.
<Subject 2> is the tavern owner, a stout man in his fifties drawn in 2D anime style: round fleshy face with heavy jowls, small sharp narrow eyes, a thin drooping black moustache, a black round skullcap, a long black queue, a dark brown cotton long gown under a black sleeveless waistcoat, often holding an abacus, as shown in <Picture 2>. This image supplies identity and costume only; its plain portrait backdrop is not the story location.
<Subject 3> is the empty Xianheng tavern in the water town of Luzhen, drawn as a 2D anime watercolor background: a large L-shaped wooden counter opens directly onto the grey flagstone street; behind the counter sit a low clay stove with a copper pot of hot water for warming wine, rows of large brown earthenware wine jars with red cloth-and-clay seals, and small tin wine pots; a dark lacquered wooden shopfront with removable plank doors, black-tiled eaves above, whitewashed walls; through a doorway at the back is a dim inner room with square tables; a small dark chalk slate hangs on a pillar by the counter covered only with faint smudges; a raised wooden threshold below the counter's street side, as shown in <Picture 3>. This image supplies the set layout, materials and colors, not a locked camera angle; it is empty and contains no people.
summary:
[reference generation] A continuous 12.2500-second 2D anime shot from an adaptation of Lu Xun's story set in a late-Qing village tavern. Kong Yiji counts several copper coins onto the counter.
retention_analysis:
<Subject 1> (appears in [Shot 1]): fully_preserved - Keep his tall thin build, long pale wrinkled face with the small scars, scruffy grey-white beard, grey queue and the filthy torn long grey-white scholar's gown. Keep the 2D anime character design while performing the directed actions.
<Subject 2> (appears in [Shot 1]): fully_preserved - Keep his stout build, round fleshy face, small narrow eyes, thin moustache, black skullcap and the dark brown gown with black waistcoat. Keep the 2D anime character design while performing the directed actions.
<Subject 3> (appears in [Shot 1]): fully_preserved - Keep the L-shaped counter facing the street, the stove and copper pot behind it, the earthenware jars, the chalk slate on the pillar, the dark shopfront and black-tiled eaves in the same spatial arrangement. Lighting follows the time of day described in this shot.
detailed_description:
Chinese 2D hand-drawn anime film style: clean confident ink-like linework, soft cel shading with two-tone shadows, hand-painted watercolor backgrounds with visible paper grain, late-Qing Shaoxing water-town atmosphere. Restrained palette of ink black, faded indigo, grey-white plaster and weathered wood, with warm amber firelight and wine steam inside the counter against cool grey-blue daylight in the street. Expressive but grounded anime acting, consistent character model sheets, stable line weight. No readable text, no Chinese characters, no subtitles, no captions, no logos, no watermarks anywhere in the frame.
Autumn afternoon: warm amber stove glow and wine steam behind the counter, cool grey-blue daylight on the street flagstones.
[Shot 1] One continuous 12.2500-second shot with no internal cut. Medium shot across the counter: the owner behind it on the right beside the chalk slate on the pillar, Kong Yiji outside on the left.
The location is <Subject 3>; keep its counter, stove, jars and shopfront in their established places. Referenced characters on screen: <Subject 1>, <Subject 2>. Each referenced character appears exactly once, never duplicated; extras, if any, are simpler background figures who do not resemble the referenced characters.
Kong Yiji counts several copper coins onto the counter. The owner sweeps them into his palm, then turns and wipes a few faint chalk marks off the dark slate with his sleeve; the slate stays without any readable writing. The owner gives a short satisfied grunt. Both keep their lips closed during the narration.
The unseen narrator, in the calm, low, slightly weary voice of a man in his forties, speaking Mandarin Chinese with measured, reflective pacing, remembering his childhood (S1), says in an off-screen voiceover, quietly, from second one to second ten: <d>[Chinese] 但他在我们店里，品行却比别人都好，就是从不拖欠。</d> All visible characters keep their lips closed during the narration.
Animation stays smooth and on-model: stable faces, consistent line weight, cloth folds and queues moving with the body, steam and firelight drifting gently. Keep anatomy correct, with no extra limbs or duplicated props. End on the wiped slate; next shot the laborers start in again.
Keep a short breath of silence after the final spoken words; no unfinished syllable at the end. Mouths move only for the written Chinese lines; no other intelligible speech, no ad-libs. Absolutely no subtitles, captions, on-screen text, Chinese characters, signboards with writing, logos or watermarks.
overall_soundscape:
Coins clink, cloth rubbing on slate, a grunt. Natural acoustic space of a small street-front tavern; no words beyond the written lines.
non_diegetic_music:
N/A
Shot 12
15 s · Reference to video · Pika H3 Flex ref2va (all-sparse), 768p
subject_definitions:
<Subject 1> is Kong Yiji, a tall gaunt scholar of about fifty, drawn in 2D anime style: long pale bluish-white face lined with deep wrinkles and two small healed scars on the left cheek and forehead, a scruffy unkempt grey-white beard, tired drooping eyes, the shaved Qing forehead with a thin greasy grey queue down his back; he wears a long grey-white scholar's gown (changshan) that is filthy, stained and torn at the hem and sleeves, never washed or mended, with cloth shoes worn through, as shown in <Picture 1>. This image supplies identity and costume only; its plain portrait backdrop is not the story location.
<Subject 2> is the empty Xianheng tavern in the water town of Luzhen, drawn as a 2D anime watercolor background: a large L-shaped wooden counter opens directly onto the grey flagstone street; behind the counter sit a low clay stove with a copper pot of hot water for warming wine, rows of large brown earthenware wine jars with red cloth-and-clay seals, and small tin wine pots; a dark lacquered wooden shopfront with removable plank doors, black-tiled eaves above, whitewashed walls; through a doorway at the back is a dim inner room with square tables; a small dark chalk slate hangs on a pillar by the counter covered only with faint smudges; a raised wooden threshold below the counter's street side, as shown in <Picture 2>. This image supplies the set layout, materials and colors, not a locked camera angle; it is empty and contains no people.
summary:
[reference generation] A continuous 15.0833-second 2D anime shot from an adaptation of Lu Xun's story set in a late-Qing village tavern. The burly laborer leans close and asks; Kong Yiji glances at him with disdain and lifts his chin.
retention_analysis:
<Subject 1> (appears in [Shot 1]): fully_preserved - Keep his tall thin build, long pale wrinkled face with the small scars, scruffy grey-white beard, grey queue and the filthy torn long grey-white scholar's gown. Keep the 2D anime character design while performing the directed actions.
<Subject 2> (appears in [Shot 1]): fully_preserved - Keep the L-shaped counter facing the street, the stove and copper pot behind it, the earthenware jars, the chalk slate on the pillar, the dark shopfront and black-tiled eaves in the same spatial arrangement. Lighting follows the time of day described in this shot.
detailed_description:
Chinese 2D hand-drawn anime film style: clean confident ink-like linework, soft cel shading with two-tone shadows, hand-painted watercolor backgrounds with visible paper grain, late-Qing Shaoxing water-town atmosphere. Restrained palette of ink black, faded indigo, grey-white plaster and weathered wood, with warm amber firelight and wine steam inside the counter against cool grey-blue daylight in the street. Expressive but grounded anime acting, consistent character model sheets, stable line weight. No readable text, no Chinese characters, no subtitles, no captions, no logos, no watermarks anywhere in the frame.
Autumn afternoon: warm amber stove glow and wine steam behind the counter, cool grey-blue daylight on the street flagstones.
[Shot 1] One continuous 15.0833-second shot with no internal cut. Medium two-shot from the street: Kong Yiji at the counter, a burly laborer in his thirties with a sun-browned face, wearing a patched black cotton short jacket beside him, other drinkers grinning behind.
The location is <Subject 2>; keep its counter, stove, jars and shopfront in their established places. Referenced characters on screen: <Subject 1>. Each referenced character appears exactly once, never duplicated; extras, if any, are simpler background figures who do not resemble the referenced characters.
The burly laborer leans close and asks; Kong Yiji glances at him with disdain and lifts his chin. The laborer presses with a second question. Kong Yiji's face turns grey and uneasy, and he murmurs indistinct classical particles under his breath, lips barely moving, then turns back to his bowl as the others laugh.
A burly laborer in his thirties with a sun-browned face, wearing a patched black cotton short jacket, in the loud, rough, cheerful voice of a laborer in his thirties, speaking Mandarin Chinese with teasing relish (S5), says with mock curiosity, from second zero to second three: <d>[Chinese] 孔乙己，你当真认识字么？</d> Only this speaker's lips move for this line. A burly laborer in his thirties with a sun-browned face, wearing a patched black cotton short jacket, in the loud, rough, cheerful voice of a laborer in his thirties, speaking Mandarin Chinese with teasing relish (S5), says jeering, from second five to second nine: <d>[Chinese] 你怎的连半个秀才也捞不到呢？</d> Only this speaker's lips move for this line.
Animation stays smooth and on-model: stable faces, consistent line weight, cloth folds and queues moving with the body, steam and firelight drifting gently. Keep anatomy correct, with no extra limbs or duplicated props. End on him turning back to the bowl; next shot he turns to the boy instead.
Keep a short breath of silence after the final spoken words; no unfinished syllable at the end. Mouths move only for the written Chinese lines; no other intelligible speech, no ad-libs. Absolutely no subtitles, captions, on-screen text, Chinese characters, signboards with writing, logos or watermarks.
overall_soundscape:
An indistinct mumble, then laughter swelling, bowls clinking. Natural acoustic space of a small street-front tavern; no words beyond the written lines.
non_diegetic_music:
N/A
Shot 13
15 s · Reference to video · Pika H3 Flex ref2va (all-sparse), 768p
subject_definitions:
<Subject 1> is Kong Yiji, a tall gaunt scholar of about fifty, drawn in 2D anime style: long pale bluish-white face lined with deep wrinkles and two small healed scars on the left cheek and forehead, a scruffy unkempt grey-white beard, tired drooping eyes, the shaved Qing forehead with a thin greasy grey queue down his back; he wears a long grey-white scholar's gown (changshan) that is filthy, stained and torn at the hem and sleeves, never washed or mended, with cloth shoes worn through, as shown in <Picture 1>. This image supplies identity and costume only; its plain portrait backdrop is not the story location.
<Subject 2> is the young waiter, a thin twelve-year-old boy drawn in 2D anime style: round boyish face, large dark attentive eyes, shaved Qing forehead with a short black queue, a plain faded indigo cotton short jacket and trousers, a stained off-white cloth apron tied at the waist, black cloth shoes, as shown in <Picture 2>. This image supplies identity and costume only; its plain portrait backdrop is not the story location.
<Subject 3> is the interior of the same tavern seen from behind the counter: the worn wooden countertop across the foreground, the clay stove with copper hot-water pot and tin wine pots, dark earthenware wine jars with red seals, the open shopfront and the grey flagstone street with whitewashed walls and black-tiled roofs beyond, as shown in <Picture 3>. This image supplies the set layout, materials and colors, not a locked camera angle; it is empty and contains no people.
summary:
[reference generation] A continuous 15.0833-second 2D anime shot from an adaptation of Lu Xun's story set in a late-Qing village tavern. Kong Yiji, unable to talk with the drinkers, leans over the counter toward the boy and asks his first question.
retention_analysis:
<Subject 1> (appears in [Shot 1]): fully_preserved - Keep his tall thin build, long pale wrinkled face with the small scars, scruffy grey-white beard, grey queue and the filthy torn long grey-white scholar's gown. Keep the 2D anime character design while performing the directed actions.
<Subject 2> (appears in [Shot 1]): fully_preserved - Keep his twelve-year-old face and size, short black queue, faded indigo short jacket and the off-white apron. Keep the 2D anime character design while performing the directed actions.
<Subject 3> (appears in [Shot 1]): fully_preserved - Keep the L-shaped counter facing the street, the stove and copper pot behind it, the earthenware jars, the chalk slate on the pillar, the dark shopfront and black-tiled eaves in the same spatial arrangement. Lighting follows the time of day described in this shot.
detailed_description:
Chinese 2D hand-drawn anime film style: clean confident ink-like linework, soft cel shading with two-tone shadows, hand-painted watercolor backgrounds with visible paper grain, late-Qing Shaoxing water-town atmosphere. Restrained palette of ink black, faded indigo, grey-white plaster and weathered wood, with warm amber firelight and wine steam inside the counter against cool grey-blue daylight in the street. Expressive but grounded anime acting, consistent character model sheets, stable line weight. No readable text, no Chinese characters, no subtitles, no captions, no logos, no watermarks anywhere in the frame.
Autumn afternoon: warm amber stove glow and wine steam behind the counter, cool grey-blue daylight on the street flagstones.
[Shot 1] One continuous 15.0833-second shot with no internal cut. Over-the-shoulder from behind the counter: the boy in the foreground right, back three-quarter to camera, Kong Yiji outside the counter on the left, leaning in.
The location is <Subject 3>; keep its counter, stove, jars and shopfront in their established places. Referenced characters on screen: <Subject 1>, <Subject 2>. Each referenced character appears exactly once, never duplicated; extras, if any, are simpler background figures who do not resemble the referenced characters.
Kong Yiji, unable to talk with the drinkers, leans over the counter toward the boy and asks his first question. The boy gives a tiny reluctant nod, lips closed. Kong Yiji brightens, raises a lecturing finger and asks the second question.
<Subject 1>, in the thin, slightly hoarse, nasal voice of a man around fifty, speaking Mandarin Chinese in a slow pedantic drawl with old-fashioned bookish cadence, rising and stammering when embarrassed (S2), says gently, from second one to second three: <d>[Chinese] 你读过书么？</d> Only this character's lips move for this line; listeners stay silent. <Subject 1>, in the thin, slightly hoarse, nasal voice of a man around fifty, speaking Mandarin Chinese in a slow pedantic drawl with old-fashioned bookish cadence, rising and stammering when embarrassed (S2), says pleased and pedantic, from second five to second thirteen: <d>[Chinese] 读过书……我便考你一考。茴香豆的茴字，怎样写的？</d> Only this character's lips move for this line; listeners stay silent.
Animation stays smooth and on-model: stable faces, consistent line weight, cloth folds and queues moving with the body, steam and firelight drifting gently. Keep anatomy correct, with no extra limbs or duplicated props. End on Kong Yiji waiting eagerly; next shot the boy turns away.
Keep a short breath of silence after the final spoken words; no unfinished syllable at the end. Mouths move only for the written Chinese lines; no other intelligible speech, no ad-libs. Absolutely no subtitles, captions, on-screen text, Chinese characters, signboards with writing, logos or watermarks.
overall_soundscape:
Fire crackle, steam, a faint laugh from the street. Natural acoustic space of a small street-front tavern; no words beyond the written lines.
non_diegetic_music:
N/A
Shot 14
12 s · Reference to video · Pika H3 Flex ref2va (all-sparse), 768p
subject_definitions:
<Subject 1> is Kong Yiji, a tall gaunt scholar of about fifty, drawn in 2D anime style: long pale bluish-white face lined with deep wrinkles and two small healed scars on the left cheek and forehead, a scruffy unkempt grey-white beard, tired drooping eyes, the shaved Qing forehead with a thin greasy grey queue down his back; he wears a long grey-white scholar's gown (changshan) that is filthy, stained and torn at the hem and sleeves, never washed or mended, with cloth shoes worn through, as shown in <Picture 1>. This image supplies identity and costume only; its plain portrait backdrop is not the story location.
<Subject 2> is the young waiter, a thin twelve-year-old boy drawn in 2D anime style: round boyish face, large dark attentive eyes, shaved Qing forehead with a short black queue, a plain faded indigo cotton short jacket and trousers, a stained off-white cloth apron tied at the waist, black cloth shoes, as shown in <Picture 2>. This image supplies identity and costume only; its plain portrait backdrop is not the story location.
<Subject 3> is the interior of the same tavern seen from behind the counter: the worn wooden countertop across the foreground, the clay stove with copper hot-water pot and tin wine pots, dark earthenware wine jars with red seals, the open shopfront and the grey flagstone street with whitewashed walls and black-tiled roofs beyond, as shown in <Picture 3>. This image supplies the set layout, materials and colors, not a locked camera angle; it is empty and contains no people.
summary:
[reference generation] A continuous 12.2500-second 2D anime shot from an adaptation of Lu Xun's story set in a late-Qing village tavern. The boy turns away and busies himself with the wine pots, lips closed, while the unseen narrator speaks his thought.
retention_analysis:
<Subject 1> (appears in [Shot 1]): fully_preserved - Keep his tall thin build, long pale wrinkled face with the small scars, scruffy grey-white beard, grey queue and the filthy torn long grey-white scholar's gown. Keep the 2D anime character design while performing the directed actions.
<Subject 2> (appears in [Shot 1]): fully_preserved - Keep his twelve-year-old face and size, short black queue, faded indigo short jacket and the off-white apron. Keep the 2D anime character design while performing the directed actions.
<Subject 3> (appears in [Shot 1]): fully_preserved - Keep the L-shaped counter facing the street, the stove and copper pot behind it, the earthenware jars, the chalk slate on the pillar, the dark shopfront and black-tiled eaves in the same spatial arrangement. Lighting follows the time of day described in this shot.
detailed_description:
Chinese 2D hand-drawn anime film style: clean confident ink-like linework, soft cel shading with two-tone shadows, hand-painted watercolor backgrounds with visible paper grain, late-Qing Shaoxing water-town atmosphere. Restrained palette of ink black, faded indigo, grey-white plaster and weathered wood, with warm amber firelight and wine steam inside the counter against cool grey-blue daylight in the street. Expressive but grounded anime acting, consistent character model sheets, stable line weight. No readable text, no Chinese characters, no subtitles, no captions, no logos, no watermarks anywhere in the frame.
Autumn afternoon: warm amber stove glow and wine steam behind the counter, cool grey-blue daylight on the street flagstones.
[Shot 1] One continuous 12.2500-second shot with no internal cut. Reverse medium shot from beside the stove: the boy in the foreground turning away, Kong Yiji behind him beyond the counter.
The location is <Subject 3>; keep its counter, stove, jars and shopfront in their established places. Referenced characters on screen: <Subject 1>, <Subject 2>. Each referenced character appears exactly once, never duplicated; extras, if any, are simpler background figures who do not resemble the referenced characters.
The boy turns away and busies himself with the wine pots, lips closed, while the unseen narrator speaks his thought. Kong Yiji waits, then leans further over the counter and speaks earnestly.
The unseen narrator, in the calm, low, slightly weary voice of a man in his forties, speaking Mandarin Chinese with measured, reflective pacing, remembering his childhood (S1), says in an off-screen voiceover, dry, a boy's contempt remembered, from second zero to second four: <d>[Chinese] 讨饭一样的人，也配考我么？</d> All visible characters keep their lips closed during the narration. <Subject 1>, in the thin, slightly hoarse, nasal voice of a man around fifty, speaking Mandarin Chinese in a slow pedantic drawl with old-fashioned bookish cadence, rising and stammering when embarrassed (S2), says earnest and kind, from second six to second ten: <d>[Chinese] 不能写罢？……我教给你，记着！</d> Only this character's lips move for this line; listeners stay silent.
Animation stays smooth and on-model: stable faces, consistent line weight, cloth folds and queues moving with the body, steam and firelight drifting gently. Keep anatomy correct, with no extra limbs or duplicated props. End on Kong Yiji dipping his fingernail toward his bowl; next shot he writes.
Keep a short breath of silence after the final spoken words; no unfinished syllable at the end. Mouths move only for the written Chinese lines; no other intelligible speech, no ad-libs. Absolutely no subtitles, captions, on-screen text, Chinese characters, signboards with writing, logos or watermarks.
overall_soundscape:
Tin pots clinking, steam. Natural acoustic space of a small street-front tavern; no words beyond the written lines.
non_diegetic_music:
N/A
Shot 15
12 s · Reference to video · Pika H3 Flex ref2va (all-sparse), 768p
subject_definitions:
<Subject 1> is Kong Yiji, a tall gaunt scholar of about fifty, drawn in 2D anime style: long pale bluish-white face lined with deep wrinkles and two small healed scars on the left cheek and forehead, a scruffy unkempt grey-white beard, tired drooping eyes, the shaved Qing forehead with a thin greasy grey queue down his back; he wears a long grey-white scholar's gown (changshan) that is filthy, stained and torn at the hem and sleeves, never washed or mended, with cloth shoes worn through, as shown in <Picture 1>. This image supplies identity and costume only; its plain portrait backdrop is not the story location.
<Subject 2> is the young waiter, a thin twelve-year-old boy drawn in 2D anime style: round boyish face, large dark attentive eyes, shaved Qing forehead with a short black queue, a plain faded indigo cotton short jacket and trousers, a stained off-white cloth apron tied at the waist, black cloth shoes, as shown in <Picture 2>. This image supplies identity and costume only; its plain portrait backdrop is not the story location.
<Subject 3> is the interior of the same tavern seen from behind the counter: the worn wooden countertop across the foreground, the clay stove with copper hot-water pot and tin wine pots, dark earthenware wine jars with red seals, the open shopfront and the grey flagstone street with whitewashed walls and black-tiled roofs beyond, as shown in <Picture 3>. This image supplies the set layout, materials and colors, not a locked camera angle; it is empty and contains no people.
summary:
[reference generation] A continuous 12.2500-second 2D anime shot from an adaptation of Lu Xun's story set in a late-Qing village tavern. Kong Yiji dips his long fingernail into the wine and draws wet strokes on the counter; the wet marks are abstract smudges, not readable characters.
retention_analysis:
<Subject 1> (appears in [Shot 1]): fully_preserved - Keep his tall thin build, long pale wrinkled face with the small scars, scruffy grey-white beard, grey queue and the filthy torn long grey-white scholar's gown. Keep the 2D anime character design while performing the directed actions.
<Subject 2> (appears in [Shot 1]): fully_preserved - Keep his twelve-year-old face and size, short black queue, faded indigo short jacket and the off-white apron. Keep the 2D anime character design while performing the directed actions.
<Subject 3> (appears in [Shot 1]): fully_preserved - Keep the L-shaped counter facing the street, the stove and copper pot behind it, the earthenware jars, the chalk slate on the pillar, the dark shopfront and black-tiled eaves in the same spatial arrangement. Lighting follows the time of day described in this shot.
detailed_description:
Chinese 2D hand-drawn anime film style: clean confident ink-like linework, soft cel shading with two-tone shadows, hand-painted watercolor backgrounds with visible paper grain, late-Qing Shaoxing water-town atmosphere. Restrained palette of ink black, faded indigo, grey-white plaster and weathered wood, with warm amber firelight and wine steam inside the counter against cool grey-blue daylight in the street. Expressive but grounded anime acting, consistent character model sheets, stable line weight. No readable text, no Chinese characters, no subtitles, no captions, no logos, no watermarks anywhere in the frame.
Autumn afternoon: warm amber stove glow and wine steam behind the counter, cool grey-blue daylight on the street flagstones.
[Shot 1] One continuous 12.2500-second shot with no internal cut. Close shot on the countertop from above the boy's shoulder: Kong Yiji's long fingernail and his bowl on the left, the boy's hand on the right.
The location is <Subject 3>; keep its counter, stove, jars and shopfront in their established places. Referenced characters on screen: <Subject 1>, <Subject 2>. Each referenced character appears exactly once, never duplicated; extras, if any, are simpler background figures who do not resemble the referenced characters.
Kong Yiji dips his long fingernail into the wine and draws wet strokes on the counter; the wet marks are abstract smudges, not readable characters. The boy, impatient, speaks without looking at the strokes. Kong Yiji's eyes light up.
<Subject 2>, in the clear, high, slightly impatient voice of a twelve-year-old boy, speaking Mandarin Chinese briskly (S3), says briskly and impatiently, from second three to second nine: <d>[Chinese] 谁要你教，不是草头底下一个来回的回字么？</d> Only this character's lips move for this line; listeners stay silent.
Animation stays smooth and on-model: stable faces, consistent line weight, cloth folds and queues moving with the body, steam and firelight drifting gently. Keep anatomy correct, with no extra limbs or duplicated props. End on Kong Yiji's delighted face; next shot he speaks.
Keep a short breath of silence after the final spoken words; no unfinished syllable at the end. Mouths move only for the written Chinese lines; no other intelligible speech, no ad-libs. Absolutely no subtitles, captions, on-screen text, Chinese characters, signboards with writing, logos or watermarks.
overall_soundscape:
Fingernail squeak on wet wood, a sniff. Natural acoustic space of a small street-front tavern; no words beyond the written lines.
non_diegetic_music:
N/A
Shot 16
12 s · Reference to video · Pika H3 Flex ref2va (all-sparse), 768p
subject_definitions:
<Subject 1> is Kong Yiji, a tall gaunt scholar of about fifty, drawn in 2D anime style: long pale bluish-white face lined with deep wrinkles and two small healed scars on the left cheek and forehead, a scruffy unkempt grey-white beard, tired drooping eyes, the shaved Qing forehead with a thin greasy grey queue down his back; he wears a long grey-white scholar's gown (changshan) that is filthy, stained and torn at the hem and sleeves, never washed or mended, with cloth shoes worn through, as shown in <Picture 1>. This image supplies identity and costume only; its plain portrait backdrop is not the story location.
<Subject 2> is the young waiter, a thin twelve-year-old boy drawn in 2D anime style: round boyish face, large dark attentive eyes, shaved Qing forehead with a short black queue, a plain faded indigo cotton short jacket and trousers, a stained off-white cloth apron tied at the waist, black cloth shoes, as shown in <Picture 2>. This image supplies identity and costume only; its plain portrait backdrop is not the story location.
<Subject 3> is the interior of the same tavern seen from behind the counter: the worn wooden countertop across the foreground, the clay stove with copper hot-water pot and tin wine pots, dark earthenware wine jars with red seals, the open shopfront and the grey flagstone street with whitewashed walls and black-tiled roofs beyond, as shown in <Picture 3>. This image supplies the set layout, materials and colors, not a locked camera angle; it is empty and contains no people.
summary:
[reference generation] A continuous 12.2500-second 2D anime shot from an adaptation of Lu Xun's story set in a late-Qing village tavern. Kong Yiji taps two long fingernails on the counter, overjoyed, and asks.
retention_analysis:
<Subject 1> (appears in [Shot 1]): fully_preserved - Keep his tall thin build, long pale wrinkled face with the small scars, scruffy grey-white beard, grey queue and the filthy torn long grey-white scholar's gown. Keep the 2D anime character design while performing the directed actions.
<Subject 2> (appears in [Shot 1]): fully_preserved - Keep his twelve-year-old face and size, short black queue, faded indigo short jacket and the off-white apron. Keep the 2D anime character design while performing the directed actions.
<Subject 3> (appears in [Shot 1]): fully_preserved - Keep the L-shaped counter facing the street, the stove and copper pot behind it, the earthenware jars, the chalk slate on the pillar, the dark shopfront and black-tiled eaves in the same spatial arrangement. Lighting follows the time of day described in this shot.
detailed_description:
Chinese 2D hand-drawn anime film style: clean confident ink-like linework, soft cel shading with two-tone shadows, hand-painted watercolor backgrounds with visible paper grain, late-Qing Shaoxing water-town atmosphere. Restrained palette of ink black, faded indigo, grey-white plaster and weathered wood, with warm amber firelight and wine steam inside the counter against cool grey-blue daylight in the street. Expressive but grounded anime acting, consistent character model sheets, stable line weight. No readable text, no Chinese characters, no subtitles, no captions, no logos, no watermarks anywhere in the frame.
Autumn afternoon: warm amber stove glow and wine steam behind the counter, cool grey-blue daylight on the street flagstones.
[Shot 1] One continuous 12.2500-second shot with no internal cut. Medium shot across the counter: Kong Yiji on the left beaming, the boy on the right.
The location is <Subject 3>; keep its counter, stove, jars and shopfront in their established places. Referenced characters on screen: <Subject 1>, <Subject 2>. Each referenced character appears exactly once, never duplicated; extras, if any, are simpler background figures who do not resemble the referenced characters.
Kong Yiji taps two long fingernails on the counter, overjoyed, and asks. The boy pouts, lips closed, and walks away toward the stove. Kong Yiji's hand freezes mid-gesture, he sighs, and looks very regretful.
<Subject 1>, in the thin, slightly hoarse, nasal voice of a man around fifty, speaking Mandarin Chinese in a slow pedantic drawl with old-fashioned bookish cadence, rising and stammering when embarrassed (S2), says delighted, eager, from second one to second seven: <d>[Chinese] 对呀对呀！……回字有四样写法，你知道么？</d> Only this character's lips move for this line; listeners stay silent.
Animation stays smooth and on-model: stable faces, consistent line weight, cloth folds and queues moving with the body, steam and firelight drifting gently. Keep anatomy correct, with no extra limbs or duplicated props. End on his regretful sigh; next shot the neighborhood children arrive.
Keep a short breath of silence after the final spoken words; no unfinished syllable at the end. Mouths move only for the written Chinese lines; no other intelligible speech, no ad-libs. Absolutely no subtitles, captions, on-screen text, Chinese characters, signboards with writing, logos or watermarks.
overall_soundscape:
Fingernails tapping, footsteps, a long sigh. Natural acoustic space of a small street-front tavern; no words beyond the written lines.
non_diegetic_music:
N/A
Shot 17
15 s · Reference to video · Pika H3 Flex ref2va (all-sparse), 768p
subject_definitions:
<Subject 1> is Kong Yiji, a tall gaunt scholar of about fifty, drawn in 2D anime style: long pale bluish-white face lined with deep wrinkles and two small healed scars on the left cheek and forehead, a scruffy unkempt grey-white beard, tired drooping eyes, the shaved Qing forehead with a thin greasy grey queue down his back; he wears a long grey-white scholar's gown (changshan) that is filthy, stained and torn at the hem and sleeves, never washed or mended, with cloth shoes worn through, as shown in <Picture 1>. This image supplies identity and costume only; its plain portrait backdrop is not the story location.
<Subject 2> is the empty Xianheng tavern in the water town of Luzhen, drawn as a 2D anime watercolor background: a large L-shaped wooden counter opens directly onto the grey flagstone street; behind the counter sit a low clay stove with a copper pot of hot water for warming wine, rows of large brown earthenware wine jars with red cloth-and-clay seals, and small tin wine pots; a dark lacquered wooden shopfront with removable plank doors, black-tiled eaves above, whitewashed walls; through a doorway at the back is a dim inner room with square tables; a small dark chalk slate hangs on a pillar by the counter covered only with faint smudges; a raised wooden threshold below the counter's street side, as shown in <Picture 2>. This image supplies the set layout, materials and colors, not a locked camera angle; it is empty and contains no people.
summary:
[reference generation] A continuous 15.0833-second 2D anime shot from an adaptation of Lu Xun's story set in a late-Qing village tavern. He gives each child one aniseed bean from his dish.
retention_analysis:
<Subject 1> (appears in [Shot 1]): fully_preserved - Keep his tall thin build, long pale wrinkled face with the small scars, scruffy grey-white beard, grey queue and the filthy torn long grey-white scholar's gown. Keep the 2D anime character design while performing the directed actions.
<Subject 2> (appears in [Shot 1]): fully_preserved - Keep the L-shaped counter facing the street, the stove and copper pot behind it, the earthenware jars, the chalk slate on the pillar, the dark shopfront and black-tiled eaves in the same spatial arrangement. Lighting follows the time of day described in this shot.
detailed_description:
Chinese 2D hand-drawn anime film style: clean confident ink-like linework, soft cel shading with two-tone shadows, hand-painted watercolor backgrounds with visible paper grain, late-Qing Shaoxing water-town atmosphere. Restrained palette of ink black, faded indigo, grey-white plaster and weathered wood, with warm amber firelight and wine steam inside the counter against cool grey-blue daylight in the street. Expressive but grounded anime acting, consistent character model sheets, stable line weight. No readable text, no Chinese characters, no subtitles, no captions, no logos, no watermarks anywhere in the frame.
Autumn afternoon: warm amber stove glow and wine steam behind the counter, cool grey-blue daylight on the street flagstones.
[Shot 1] One continuous 15.0833-second shot with no internal cut. Medium shot from the street: Kong Yiji at the counter, three small neighborhood children in simple cotton clothes and queues crowding around his knees.
The location is <Subject 2>; keep its counter, stove, jars and shopfront in their established places. Referenced characters on screen: <Subject 1>. Each referenced character appears exactly once, never duplicated; extras, if any, are simpler background figures who do not resemble the referenced characters.
He gives each child one aniseed bean from his dish. The children eat and keep staring at the dish. He covers the dish with five spread fingers, bends down and speaks; then straightens, looks at the beans, shakes his head and speaks again. The children run off laughing down the street.
<Subject 1>, in the thin, slightly hoarse, nasal voice of a man around fifty, speaking Mandarin Chinese in a slow pedantic drawl with old-fashioned bookish cadence, rising and stammering when embarrassed (S2), says flustered, bending down, from second four to second seven: <d>[Chinese] 不多了，我已经不多了。</d> Only this character's lips move for this line; listeners stay silent. <Subject 1>, in the thin, slightly hoarse, nasal voice of a man around fifty, speaking Mandarin Chinese in a slow pedantic drawl with old-fashioned bookish cadence, rising and stammering when embarrassed (S2), says shaking his head, pedantic, from second nine to second twelve: <d>[Chinese] 多乎哉？不多也。</d> Only this character's lips move for this line; listeners stay silent.
Animation stays smooth and on-model: stable faces, consistent line weight, cloth folds and queues moving with the body, steam and firelight drifting gently. Keep anatomy correct, with no extra limbs or duplicated props. End on him alone with the dish as the children vanish; next shot the counter goes on without him.
Keep a short breath of silence after the final spoken words; no unfinished syllable at the end. Mouths move only for the written Chinese lines; no other intelligible speech, no ad-libs. Absolutely no subtitles, captions, on-screen text, Chinese characters, signboards with writing, logos or watermarks.
overall_soundscape:
Children's giggles and quick footsteps, a bean dish set down. Natural acoustic space of a small street-front tavern; no words beyond the written lines.
non_diegetic_music:
N/A
Shot 18
12 s · Reference to video · Pika H3 Flex ref2va (all-sparse), 768p
subject_definitions:
<Subject 1> is the empty Xianheng tavern in the water town of Luzhen, drawn as a 2D anime watercolor background: a large L-shaped wooden counter opens directly onto the grey flagstone street; behind the counter sit a low clay stove with a copper pot of hot water for warming wine, rows of large brown earthenware wine jars with red cloth-and-clay seals, and small tin wine pots; a dark lacquered wooden shopfront with removable plank doors, black-tiled eaves above, whitewashed walls; through a doorway at the back is a dim inner room with square tables; a small dark chalk slate hangs on a pillar by the counter covered only with faint smudges; a raised wooden threshold below the counter's street side, as shown in <Picture 1>. This image supplies the set layout, materials and colors, not a locked camera angle; it is empty and contains no people.
summary:
[reference generation] A continuous 12.2500-second 2D anime shot from an adaptation of Lu Xun's story set in a late-Qing village tavern. Laborers drink as usual at the near end, laughing silently among themselves, lips closed.
retention_analysis:
<Subject 1> (appears in [Shot 1]): fully_preserved - Keep the L-shaped counter facing the street, the stove and copper pot behind it, the earthenware jars, the chalk slate on the pillar, the dark shopfront and black-tiled eaves in the same spatial arrangement. Lighting follows the time of day described in this shot.
detailed_description:
Chinese 2D hand-drawn anime film style: clean confident ink-like linework, soft cel shading with two-tone shadows, hand-painted watercolor backgrounds with visible paper grain, late-Qing Shaoxing water-town atmosphere. Restrained palette of ink black, faded indigo, grey-white plaster and weathered wood, with warm amber firelight and wine steam inside the counter against cool grey-blue daylight in the street. Expressive but grounded anime acting, consistent character model sheets, stable line weight. No readable text, no Chinese characters, no subtitles, no captions, no logos, no watermarks anywhere in the frame.
Autumn afternoon: warm amber stove glow and wine steam behind the counter, cool grey-blue daylight on the street flagstones.
[Shot 1] One continuous 12.2500-second shot with no internal cut. Wide view of the counter from the street, matching the reference; the far end of the counter, where Kong Yiji used to stand, is empty.
The location is <Subject 1>; keep its counter, stove, jars and shopfront in their established places. Referenced characters on screen: none. Each referenced character appears exactly once, never duplicated; extras, if any, are simpler background figures who do not resemble the referenced characters.
Laborers drink as usual at the near end, laughing silently among themselves, lips closed. The empty end of the counter holds an empty bowl. The camera slowly drifts toward that empty spot.
The unseen narrator, in the calm, low, slightly weary voice of a man in his forties, speaking Mandarin Chinese with measured, reflective pacing, remembering his childhood (S1), says in an off-screen voiceover, quietly, from second one to second ten: <d>[Chinese] 孔乙己是这样的使人快活，可是没有他，别人也便这么过。</d> All visible characters keep their lips closed during the narration.
Animation stays smooth and on-model: stable faces, consistent line weight, cloth folds and queues moving with the body, steam and firelight drifting gently. Keep anatomy correct, with no extra limbs or duplicated props. End on the empty bowl; next shot time passes to mid-autumn.
Keep a short breath of silence after the final spoken words; no unfinished syllable at the end. Mouths move only for the written Chinese lines; no other intelligible speech, no ad-libs. Absolutely no subtitles, captions, on-screen text, Chinese characters, signboards with writing, logos or watermarks.
overall_soundscape:
Muted street sounds and murmurs, wind. Natural acoustic space of a small street-front tavern; no words beyond the written lines.
non_diegetic_music:
N/A
Shot 19
12 s · Reference to video · Pika H3 Flex ref2va (all-sparse), 768p
subject_definitions:
<Subject 1> is the tavern owner, a stout man in his fifties drawn in 2D anime style: round fleshy face with heavy jowls, small sharp narrow eyes, a thin drooping black moustache, a black round skullcap, a long black queue, a dark brown cotton long gown under a black sleeveless waistcoat, often holding an abacus, as shown in <Picture 1>. This image supplies identity and costume only; its plain portrait backdrop is not the story location.
<Subject 2> is the interior of the same tavern seen from behind the counter: the worn wooden countertop across the foreground, the clay stove with copper hot-water pot and tin wine pots, dark earthenware wine jars with red seals, the open shopfront and the grey flagstone street with whitewashed walls and black-tiled roofs beyond, as shown in <Picture 2>. This image supplies the set layout, materials and colors, not a locked camera angle; it is empty and contains no people.
summary:
[reference generation] A continuous 12.2500-second 2D anime shot from an adaptation of Lu Xun's story set in a late-Qing village tavern. The owner flicks abacus beads, then pauses, takes down the chalk slate and squints at the faint smudges on it (no readable writing), and speaks to himself..
retention_analysis:
<Subject 1> (appears in [Shot 1]): fully_preserved - Keep his stout build, round fleshy face, small narrow eyes, thin moustache, black skullcap and the dark brown gown with black waistcoat. Keep the 2D anime character design while performing the directed actions.
<Subject 2> (appears in [Shot 1]): fully_preserved - Keep the L-shaped counter facing the street, the stove and copper pot behind it, the earthenware jars, the chalk slate on the pillar, the dark shopfront and black-tiled eaves in the same spatial arrangement. Lighting follows the time of day described in this shot.
detailed_description:
Chinese 2D hand-drawn anime film style: clean confident ink-like linework, soft cel shading with two-tone shadows, hand-painted watercolor backgrounds with visible paper grain, late-Qing Shaoxing water-town atmosphere. Restrained palette of ink black, faded indigo, grey-white plaster and weathered wood, with warm amber firelight and wine steam inside the counter against cool grey-blue daylight in the street. Expressive but grounded anime acting, consistent character model sheets, stable line weight. No readable text, no Chinese characters, no subtitles, no captions, no logos, no watermarks anywhere in the frame.
Mid-autumn evening: an oil lamp beside the counter casts a small warm pool, the street outside sinks into deep blue dusk, colors slightly desaturated.
[Shot 1] One continuous 12.2500-second shot with no internal cut. Medium shot inside behind the counter at dusk: the owner beside the pillar with the dark chalk slate, an abacus on the counter.
The location is <Subject 2>; keep its counter, stove, jars and shopfront in their established places. Referenced characters on screen: <Subject 1>. Each referenced character appears exactly once, never duplicated; extras, if any, are simpler background figures who do not resemble the referenced characters.
The owner flicks abacus beads, then pauses, takes down the chalk slate and squints at the faint smudges on it (no readable writing), and speaks to himself.
<Subject 1>, in the thick, deep, gruff voice of a stout man in his fifties, speaking Mandarin Chinese bluntly with a mocking edge (S4), says thoughtfully, then grumbling, from second three to second nine: <d>[Chinese] 孔乙己长久没有来了。还欠十九个钱呢！</d> Only this character's lips move for this line; listeners stay silent.
Animation stays smooth and on-model: stable faces, consistent line weight, cloth folds and queues moving with the body, steam and firelight drifting gently. Keep anatomy correct, with no extra limbs or duplicated props. End on the owner holding the slate; next shot a drinker answers.
Keep a short breath of silence after the final spoken words; no unfinished syllable at the end. Mouths move only for the written Chinese lines; no other intelligible speech, no ad-libs. Absolutely no subtitles, captions, on-screen text, Chinese characters, signboards with writing, logos or watermarks.
overall_soundscape:
Abacus beads clacking, evening insects, a lamp wick sputtering. Natural acoustic space of a small street-front tavern; no words beyond the written lines.
non_diegetic_music:
N/A
Shot 20
12 s · Reference to video · Pika H3 Flex ref2va (all-sparse), 768p
subject_definitions:
<Subject 1> is the tavern owner, a stout man in his fifties drawn in 2D anime style: round fleshy face with heavy jowls, small sharp narrow eyes, a thin drooping black moustache, a black round skullcap, a long black queue, a dark brown cotton long gown under a black sleeveless waistcoat, often holding an abacus, as shown in <Picture 1>. This image supplies identity and costume only; its plain portrait backdrop is not the story location.
<Subject 2> is the empty Xianheng tavern in the water town of Luzhen, drawn as a 2D anime watercolor background: a large L-shaped wooden counter opens directly onto the grey flagstone street; behind the counter sit a low clay stove with a copper pot of hot water for warming wine, rows of large brown earthenware wine jars with red cloth-and-clay seals, and small tin wine pots; a dark lacquered wooden shopfront with removable plank doors, black-tiled eaves above, whitewashed walls; through a doorway at the back is a dim inner room with square tables; a small dark chalk slate hangs on a pillar by the counter covered only with faint smudges; a raised wooden threshold below the counter's street side, as shown in <Picture 2>. This image supplies the set layout, materials and colors, not a locked camera angle; it is empty and contains no people.
summary:
[reference generation] A continuous 12.2500-second 2D anime shot from an adaptation of Lu Xun's story set in a late-Qing village tavern. The wiry laborer puts down his bowl and answers offhandedly.
retention_analysis:
<Subject 1> (appears in [Shot 1]): fully_preserved - Keep his stout build, round fleshy face, small narrow eyes, thin moustache, black skullcap and the dark brown gown with black waistcoat. Keep the 2D anime character design while performing the directed actions.
<Subject 2> (appears in [Shot 1]): fully_preserved - Keep the L-shaped counter facing the street, the stove and copper pot behind it, the earthenware jars, the chalk slate on the pillar, the dark shopfront and black-tiled eaves in the same spatial arrangement. Lighting follows the time of day described in this shot.
detailed_description:
Chinese 2D hand-drawn anime film style: clean confident ink-like linework, soft cel shading with two-tone shadows, hand-painted watercolor backgrounds with visible paper grain, late-Qing Shaoxing water-town atmosphere. Restrained palette of ink black, faded indigo, grey-white plaster and weathered wood, with warm amber firelight and wine steam inside the counter against cool grey-blue daylight in the street. Expressive but grounded anime acting, consistent character model sheets, stable line weight. No readable text, no Chinese characters, no subtitles, no captions, no logos, no watermarks anywhere in the frame.
Mid-autumn evening: an oil lamp beside the counter casts a small warm pool, the street outside sinks into deep blue dusk, colors slightly desaturated.
[Shot 1] One continuous 12.2500-second shot with no internal cut. Medium two-shot across the counter at dusk: the owner behind the counter on the right, a wiry laborer in his forties with a thin moustache, wearing a faded brown short jacket and a straw-rope belt drinking on the street side on the left.
The location is <Subject 2>; keep its counter, stove, jars and shopfront in their established places. Referenced characters on screen: <Subject 1>. Each referenced character appears exactly once, never duplicated; extras, if any, are simpler background figures who do not resemble the referenced characters.
The wiry laborer puts down his bowl and answers offhandedly. The owner raises his eyebrows, surprised, and asks.
A wiry laborer in his forties with a thin moustache, wearing a faded brown short jacket and a straw-rope belt, in the raspy, mid-pitched voice of a wiry laborer in his forties, speaking Mandarin Chinese flatly and offhandedly (S6), says flatly, from second one to second five: <d>[Chinese] 他怎么会来？……他打折了腿了。</d> Only this speaker's lips move for this line. <Subject 1>, in the thick, deep, gruff voice of a stout man in his fifties, speaking Mandarin Chinese bluntly with a mocking edge (S4), says surprised, from second seven to second nine: <d>[Chinese] 哦！后来呢？</d> Only this character's lips move for this line; listeners stay silent.
Animation stays smooth and on-model: stable faces, consistent line weight, cloth folds and queues moving with the body, steam and firelight drifting gently. Keep anatomy correct, with no extra limbs or duplicated props. End on the owner leaning in; next shot the laborer tells the story.
Keep a short breath of silence after the final spoken words; no unfinished syllable at the end. Mouths move only for the written Chinese lines; no other intelligible speech, no ad-libs. Absolutely no subtitles, captions, on-screen text, Chinese characters, signboards with writing, logos or watermarks.
overall_soundscape:
A bowl set down, evening insects. Natural acoustic space of a small street-front tavern; no words beyond the written lines.
non_diegetic_music:
N/A
Shot 21
15 s · Reference to video · Pika H3 Flex ref2va (all-sparse), 768p
subject_definitions:
<Subject 1> is the tavern owner, a stout man in his fifties drawn in 2D anime style: round fleshy face with heavy jowls, small sharp narrow eyes, a thin drooping black moustache, a black round skullcap, a long black queue, a dark brown cotton long gown under a black sleeveless waistcoat, often holding an abacus, as shown in <Picture 1>. This image supplies identity and costume only; its plain portrait backdrop is not the story location.
<Subject 2> is the empty Xianheng tavern in the water town of Luzhen, drawn as a 2D anime watercolor background: a large L-shaped wooden counter opens directly onto the grey flagstone street; behind the counter sit a low clay stove with a copper pot of hot water for warming wine, rows of large brown earthenware wine jars with red cloth-and-clay seals, and small tin wine pots; a dark lacquered wooden shopfront with removable plank doors, black-tiled eaves above, whitewashed walls; through a doorway at the back is a dim inner room with square tables; a small dark chalk slate hangs on a pillar by the counter covered only with faint smudges; a raised wooden threshold below the counter's street side, as shown in <Picture 2>. This image supplies the set layout, materials and colors, not a locked camera angle; it is empty and contains no people.
summary:
[reference generation] A continuous 15.0833-second 2D anime shot from an adaptation of Lu Xun's story set in a late-Qing village tavern. The laborer tells it casually between sips, miming a beating with his hand.
retention_analysis:
<Subject 1> (appears in [Shot 1]): fully_preserved - Keep his stout build, round fleshy face, small narrow eyes, thin moustache, black skullcap and the dark brown gown with black waistcoat. Keep the 2D anime character design while performing the directed actions.
<Subject 2> (appears in [Shot 1]): fully_preserved - Keep the L-shaped counter facing the street, the stove and copper pot behind it, the earthenware jars, the chalk slate on the pillar, the dark shopfront and black-tiled eaves in the same spatial arrangement. Lighting follows the time of day described in this shot.
detailed_description:
Chinese 2D hand-drawn anime film style: clean confident ink-like linework, soft cel shading with two-tone shadows, hand-painted watercolor backgrounds with visible paper grain, late-Qing Shaoxing water-town atmosphere. Restrained palette of ink black, faded indigo, grey-white plaster and weathered wood, with warm amber firelight and wine steam inside the counter against cool grey-blue daylight in the street. Expressive but grounded anime acting, consistent character model sheets, stable line weight. No readable text, no Chinese characters, no subtitles, no captions, no logos, no watermarks anywhere in the frame.
Mid-autumn evening: an oil lamp beside the counter casts a small warm pool, the street outside sinks into deep blue dusk, colors slightly desaturated.
[Shot 1] One continuous 15.0833-second shot with no internal cut. Close medium shot on a wiry laborer in his forties with a thin moustache, wearing a faded brown short jacket and a straw-rope belt at the counter in lamplight, the owner on the right edge.
The location is <Subject 2>; keep its counter, stove, jars and shopfront in their established places. Referenced characters on screen: <Subject 1>. Each referenced character appears exactly once, never duplicated; extras, if any, are simpler background figures who do not resemble the referenced characters.
The laborer tells it casually between sips, miming a beating with his hand. The owner asks again; the laborer shrugs.
A wiry laborer in his forties with a thin moustache, wearing a faded brown short jacket and a straw-rope belt, in the raspy, mid-pitched voice of a wiry laborer in his forties, speaking Mandarin Chinese flatly and offhandedly (S6), says casually, from second zero to second seven: <d>[Chinese] 偷到丁举人家里去了。打了大半夜，再打折了腿。</d> Only this speaker's lips move for this line. <Subject 1>, in the thick, deep, gruff voice of a stout man in his fifties, speaking Mandarin Chinese bluntly with a mocking edge (S4), says from second eight to second ten: <d>[Chinese] 后来怎么样？</d> Only this character's lips move for this line; listeners stay silent. A wiry laborer in his forties with a thin moustache, wearing a faded brown short jacket and a straw-rope belt, in the raspy, mid-pitched voice of a wiry laborer in his forties, speaking Mandarin Chinese flatly and offhandedly (S6), says with a shrug, from second eleven to second fourteen: <d>[Chinese] 谁晓得？许是死了。</d> Only this speaker's lips move for this line.
Animation stays smooth and on-model: stable faces, consistent line weight, cloth folds and queues moving with the body, steam and firelight drifting gently. Keep anatomy correct, with no extra limbs or duplicated props. End on the shrug; next shot the owner goes back to his accounts.
Keep a short breath of silence after the final spoken words; no unfinished syllable at the end. Mouths move only for the written Chinese lines; no other intelligible speech, no ad-libs. Absolutely no subtitles, captions, on-screen text, Chinese characters, signboards with writing, logos or watermarks.
overall_soundscape:
Lamp sputter, a sip. Natural acoustic space of a small street-front tavern; no words beyond the written lines.
non_diegetic_music:
N/A
Shot 22
10 s · Reference to video · Pika H3 Flex ref2va (all-sparse), 768p
subject_definitions:
<Subject 1> is the tavern owner, a stout man in his fifties drawn in 2D anime style: round fleshy face with heavy jowls, small sharp narrow eyes, a thin drooping black moustache, a black round skullcap, a long black queue, a dark brown cotton long gown under a black sleeveless waistcoat, often holding an abacus, as shown in <Picture 1>. This image supplies identity and costume only; its plain portrait backdrop is not the story location.
<Subject 2> is the interior of the same tavern seen from behind the counter: the worn wooden countertop across the foreground, the clay stove with copper hot-water pot and tin wine pots, dark earthenware wine jars with red seals, the open shopfront and the grey flagstone street with whitewashed walls and black-tiled roofs beyond, as shown in <Picture 2>. This image supplies the set layout, materials and colors, not a locked camera angle; it is empty and contains no people.
summary:
[reference generation] A continuous 10.1250-second 2D anime shot from an adaptation of Lu Xun's story set in a late-Qing village tavern. The owner no longer asks; he hangs the slate back on the pillar and slowly flicks abacus beads.
retention_analysis:
<Subject 1> (appears in [Shot 1]): fully_preserved - Keep his stout build, round fleshy face, small narrow eyes, thin moustache, black skullcap and the dark brown gown with black waistcoat. Keep the 2D anime character design while performing the directed actions.
<Subject 2> (appears in [Shot 1]): fully_preserved - Keep the L-shaped counter facing the street, the stove and copper pot behind it, the earthenware jars, the chalk slate on the pillar, the dark shopfront and black-tiled eaves in the same spatial arrangement. Lighting follows the time of day described in this shot.
detailed_description:
Chinese 2D hand-drawn anime film style: clean confident ink-like linework, soft cel shading with two-tone shadows, hand-painted watercolor backgrounds with visible paper grain, late-Qing Shaoxing water-town atmosphere. Restrained palette of ink black, faded indigo, grey-white plaster and weathered wood, with warm amber firelight and wine steam inside the counter against cool grey-blue daylight in the street. Expressive but grounded anime acting, consistent character model sheets, stable line weight. No readable text, no Chinese characters, no subtitles, no captions, no logos, no watermarks anywhere in the frame.
Mid-autumn evening: an oil lamp beside the counter casts a small warm pool, the street outside sinks into deep blue dusk, colors slightly desaturated.
[Shot 1] One continuous 10.1250-second shot with no internal cut. Medium shot of the owner behind the counter in lamplight, the dark street beyond.
The location is <Subject 2>; keep its counter, stove, jars and shopfront in their established places. Referenced characters on screen: <Subject 1>. Each referenced character appears exactly once, never duplicated; extras, if any, are simpler background figures who do not resemble the referenced characters.
The owner no longer asks; he hangs the slate back on the pillar and slowly flicks abacus beads. Lips closed while the narrator speaks.
The unseen narrator, in the calm, low, slightly weary voice of a man in his forties, speaking Mandarin Chinese with measured, reflective pacing, remembering his childhood (S1), says in an off-screen voiceover, quietly, from second one to second seven: <d>[Chinese] 掌柜也不再问，仍然慢慢的算他的账。</d> All visible characters keep their lips closed during the narration.
Animation stays smooth and on-model: stable faces, consistent line weight, cloth folds and queues moving with the body, steam and firelight drifting gently. Keep anatomy correct, with no extra limbs or duplicated props. End on the lamp flame; next shot the season turns cold.
Keep a short breath of silence after the final spoken words; no unfinished syllable at the end. Mouths move only for the written Chinese lines; no other intelligible speech, no ad-libs. Absolutely no subtitles, captions, on-screen text, Chinese characters, signboards with writing, logos or watermarks.
overall_soundscape:
Abacus beads slowly clacking. Natural acoustic space of a small street-front tavern; no words beyond the written lines.
non_diegetic_music:
N/A
Shot 23
12 s · Reference to video · Pika H3 Flex ref2va (all-sparse), 768p
subject_definitions:
<Subject 1> is the young waiter, a thin twelve-year-old boy drawn in 2D anime style: round boyish face, large dark attentive eyes, shaved Qing forehead with a short black queue, a plain faded indigo cotton short jacket and trousers, a stained off-white cloth apron tied at the waist, black cloth shoes, as shown in <Picture 1>. This image supplies identity and costume only; its plain portrait backdrop is not the story location.
<Subject 2> is the interior of the same tavern seen from behind the counter: the worn wooden countertop across the foreground, the clay stove with copper hot-water pot and tin wine pots, dark earthenware wine jars with red seals, the open shopfront and the grey flagstone street with whitewashed walls and black-tiled roofs beyond, as shown in <Picture 2>. This image supplies the set layout, materials and colors, not a locked camera angle; it is empty and contains no people.
summary:
[reference generation] A continuous 12.2500-second 2D anime shot from an adaptation of Lu Xun's story set in a late-Qing village tavern. Dry leaves blow across the empty street.
retention_analysis:
<Subject 1> (appears in [Shot 1]): fully_preserved - Keep his twelve-year-old face and size, short black queue, faded indigo short jacket and the off-white apron. Keep the 2D anime character design while performing the directed actions.
<Subject 2> (appears in [Shot 1]): fully_preserved - Keep the L-shaped counter facing the street, the stove and copper pot behind it, the earthenware jars, the chalk slate on the pillar, the dark shopfront and black-tiled eaves in the same spatial arrangement. Lighting follows the time of day described in this shot.
detailed_description:
Chinese 2D hand-drawn anime film style: clean confident ink-like linework, soft cel shading with two-tone shadows, hand-painted watercolor backgrounds with visible paper grain, late-Qing Shaoxing water-town atmosphere. Restrained palette of ink black, faded indigo, grey-white plaster and weathered wood, with warm amber firelight and wine steam inside the counter against cool grey-blue daylight in the street. Expressive but grounded anime acting, consistent character model sheets, stable line weight. No readable text, no Chinese characters, no subtitles, no captions, no logos, no watermarks anywhere in the frame.
Early winter, overcast and cold: the street is washed in pale desaturated grey-blue, dry leaves and grit blow in a thin wind, only the stove fire inside keeps a small warm orange glow; breath is faintly visible.
[Shot 1] One continuous 12.2500-second shot with no internal cut. Medium shot from inside behind the counter, the cold grey street beyond; the boy beside the stove.
The location is <Subject 2>; keep its counter, stove, jars and shopfront in their established places. Referenced characters on screen: <Subject 1>. Each referenced character appears exactly once, never duplicated; extras, if any, are simpler background figures who do not resemble the referenced characters.
Dry leaves blow across the empty street. The boy huddles close to the stove, hands held out to the fire, and shivers. Lips closed during the narration.
The unseen narrator, in the calm, low, slightly weary voice of a man in his forties, speaking Mandarin Chinese with measured, reflective pacing, remembering his childhood (S1), says in an off-screen voiceover, slowly, from second one to second ten: <d>[Chinese] 中秋过后，秋风是一天凉比一天，看看将近初冬。</d> All visible characters keep their lips closed during the narration.
Animation stays smooth and on-model: stable faces, consistent line weight, cloth folds and queues moving with the body, steam and firelight drifting gently. Keep anatomy correct, with no extra limbs or duplicated props. End on the boy dozing by the fire; next shot a voice comes from below the counter.
Keep a short breath of silence after the final spoken words; no unfinished syllable at the end. Mouths move only for the written Chinese lines; no other intelligible speech, no ad-libs. Absolutely no subtitles, captions, on-screen text, Chinese characters, signboards with writing, logos or watermarks.
overall_soundscape:
Cold wind whistling, leaves scraping on stone, fire crackle. Natural acoustic space of a small street-front tavern; no words beyond the written lines.
non_diegetic_music:
A single soft low guqin note sustained, very quiet.
Shot 24
10 s · Reference to video · Pika H3 Flex ref2va (all-sparse), 768p
subject_definitions:
<Subject 1> is the young waiter, a thin twelve-year-old boy drawn in 2D anime style: round boyish face, large dark attentive eyes, shaved Qing forehead with a short black queue, a plain faded indigo cotton short jacket and trousers, a stained off-white cloth apron tied at the waist, black cloth shoes, as shown in <Picture 1>. This image supplies identity and costume only; its plain portrait backdrop is not the story location.
<Subject 2> is the interior of the same tavern seen from behind the counter: the worn wooden countertop across the foreground, the clay stove with copper hot-water pot and tin wine pots, dark earthenware wine jars with red seals, the open shopfront and the grey flagstone street with whitewashed walls and black-tiled roofs beyond, as shown in <Picture 2>. This image supplies the set layout, materials and colors, not a locked camera angle; it is empty and contains no people.
summary:
[reference generation] A continuous 10.1250-second 2D anime shot from an adaptation of Lu Xun's story set in a late-Qing village tavern. A weak voice comes from somewhere below the counter on the street side; the speaker is not visible.
retention_analysis:
<Subject 1> (appears in [Shot 1]): fully_preserved - Keep his twelve-year-old face and size, short black queue, faded indigo short jacket and the off-white apron. Keep the 2D anime character design while performing the directed actions.
<Subject 2> (appears in [Shot 1]): fully_preserved - Keep the L-shaped counter facing the street, the stove and copper pot behind it, the earthenware jars, the chalk slate on the pillar, the dark shopfront and black-tiled eaves in the same spatial arrangement. Lighting follows the time of day described in this shot.
detailed_description:
Chinese 2D hand-drawn anime film style: clean confident ink-like linework, soft cel shading with two-tone shadows, hand-painted watercolor backgrounds with visible paper grain, late-Qing Shaoxing water-town atmosphere. Restrained palette of ink black, faded indigo, grey-white plaster and weathered wood, with warm amber firelight and wine steam inside the counter against cool grey-blue daylight in the street. Expressive but grounded anime acting, consistent character model sheets, stable line weight. No readable text, no Chinese characters, no subtitles, no captions, no logos, no watermarks anywhere in the frame.
Early winter, overcast and cold: the street is washed in pale desaturated grey-blue, dry leaves and grit blow in a thin wind, only the stove fire inside keeps a small warm orange glow; breath is faintly visible.
[Shot 1] One continuous 10.1250-second shot with no internal cut. Medium shot of the boy behind the counter, looking out over the empty countertop toward the street.
The location is <Subject 2>; keep its counter, stove, jars and shopfront in their established places. Referenced characters on screen: <Subject 1>. Each referenced character appears exactly once, never duplicated; extras, if any, are simpler background figures who do not resemble the referenced characters.
A weak voice comes from somewhere below the counter on the street side; the speaker is not visible. The boy lifts his head, looks out over the counter, sees no one at first, and frowns, puzzled. His lips stay closed.
An unseen man below the counter, in the thin, slightly hoarse, nasal voice of a man around fifty, speaking Mandarin Chinese, now weak, low and breathy, pleading (S2), says from off-screen weak and hoarse, off-screen from below the counter, from second two to second four: <d>[Chinese] 温一碗酒。</d> Only this speaker's lips move for this line.
Animation stays smooth and on-model: stable faces, consistent line weight, cloth folds and queues moving with the body, steam and firelight drifting gently. Keep anatomy correct, with no extra limbs or duplicated props. End on the boy standing on tiptoe to look down over the counter; next shot reveals Kong Yiji.
Keep a short breath of silence after the final spoken words; no unfinished syllable at the end. Mouths move only for the written Chinese lines; no other intelligible speech, no ad-libs. Absolutely no subtitles, captions, on-screen text, Chinese characters, signboards with writing, logos or watermarks.
overall_soundscape:
Wind, a faint scrape on stone. Natural acoustic space of a small street-front tavern; no words beyond the written lines.
non_diegetic_music:
N/A
Shot 25
15 s · Reference to video · Pika H3 Flex ref2va (all-sparse), 768p
subject_definitions:
<Subject 1> is Kong Yiji in his last appearance, drawn in 2D anime style: the same tall gaunt scholar of about fifty with the same scruffy grey-white beard and grey queue, but his face now darkened, hollow and much thinner; he no longer wears the long gown but a ragged, patched, dirty dark-grey lined short jacket; his broken legs are folded uselessly beneath him and he sits on a flattened woven cattail-straw mat bag, which hangs from his shoulders by a straw rope; his hands are caked with mud, as shown in <Picture 1>. This image supplies identity and costume only; its plain portrait backdrop is not the story location.
<Subject 2> is the tavern owner, a stout man in his fifties drawn in 2D anime style: round fleshy face with heavy jowls, small sharp narrow eyes, a thin drooping black moustache, a black round skullcap, a long black queue, a dark brown cotton long gown under a black sleeveless waistcoat, often holding an abacus, as shown in <Picture 2>. This image supplies identity and costume only; its plain portrait backdrop is not the story location.
<Subject 3> is the empty Xianheng tavern in the water town of Luzhen, drawn as a 2D anime watercolor background: a large L-shaped wooden counter opens directly onto the grey flagstone street; behind the counter sit a low clay stove with a copper pot of hot water for warming wine, rows of large brown earthenware wine jars with red cloth-and-clay seals, and small tin wine pots; a dark lacquered wooden shopfront with removable plank doors, black-tiled eaves above, whitewashed walls; through a doorway at the back is a dim inner room with square tables; a small dark chalk slate hangs on a pillar by the counter covered only with faint smudges; a raised wooden threshold below the counter's street side, as shown in <Picture 3>. This image supplies the set layout, materials and colors, not a locked camera angle; it is empty and contains no people.
summary:
[reference generation] A continuous 15.0833-second 2D anime shot from an adaptation of Lu Xun's story set in a late-Qing village tavern. Kong Yiji sits on his straw mat at the foot of the counter, legs folded beneath him, looking up, thin and dark-faced.
retention_analysis:
<Subject 1> (appears in [Shot 1]): fully_preserved - Keep his face identity, grey-white beard and queue, the darkened thin face, the ragged dark-grey lined short jacket, the straw mat he sits on with the straw rope over his shoulders, and the muddy hands; his legs stay folded beneath him and he never stands. Keep the 2D anime character design while performing the directed actions.
<Subject 2> (appears in [Shot 1]): fully_preserved - Keep his stout build, round fleshy face, small narrow eyes, thin moustache, black skullcap and the dark brown gown with black waistcoat. Keep the 2D anime character design while performing the directed actions.
<Subject 3> (appears in [Shot 1]): fully_preserved - Keep the L-shaped counter facing the street, the stove and copper pot behind it, the earthenware jars, the chalk slate on the pillar, the dark shopfront and black-tiled eaves in the same spatial arrangement. Lighting follows the time of day described in this shot.
detailed_description:
Chinese 2D hand-drawn anime film style: clean confident ink-like linework, soft cel shading with two-tone shadows, hand-painted watercolor backgrounds with visible paper grain, late-Qing Shaoxing water-town atmosphere. Restrained palette of ink black, faded indigo, grey-white plaster and weathered wood, with warm amber firelight and wine steam inside the counter against cool grey-blue daylight in the street. Expressive but grounded anime acting, consistent character model sheets, stable line weight. No readable text, no Chinese characters, no subtitles, no captions, no logos, no watermarks anywhere in the frame.
Early winter, overcast and cold: the street is washed in pale desaturated grey-blue, dry leaves and grit blow in a thin wind, only the stove fire inside keeps a small warm orange glow; breath is faintly visible.
[Shot 1] One continuous 15.0833-second shot with no internal cut. High angle looking down over the counter edge onto the threshold below: Kong Yiji sitting on the ground in front of the counter, the owner leaning out over the counter at the top right.
The location is <Subject 3>; keep its counter, stove, jars and shopfront in their established places. Referenced characters on screen: <Subject 1>, <Subject 2>. Each referenced character appears exactly once, never duplicated; extras, if any, are simpler background figures who do not resemble the referenced characters.
Kong Yiji sits on his straw mat at the foot of the counter, legs folded beneath him, looking up, thin and dark-faced. He repeats his request. The owner leans out over the counter, peers down and speaks with mocking surprise.
<Subject 1>, in the thin, slightly hoarse, nasal voice of a man around fifty, speaking Mandarin Chinese, now weak, low and breathy, pleading (S2), says weakly, from second two to second four: <d>[Chinese] 温一碗酒。</d> Only this character's lips move for this line; listeners stay silent. <Subject 2>, in the thick, deep, gruff voice of a stout man in his fifties, speaking Mandarin Chinese bluntly with a mocking edge (S4), says mocking surprise, from second seven to second eleven: <d>[Chinese] 孔乙己么？你还欠十九个钱呢！</d> Only this character's lips move for this line; listeners stay silent.
Animation stays smooth and on-model: stable faces, consistent line weight, cloth folds and queues moving with the body, steam and firelight drifting gently. Keep anatomy correct, with no extra limbs or duplicated props. End on Kong Yiji's dejected upward look; next shot he answers.
Keep a short breath of silence after the final spoken words; no unfinished syllable at the end. Mouths move only for the written Chinese lines; no other intelligible speech, no ad-libs. Absolutely no subtitles, captions, on-screen text, Chinese characters, signboards with writing, logos or watermarks.
overall_soundscape:
Cold wind, a straw mat rustling. Natural acoustic space of a small street-front tavern; no words beyond the written lines.
non_diegetic_music:
A single soft low guqin note sustained, very quiet.
Shot 26
15 s · Reference to video · Pika H3 Flex ref2va (all-sparse), 768p
subject_definitions:
<Subject 1> is Kong Yiji in his last appearance, drawn in 2D anime style: the same tall gaunt scholar of about fifty with the same scruffy grey-white beard and grey queue, but his face now darkened, hollow and much thinner; he no longer wears the long gown but a ragged, patched, dirty dark-grey lined short jacket; his broken legs are folded uselessly beneath him and he sits on a flattened woven cattail-straw mat bag, which hangs from his shoulders by a straw rope; his hands are caked with mud, as shown in <Picture 1>. This image supplies identity and costume only; its plain portrait backdrop is not the story location.
<Subject 2> is the tavern owner, a stout man in his fifties drawn in 2D anime style: round fleshy face with heavy jowls, small sharp narrow eyes, a thin drooping black moustache, a black round skullcap, a long black queue, a dark brown cotton long gown under a black sleeveless waistcoat, often holding an abacus, as shown in <Picture 2>. This image supplies identity and costume only; its plain portrait backdrop is not the story location.
<Subject 3> is the empty Xianheng tavern in the water town of Luzhen, drawn as a 2D anime watercolor background: a large L-shaped wooden counter opens directly onto the grey flagstone street; behind the counter sit a low clay stove with a copper pot of hot water for warming wine, rows of large brown earthenware wine jars with red cloth-and-clay seals, and small tin wine pots; a dark lacquered wooden shopfront with removable plank doors, black-tiled eaves above, whitewashed walls; through a doorway at the back is a dim inner room with square tables; a small dark chalk slate hangs on a pillar by the counter covered only with faint smudges; a raised wooden threshold below the counter's street side, as shown in <Picture 3>. This image supplies the set layout, materials and colors, not a locked camera angle; it is empty and contains no people.
summary:
[reference generation] A continuous 15.0833-second 2D anime shot from an adaptation of Lu Xun's story set in a late-Qing village tavern. Kong Yiji looks up, very dejected, and answers haltingly.
retention_analysis:
<Subject 1> (appears in [Shot 1]): fully_preserved - Keep his face identity, grey-white beard and queue, the darkened thin face, the ragged dark-grey lined short jacket, the straw mat he sits on with the straw rope over his shoulders, and the muddy hands; his legs stay folded beneath him and he never stands. Keep the 2D anime character design while performing the directed actions.
<Subject 2> (appears in [Shot 1]): fully_preserved - Keep his stout build, round fleshy face, small narrow eyes, thin moustache, black skullcap and the dark brown gown with black waistcoat. Keep the 2D anime character design while performing the directed actions.
<Subject 3> (appears in [Shot 1]): fully_preserved - Keep the L-shaped counter facing the street, the stove and copper pot behind it, the earthenware jars, the chalk slate on the pillar, the dark shopfront and black-tiled eaves in the same spatial arrangement. Lighting follows the time of day described in this shot.
detailed_description:
Chinese 2D hand-drawn anime film style: clean confident ink-like linework, soft cel shading with two-tone shadows, hand-painted watercolor backgrounds with visible paper grain, late-Qing Shaoxing water-town atmosphere. Restrained palette of ink black, faded indigo, grey-white plaster and weathered wood, with warm amber firelight and wine steam inside the counter against cool grey-blue daylight in the street. Expressive but grounded anime acting, consistent character model sheets, stable line weight. No readable text, no Chinese characters, no subtitles, no captions, no logos, no watermarks anywhere in the frame.
Early winter, overcast and cold: the street is washed in pale desaturated grey-blue, dry leaves and grit blow in a thin wind, only the stove fire inside keeps a small warm orange glow; breath is faintly visible.
[Shot 1] One continuous 15.0833-second shot with no internal cut. Low angle from the street level beside Kong Yiji: he sits on the mat in the foreground left, looking up; the owner leans over the counter above on the right.
The location is <Subject 3>; keep its counter, stove, jars and shopfront in their established places. Referenced characters on screen: <Subject 1>, <Subject 2>. Each referenced character appears exactly once, never duplicated; extras, if any, are simpler background figures who do not resemble the referenced characters.
Kong Yiji looks up, very dejected, and answers haltingly. The owner smiles a broad smile and teases him, as usual.
<Subject 1>, in the thin, slightly hoarse, nasal voice of a man around fifty, speaking Mandarin Chinese, now weak, low and breathy, pleading (S2), says dejected, halting, from second one to second eight: <d>[Chinese] 这……下回还清罢。这一回是现钱，酒要好。</d> Only this character's lips move for this line; listeners stay silent. <Subject 2>, in the thick, deep, gruff voice of a stout man in his fifties, speaking Mandarin Chinese bluntly with a mocking edge (S4), says grinning, from second ten to second thirteen: <d>[Chinese] 孔乙己，你又偷了东西了！</d> Only this character's lips move for this line; listeners stay silent.
Animation stays smooth and on-model: stable faces, consistent line weight, cloth folds and queues moving with the body, steam and firelight drifting gently. Keep anatomy correct, with no extra limbs or duplicated props. End on the owner's grin; next shot Kong Yiji pleads.
Keep a short breath of silence after the final spoken words; no unfinished syllable at the end. Mouths move only for the written Chinese lines; no other intelligible speech, no ad-libs. Absolutely no subtitles, captions, on-screen text, Chinese characters, signboards with writing, logos or watermarks.
overall_soundscape:
Wind, a chuckle from a laborer off-screen. Natural acoustic space of a small street-front tavern; no words beyond the written lines.
non_diegetic_music:
N/A
Shot 27
15 s · Reference to video · Pika H3 Flex ref2va (all-sparse), 768p
subject_definitions:
<Subject 1> is Kong Yiji in his last appearance, drawn in 2D anime style: the same tall gaunt scholar of about fifty with the same scruffy grey-white beard and grey queue, but his face now darkened, hollow and much thinner; he no longer wears the long gown but a ragged, patched, dirty dark-grey lined short jacket; his broken legs are folded uselessly beneath him and he sits on a flattened woven cattail-straw mat bag, which hangs from his shoulders by a straw rope; his hands are caked with mud, as shown in <Picture 1>. This image supplies identity and costume only; its plain portrait backdrop is not the story location.
<Subject 2> is the tavern owner, a stout man in his fifties drawn in 2D anime style: round fleshy face with heavy jowls, small sharp narrow eyes, a thin drooping black moustache, a black round skullcap, a long black queue, a dark brown cotton long gown under a black sleeveless waistcoat, often holding an abacus, as shown in <Picture 2>. This image supplies identity and costume only; its plain portrait backdrop is not the story location.
<Subject 3> is the empty Xianheng tavern in the water town of Luzhen, drawn as a 2D anime watercolor background: a large L-shaped wooden counter opens directly onto the grey flagstone street; behind the counter sit a low clay stove with a copper pot of hot water for warming wine, rows of large brown earthenware wine jars with red cloth-and-clay seals, and small tin wine pots; a dark lacquered wooden shopfront with removable plank doors, black-tiled eaves above, whitewashed walls; through a doorway at the back is a dim inner room with square tables; a small dark chalk slate hangs on a pillar by the counter covered only with faint smudges; a raised wooden threshold below the counter's street side, as shown in <Picture 3>. This image supplies the set layout, materials and colors, not a locked camera angle; it is empty and contains no people.
summary:
[reference generation] A continuous 15.0833-second 2D anime shot from an adaptation of Lu Xun's story set in a late-Qing village tavern. Kong Yiji does not argue much, only says weakly.
retention_analysis:
<Subject 1> (appears in [Shot 1]): fully_preserved - Keep his face identity, grey-white beard and queue, the darkened thin face, the ragged dark-grey lined short jacket, the straw mat he sits on with the straw rope over his shoulders, and the muddy hands; his legs stay folded beneath him and he never stands. Keep the 2D anime character design while performing the directed actions.
<Subject 2> (appears in [Shot 1]): fully_preserved - Keep his stout build, round fleshy face, small narrow eyes, thin moustache, black skullcap and the dark brown gown with black waistcoat. Keep the 2D anime character design while performing the directed actions.
<Subject 3> (appears in [Shot 1]): fully_preserved - Keep the L-shaped counter facing the street, the stove and copper pot behind it, the earthenware jars, the chalk slate on the pillar, the dark shopfront and black-tiled eaves in the same spatial arrangement. Lighting follows the time of day described in this shot.
detailed_description:
Chinese 2D hand-drawn anime film style: clean confident ink-like linework, soft cel shading with two-tone shadows, hand-painted watercolor backgrounds with visible paper grain, late-Qing Shaoxing water-town atmosphere. Restrained palette of ink black, faded indigo, grey-white plaster and weathered wood, with warm amber firelight and wine steam inside the counter against cool grey-blue daylight in the street. Expressive but grounded anime acting, consistent character model sheets, stable line weight. No readable text, no Chinese characters, no subtitles, no captions, no logos, no watermarks anywhere in the frame.
Early winter, overcast and cold: the street is washed in pale desaturated grey-blue, dry leaves and grit blow in a thin wind, only the stove fire inside keeps a small warm orange glow; breath is faintly visible.
[Shot 1] One continuous 15.0833-second shot with no internal cut. Close shot on Kong Yiji on the mat at the threshold, the owner's face partly visible above.
The location is <Subject 3>; keep its counter, stove, jars and shopfront in their established places. Referenced characters on screen: <Subject 1>, <Subject 2>. Each referenced character appears exactly once, never duplicated; extras, if any, are simpler background figures who do not resemble the referenced characters.
Kong Yiji does not argue much, only says weakly. The owner presses. Kong Yiji lowers his voice, eyes pleading the owner to stop, stammering. Laborers off-screen laugh.
<Subject 1>, in the thin, slightly hoarse, nasal voice of a man around fifty, speaking Mandarin Chinese, now weak, low and breathy, pleading (S2), says weakly, from second zero to second two: <d>[Chinese] 不要取笑！</d> Only this character's lips move for this line; listeners stay silent. <Subject 2>, in the thick, deep, gruff voice of a stout man in his fifties, speaking Mandarin Chinese bluntly with a mocking edge (S4), says mocking, from second three to second eight: <d>[Chinese] 取笑？要是不偷，怎么会打断腿？</d> Only this character's lips move for this line; listeners stay silent. <Subject 1>, in the thin, slightly hoarse, nasal voice of a man around fifty, speaking Mandarin Chinese, now weak, low and breathy, pleading (S2), says very low, pleading, from second nine to second thirteen: <d>[Chinese] 跌断，跌，跌……</d> Only this character's lips move for this line; listeners stay silent.
Animation stays smooth and on-model: stable faces, consistent line weight, cloth folds and queues moving with the body, steam and firelight drifting gently. Keep anatomy correct, with no extra limbs or duplicated props. End on his pleading eyes; next shot the boy brings the wine.
Keep a short breath of silence after the final spoken words; no unfinished syllable at the end. Mouths move only for the written Chinese lines; no other intelligible speech, no ad-libs. Absolutely no subtitles, captions, on-screen text, Chinese characters, signboards with writing, logos or watermarks.
overall_soundscape:
Several men laughing off-screen from second thirteen, wind. Natural acoustic space of a small street-front tavern; no words beyond the written lines.
non_diegetic_music:
N/A
Shot 28
12 s · Reference to video · Pika H3 Flex ref2va (all-sparse), 768p
subject_definitions:
<Subject 1> is Kong Yiji in his last appearance, drawn in 2D anime style: the same tall gaunt scholar of about fifty with the same scruffy grey-white beard and grey queue, but his face now darkened, hollow and much thinner; he no longer wears the long gown but a ragged, patched, dirty dark-grey lined short jacket; his broken legs are folded uselessly beneath him and he sits on a flattened woven cattail-straw mat bag, which hangs from his shoulders by a straw rope; his hands are caked with mud, as shown in <Picture 1>. This image supplies identity and costume only; its plain portrait backdrop is not the story location.
<Subject 2> is the young waiter, a thin twelve-year-old boy drawn in 2D anime style: round boyish face, large dark attentive eyes, shaved Qing forehead with a short black queue, a plain faded indigo cotton short jacket and trousers, a stained off-white cloth apron tied at the waist, black cloth shoes, as shown in <Picture 2>. This image supplies identity and costume only; its plain portrait backdrop is not the story location.
<Subject 3> is the empty Xianheng tavern in the water town of Luzhen, drawn as a 2D anime watercolor background: a large L-shaped wooden counter opens directly onto the grey flagstone street; behind the counter sit a low clay stove with a copper pot of hot water for warming wine, rows of large brown earthenware wine jars with red cloth-and-clay seals, and small tin wine pots; a dark lacquered wooden shopfront with removable plank doors, black-tiled eaves above, whitewashed walls; through a doorway at the back is a dim inner room with square tables; a small dark chalk slate hangs on a pillar by the counter covered only with faint smudges; a raised wooden threshold below the counter's street side, as shown in <Picture 3>. This image supplies the set layout, materials and colors, not a locked camera angle; it is empty and contains no people.
summary:
[reference generation] A continuous 12.2500-second 2D anime shot from an adaptation of Lu Xun's story set in a late-Qing village tavern. The boy sets a bowl of warm wine on the threshold.
retention_analysis:
<Subject 1> (appears in [Shot 1]): fully_preserved - Keep his face identity, grey-white beard and queue, the darkened thin face, the ragged dark-grey lined short jacket, the straw mat he sits on with the straw rope over his shoulders, and the muddy hands; his legs stay folded beneath him and he never stands. Keep the 2D anime character design while performing the directed actions.
<Subject 2> (appears in [Shot 1]): fully_preserved - Keep his twelve-year-old face and size, short black queue, faded indigo short jacket and the off-white apron. Keep the 2D anime character design while performing the directed actions.
<Subject 3> (appears in [Shot 1]): fully_preserved - Keep the L-shaped counter facing the street, the stove and copper pot behind it, the earthenware jars, the chalk slate on the pillar, the dark shopfront and black-tiled eaves in the same spatial arrangement. Lighting follows the time of day described in this shot.
detailed_description:
Chinese 2D hand-drawn anime film style: clean confident ink-like linework, soft cel shading with two-tone shadows, hand-painted watercolor backgrounds with visible paper grain, late-Qing Shaoxing water-town atmosphere. Restrained palette of ink black, faded indigo, grey-white plaster and weathered wood, with warm amber firelight and wine steam inside the counter against cool grey-blue daylight in the street. Expressive but grounded anime acting, consistent character model sheets, stable line weight. No readable text, no Chinese characters, no subtitles, no captions, no logos, no watermarks anywhere in the frame.
Early winter, overcast and cold: the street is washed in pale desaturated grey-blue, dry leaves and grit blow in a thin wind, only the stove fire inside keeps a small warm orange glow; breath is faintly visible.
[Shot 1] One continuous 12.2500-second shot with no internal cut. Medium shot at the threshold from the street side: the boy crouching beside Kong Yiji on the stone threshold.
The location is <Subject 3>; keep its counter, stove, jars and shopfront in their established places. Referenced characters on screen: <Subject 1>, <Subject 2>. Each referenced character appears exactly once, never duplicated; extras, if any, are simpler background figures who do not resemble the referenced characters.
The boy sets a bowl of warm wine on the threshold. Kong Yiji fishes four copper coins from his ragged pocket and places them in the boy's palm; his fingers are caked with mud. The boy stares at the muddy hand. Lips closed during the narration.
The unseen narrator, in the calm, low, slightly weary voice of a man in his forties, speaking Mandarin Chinese with measured, reflective pacing, remembering his childhood (S1), says in an off-screen voiceover, quietly, slowly, from second four to second eleven: <d>[Chinese] 见他满手是泥，原来他便用这手走来的。</d> All visible characters keep their lips closed during the narration.
Animation stays smooth and on-model: stable faces, consistent line weight, cloth folds and queues moving with the body, steam and firelight drifting gently. Keep anatomy correct, with no extra limbs or duplicated props. End on the muddy hand; next shot he leaves.
Keep a short breath of silence after the final spoken words; no unfinished syllable at the end. Mouths move only for the written Chinese lines; no other intelligible speech, no ad-libs. Absolutely no subtitles, captions, on-screen text, Chinese characters, signboards with writing, logos or watermarks.
overall_soundscape:
Coins in a palm, wind, a soft sip. Natural acoustic space of a small street-front tavern; no words beyond the written lines.
non_diegetic_music:
N/A
Shot 29
15 s · Reference to video · Pika H3 Flex ref2va (all-sparse), 768p
subject_definitions:
<Subject 1> is Kong Yiji in his last appearance, drawn in 2D anime style: the same tall gaunt scholar of about fifty with the same scruffy grey-white beard and grey queue, but his face now darkened, hollow and much thinner; he no longer wears the long gown but a ragged, patched, dirty dark-grey lined short jacket; his broken legs are folded uselessly beneath him and he sits on a flattened woven cattail-straw mat bag, which hangs from his shoulders by a straw rope; his hands are caked with mud, as shown in <Picture 1>. This image supplies identity and costume only; its plain portrait backdrop is not the story location.
<Subject 2> is the empty Xianheng tavern in the water town of Luzhen, drawn as a 2D anime watercolor background: a large L-shaped wooden counter opens directly onto the grey flagstone street; behind the counter sit a low clay stove with a copper pot of hot water for warming wine, rows of large brown earthenware wine jars with red cloth-and-clay seals, and small tin wine pots; a dark lacquered wooden shopfront with removable plank doors, black-tiled eaves above, whitewashed walls; through a doorway at the back is a dim inner room with square tables; a small dark chalk slate hangs on a pillar by the counter covered only with faint smudges; a raised wooden threshold below the counter's street side, as shown in <Picture 2>. This image supplies the set layout, materials and colors, not a locked camera angle; it is empty and contains no people.
summary:
[reference generation] A continuous 15.0833-second 2D anime shot from an adaptation of Lu Xun's story set in a late-Qing village tavern. Having drained the bowl, Kong Yiji presses both muddy hands on the flagstones and slowly drags himself away, sitting on the mat, the straw rope taut over his shoulders.
retention_analysis:
<Subject 1> (appears in [Shot 1]): fully_preserved - Keep his face identity, grey-white beard and queue, the darkened thin face, the ragged dark-grey lined short jacket, the straw mat he sits on with the straw rope over his shoulders, and the muddy hands; his legs stay folded beneath him and he never stands. Keep the 2D anime character design while performing the directed actions.
<Subject 2> (appears in [Shot 1]): fully_preserved - Keep the L-shaped counter facing the street, the stove and copper pot behind it, the earthenware jars, the chalk slate on the pillar, the dark shopfront and black-tiled eaves in the same spatial arrangement. Lighting follows the time of day described in this shot.
detailed_description:
Chinese 2D hand-drawn anime film style: clean confident ink-like linework, soft cel shading with two-tone shadows, hand-painted watercolor backgrounds with visible paper grain, late-Qing Shaoxing water-town atmosphere. Restrained palette of ink black, faded indigo, grey-white plaster and weathered wood, with warm amber firelight and wine steam inside the counter against cool grey-blue daylight in the street. Expressive but grounded anime acting, consistent character model sheets, stable line weight. No readable text, no Chinese characters, no subtitles, no captions, no logos, no watermarks anywhere in the frame.
Early winter, overcast and cold: the street is washed in pale desaturated grey-blue, dry leaves and grit blow in a thin wind, only the stove fire inside keeps a small warm orange glow; breath is faintly visible.
[Shot 1] One continuous 15.0833-second shot with no internal cut. Wide shot from inside the shopfront looking out to the cold grey street: Kong Yiji on his straw mat in the middle ground, moving away.
The location is <Subject 2>; keep its counter, stove, jars and shopfront in their established places. Referenced characters on screen: <Subject 1>. Each referenced character appears exactly once, never duplicated; extras, if any, are simpler background figures who do not resemble the referenced characters.
Having drained the bowl, Kong Yiji presses both muddy hands on the flagstones and slowly drags himself away, sitting on the mat, the straw rope taut over his shoulders. Off-screen laughter rises and fades. He grows smaller down the empty street as leaves blow past. No one speaks.
Animation stays smooth and on-model: stable faces, consistent line weight, cloth folds and queues moving with the body, steam and firelight drifting gently. Keep anatomy correct, with no extra limbs or duplicated props. End on the tiny figure far down the street; next shot years-end.
Keep a short breath of silence after the final spoken words; no unfinished syllable at the end. Mouths move only for the written Chinese lines; no other intelligible speech, no ad-libs. Absolutely no subtitles, captions, on-screen text, Chinese characters, signboards with writing, logos or watermarks.
overall_soundscape:
Hands scraping on stone, straw mat dragging, laughter fading, cold wind. Natural acoustic space of a small street-front tavern; no words beyond the written lines.
non_diegetic_music:
A sparse, soft solo guqin phrase, slow and fading, very quiet.
Shot 30
10 s · Reference to video · Pika H3 Flex ref2va (all-sparse), 768p
subject_definitions:
<Subject 1> is the tavern owner, a stout man in his fifties drawn in 2D anime style: round fleshy face with heavy jowls, small sharp narrow eyes, a thin drooping black moustache, a black round skullcap, a long black queue, a dark brown cotton long gown under a black sleeveless waistcoat, often holding an abacus, as shown in <Picture 1>. This image supplies identity and costume only; its plain portrait backdrop is not the story location.
<Subject 2> is the interior of the same tavern seen from behind the counter: the worn wooden countertop across the foreground, the clay stove with copper hot-water pot and tin wine pots, dark earthenware wine jars with red seals, the open shopfront and the grey flagstone street with whitewashed walls and black-tiled roofs beyond, as shown in <Picture 2>. This image supplies the set layout, materials and colors, not a locked camera angle; it is empty and contains no people.
summary:
[reference generation] A continuous 10.1250-second 2D anime shot from an adaptation of Lu Xun's story set in a late-Qing village tavern. The owner takes down the dark slate, squints at it (no readable writing) and grumbles, then hangs it back..
retention_analysis:
<Subject 1> (appears in [Shot 1]): fully_preserved - Keep his stout build, round fleshy face, small narrow eyes, thin moustache, black skullcap and the dark brown gown with black waistcoat. Keep the 2D anime character design while performing the directed actions.
<Subject 2> (appears in [Shot 1]): fully_preserved - Keep the L-shaped counter facing the street, the stove and copper pot behind it, the earthenware jars, the chalk slate on the pillar, the dark shopfront and black-tiled eaves in the same spatial arrangement. Lighting follows the time of day described in this shot.
detailed_description:
Chinese 2D hand-drawn anime film style: clean confident ink-like linework, soft cel shading with two-tone shadows, hand-painted watercolor backgrounds with visible paper grain, late-Qing Shaoxing water-town atmosphere. Restrained palette of ink black, faded indigo, grey-white plaster and weathered wood, with warm amber firelight and wine steam inside the counter against cool grey-blue daylight in the street. Expressive but grounded anime acting, consistent character model sheets, stable line weight. No readable text, no Chinese characters, no subtitles, no captions, no logos, no watermarks anywhere in the frame.
Early winter, overcast and cold: the street is washed in pale desaturated grey-blue, dry leaves and grit blow in a thin wind, only the stove fire inside keeps a small warm orange glow; breath is faintly visible.
[Shot 1] One continuous 10.1250-second shot with no internal cut. Medium shot of the owner behind the counter in the cold grey light, holding the chalk slate.
The location is <Subject 2>; keep its counter, stove, jars and shopfront in their established places. Referenced characters on screen: <Subject 1>. Each referenced character appears exactly once, never duplicated; extras, if any, are simpler background figures who do not resemble the referenced characters.
The owner takes down the dark slate, squints at it (no readable writing) and grumbles, then hangs it back.
<Subject 1>, in the thick, deep, gruff voice of a stout man in his fifties, speaking Mandarin Chinese bluntly with a mocking edge (S4), says grumbling, from second two to second six: <d>[Chinese] 孔乙己还欠十九个钱呢！</d> Only this character's lips move for this line; listeners stay silent.
Animation stays smooth and on-model: stable faces, consistent line weight, cloth folds and queues moving with the body, steam and firelight drifting gently. Keep anatomy correct, with no extra limbs or duplicated props. End on the slate hung back on the pillar; final shot the empty street.
Keep a short breath of silence after the final spoken words; no unfinished syllable at the end. Mouths move only for the written Chinese lines; no other intelligible speech, no ad-libs. Absolutely no subtitles, captions, on-screen text, Chinese characters, signboards with writing, logos or watermarks.
overall_soundscape:
Wind, a slate knocking on wood. Natural acoustic space of a small street-front tavern; no words beyond the written lines.
non_diegetic_music:
N/A
Shot 31
15 s · Reference to video · Pika H3 Flex ref2va (all-sparse), 768p
subject_definitions:
<Subject 1> is the young waiter, a thin twelve-year-old boy drawn in 2D anime style: round boyish face, large dark attentive eyes, shaved Qing forehead with a short black queue, a plain faded indigo cotton short jacket and trousers, a stained off-white cloth apron tied at the waist, black cloth shoes, as shown in <Picture 1>. This image supplies identity and costume only; its plain portrait backdrop is not the story location.
<Subject 2> is the interior of the same tavern seen from behind the counter: the worn wooden countertop across the foreground, the clay stove with copper hot-water pot and tin wine pots, dark earthenware wine jars with red seals, the open shopfront and the grey flagstone street with whitewashed walls and black-tiled roofs beyond, as shown in <Picture 2>. This image supplies the set layout, materials and colors, not a locked camera angle; it is empty and contains no people.
summary:
[reference generation] A continuous 15.0833-second 2D anime shot from an adaptation of Lu Xun's story set in a late-Qing village tavern. The boy looks out at the empty street where Kong Yiji used to stand.
retention_analysis:
<Subject 1> (appears in [Shot 1]): fully_preserved - Keep his twelve-year-old face and size, short black queue, faded indigo short jacket and the off-white apron. Keep the 2D anime character design while performing the directed actions.
<Subject 2> (appears in [Shot 1]): fully_preserved - Keep the L-shaped counter facing the street, the stove and copper pot behind it, the earthenware jars, the chalk slate on the pillar, the dark shopfront and black-tiled eaves in the same spatial arrangement. Lighting follows the time of day described in this shot.
detailed_description:
Chinese 2D hand-drawn anime film style: clean confident ink-like linework, soft cel shading with two-tone shadows, hand-painted watercolor backgrounds with visible paper grain, late-Qing Shaoxing water-town atmosphere. Restrained palette of ink black, faded indigo, grey-white plaster and weathered wood, with warm amber firelight and wine steam inside the counter against cool grey-blue daylight in the street. Expressive but grounded anime acting, consistent character model sheets, stable line weight. No readable text, no Chinese characters, no subtitles, no captions, no logos, no watermarks anywhere in the frame.
Early winter, overcast and cold: the street is washed in pale desaturated grey-blue, dry leaves and grit blow in a thin wind, only the stove fire inside keeps a small warm orange glow; breath is faintly visible.
[Shot 1] One continuous 15.0833-second shot with no internal cut. Wide shot from behind the counter: the boy small in the right foreground by the stove, the empty cold grey street beyond the empty counter. A very slow push toward the empty street.
The location is <Subject 2>; keep its counter, stove, jars and shopfront in their established places. Referenced characters on screen: <Subject 1>. Each referenced character appears exactly once, never duplicated; extras, if any, are simpler background figures who do not resemble the referenced characters.
The boy looks out at the empty street where Kong Yiji used to stand. Leaves drift past. He turns back to the stove. Lips closed. The shot lingers on the empty street at the end.
The unseen narrator, in the calm, low, slightly weary voice of a man in his forties, speaking Mandarin Chinese with measured, reflective pacing, remembering his childhood (S1), says in an off-screen voiceover, quiet and slow, with a long pause at the dash, from second three to second twelve: <d>[Chinese] 我到现在终于没有见——大约孔乙己的确死了。</d> All visible characters keep their lips closed during the narration.
Animation stays smooth and on-model: stable faces, consistent line weight, cloth folds and queues moving with the body, steam and firelight drifting gently. Keep anatomy correct, with no extra limbs or duplicated props. End of film on the empty street.
Keep a short breath of silence after the final spoken words; no unfinished syllable at the end. Mouths move only for the written Chinese lines; no other intelligible speech, no ad-libs. Absolutely no subtitles, captions, on-screen text, Chinese characters, signboards with writing, logos or watermarks.
overall_soundscape:
Cold wind, a single leaf scraping on stone, the stove fire. Natural acoustic space of a small street-front tavern; no words beyond the written lines.
non_diegetic_music:
A sparse, soft solo guqin closing phrase, very quiet, ending on a long fading note.
H3 Gallery
