A man nails an iron spike to his roof and warns the town. They ring church bells against lightning instead, in a tower full of gunpowder. A fable inspired by the 1769 Brescia disaster, when lightning struck the Church of San Nazaro and its gunpowder store destroyed about a sixth of the city. The inventor is fictional.
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 inventor
Character
Image prompt
Hand-drawn 2D cartoon character with a large peach-colored bean-shaped head, thick black outlines, tiny dot eyes and big expressive eyebrows, short rounded body and simple stick-like arms and legs, flat clean colors with soft cel shading. Full-body character reference of the inventor, a thin man with round wire spectacles, messy bright orange hair, a long bottle-green coat with brass buttons, brown boots, holding a small hammer in his right hand. Neutral confident expression, standing front three-quarter view, plain light grey background, no text.
The mayor
Character
Image prompt
Hand-drawn 2D cartoon character with a large peach-colored bean-shaped head, thick black outlines, tiny dot eyes and big expressive eyebrows, short rounded body and simple stick-like arms and legs, flat clean colors with soft cel shading. Full-body character reference of the mayor, a very round man in a crimson coat with gold trim, a black tricorn hat and a huge curled white mustache, smug half-closed eyes, hands behind his back. Standing front three-quarter view, plain light grey background, no text.
The bell ringer
Character
Image prompt
Hand-drawn 2D cartoon character with a large peach-colored bean-shaped head, thick black outlines, tiny dot eyes and big expressive eyebrows, short rounded body and simple stick-like arms and legs, flat clean colors with soft cel shading. Full-body character reference of a church bell ringer, a stocky man with a bald head and a ring of dark hair, a plain brown hooded robe tied with a rope belt, holding a thick bell rope. Neutral expression, standing front three-quarter view, plain light grey background, no text.
The inventor's cottage
Location
Image prompt
Richly painted storybook background art of an 18th-century Italian hill town with terracotta roofs and warm stone, detailed game-art illustration with volumetric light and soft depth, no people, no text. A small stone cottage with a terracotta roof on a grassy hill outside the town, a tall black iron spike standing on the roof ridge, a copper wire running from the spike down the side wall into the ground, a vegetable patch in front. Below in the distance, the walled town and its tall church tower. Bright afternoon light. Vertical composition.
Town square
Location
Image prompt
Richly painted storybook background art of an 18th-century Italian hill town with terracotta roofs and warm stone, detailed game-art illustration with volumetric light and soft depth, no people, no text. A sunny cobbled town square seen from street level, a stone fountain in the center, market stalls with red and white striped awnings on both sides, and a tall stone church tower with an open bell chamber holding bronze bells rising at the back. Warm ochre sunlight and clear blue sky. Vertical composition.
Bell chamber
Location
Image prompt
Richly painted storybook background art of an 18th-century Italian hill town with terracotta roofs and warm stone, detailed game-art illustration with volumetric light and soft depth, no people, no text. Interior of the bell chamber at the top of a stone church tower, open stone arches showing the sky and town roofs, three big glossy bronze bells hanging from heavy wooden beams, thick ropes hanging down, an open wooden hatch in the floor revealing stacked gunpowder barrels in the room below. Warm daylight through the arches. Vertical composition.
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.
He nailed an iron spike to his roof and told the town, "Put one on every tower, or lightning will burn this city down." The whole market burst out laughing.
Video prompt
subject_definitions:
<Subject 1> is the inventor, a thin 2D cartoon man with a large peach-colored head, round wire spectacles, messy bright orange hair, a long bottle-green coat with brass buttons and brown boots, shown in <Picture 1>. This image supplies his identity and costume only.
<Subject 2> is the inventor's small stone cottage on a grassy hill shown in <Picture 2>, with a terracotta roof, a tall black iron spike on the roof ridge, a copper wire running down the side wall into the ground, a vegetable patch in front and the walled town with its church tower in the distance below.
<Subject 3> is the town square shown in <Picture 3>: sunlit cobbles, a stone fountain in the center, market stalls with red and white striped awnings and a tall stone church tower with an open bell chamber rising at the back.
<Subject 4> is the mayor, a very round 2D cartoon man with a large peach-colored egg-shaped head, smug half-closed eyes, a huge curled white mustache, a black tricorn hat and a crimson coat with gold trim, shown in <Picture 4>. This image supplies his identity and costume only.
summary:
[reference generation] A 10.125-second vertical sequence of 3 shots with continuous off-screen narration by one narrator. The inventor, a single man working alone, fixes an iron spike to his roof, urgently warns the town below, and the whole market bursts out laughing. He never appears twice in the same frame.
retention_analysis:
<Subject 1> (appears in [Shot 1], [Shot 2]): fully_preserved - Keep his peach head, round spectacles, messy orange hair and long green coat; the plain reference background is not part of his role.
<Subject 2> (appears in [Shot 1]): fully_preserved - Keep the cottage, the spike on the ridge, the copper wire down the wall and the distant town and tower in the same layout.
<Subject 3> (appears in [Shot 3]): fully_preserved - Keep the fountain, striped stalls and church tower in the same layout and painted storybook look.
<Subject 4> (appears in [Shot 3]): fully_preserved - Keep his round body, tricorn hat, white curled mustache and crimson coat; the plain reference background is not part of his role.
detailed_description:
Hand-drawn 2D cartoon characters with large peach-colored heads, thick black outlines, tiny dot eyes and big expressive eyebrows, short rounded bodies and simple stick-like arms and legs, placed in richly painted storybook backgrounds of an 18th-century Italian hill town with terracotta roofs, warm stone and volumetric light. Key props such as the iron spike, the copper wire, the bells and the barrels are rendered as glossy 3D objects with metallic highlights and shallow depth of field. Comic emotion marks such as sweat drops, motion lines and laughing tears. Vertical 9:16 framing with the main action in the upper two thirds. Energetic, constantly moving camera. No readable text, letters, subtitles, captions, logos or watermarks anywhere. The only voice is one unseen off-screen narrator; on-screen characters never speak and never lip-sync, their mouths stay closed except for silent open-mouthed laughter or gasps. Warm sunny afternoon: ochre, terracotta and clear sky blue.
[Shot 1] A close low-angle view of the roof ridge of <Subject 2>. Exactly one person is on the roof: <Subject 1>, alone, kneels on the terracotta tiles and strikes the base of the tall black iron spike twice with his hammer; small sparks of light jump at each blow and the glossy spike glints. The camera tilts up along the spike toward the bright sky. The unseen narrator, in the warm, slightly raspy voice of a middle-aged man, speaking English with a neutral American accent, a brisk, fast storytelling pace and playful, dramatic energy (S1), says in an off-screen voiceover: <d>[English] He nailed an iron spike to his roof and told the town,</d> All on-screen lips stay closed during the narration.
[Shot 2] A medium view of <Subject 1> standing on the roof ridge beside the spike, with the distant town and church tower below. He thrusts his whole arm toward the town, eyebrows raised high in alarm, a sweat drop beside his head, mouth closed. The camera pushes in toward him. The unseen narrator, in the warm, slightly raspy voice of a middle-aged man, speaking English with a neutral American accent, a brisk, fast storytelling pace and playful, dramatic energy (S1), says in an off-screen voiceover: <d>[English] "Put one on every tower, or lightning will burn this city down."</d> All on-screen lips stay closed during the narration.
[Shot 3] A wide view of the market in <Subject 3>. A crowd of townspeople and <Subject 4> in front of the fountain burst into open-mouthed silent laughter, holding their bellies; the mayor's mustache bounces and a vendor drops a basket of apples that roll across the cobbles. The camera slowly pulls back. The unseen narrator, in the warm, slightly raspy voice of a middle-aged man, speaking English with a neutral American accent, a brisk, fast storytelling pace and playful, dramatic energy (S1), says in an off-screen voiceover: <d>[English] The whole market burst out laughing.</d> All on-screen lips stay closed during the narration.
The narration flows continuously from shot to shot without long pauses and finishes before the final half second.
overall_soundscape:
Two sharp metallic hammer clangs, wind on the hill, then a whole crowd's roaring laughter, apples bouncing on cobblestones, market bustle. Effects stay below the narration; no on-screen character speaks.
non_diegetic_music:
N/A
You see, everyone knew how to stop lightning. When a storm came, the bell ringers climbed the church tower and rang the bells as loud as they could.
Video prompt
subject_definitions:
<Subject 1> is the head bell ringer, a stocky 2D cartoon man with a large peach-colored bald head with a ring of dark hair and a plain brown hooded robe tied with a rope belt, shown in <Picture 1>. This image supplies his identity and costume only.
<Subject 2> is the bell chamber at the top of the church tower shown in <Picture 2>: open stone arches over the town roofs, three big bronze bells on heavy wooden beams with thick ropes, and an open wooden floor hatch showing gunpowder barrels stacked in the room below.
<Subject 3> is the town square shown in <Picture 3>: sunlit cobbles, a stone fountain in the center, market stalls with red and white striped awnings and a tall stone church tower with an open bell chamber rising at the back.
summary:
[reference generation] A 10.125-second vertical sequence of 3 shots with continuous off-screen narration by one narrator. A storm rolls over the town and the bell ringers ring the church bells as hard as they can to scare the lightning away.
retention_analysis:
<Subject 1> (appears in [Shot 2], [Shot 3]): fully_preserved - Keep his bald head, ring of dark hair and brown robe with rope belt; the plain reference background is not part of his role.
<Subject 2> (appears in [Shot 2]): fully_preserved - Keep the arches, the three bells on the beam, the ropes and the floor hatch with barrels below.
<Subject 3> (appears in [Shot 1]): fully_preserved - Keep the fountain, striped stalls and church tower in the same layout and painted storybook look.
detailed_description:
Hand-drawn 2D cartoon characters with large peach-colored heads, thick black outlines, tiny dot eyes and big expressive eyebrows, short rounded bodies and simple stick-like arms and legs, placed in richly painted storybook backgrounds of an 18th-century Italian hill town with terracotta roofs, warm stone and volumetric light. Key props such as the iron spike, the copper wire, the bells and the barrels are rendered as glossy 3D objects with metallic highlights and shallow depth of field. Comic emotion marks such as sweat drops, motion lines and laughing tears. Vertical 9:16 framing with the main action in the upper two thirds. Energetic, constantly moving camera. No readable text, letters, subtitles, captions, logos or watermarks anywhere. The only voice is one unseen off-screen narrator; on-screen characters never speak and never lip-sync, their mouths stay closed except for silent open-mouthed laughter or gasps. Darkening storm afternoon: slate grey and green with warm stone inside the tower.
[Shot 1] A wide view of <Subject 3> as heavy dark storm clouds roll over the church tower; rain begins and townspeople hurry across the square holding baskets over their heads. The camera tilts up toward the bell chamber at the top of the tower. The unseen narrator, in the warm, slightly raspy voice of a middle-aged man, speaking English with a neutral American accent, a brisk, fast storytelling pace and playful, dramatic energy (S1), says in an off-screen voiceover: <d>[English] You see, everyone knew how to stop lightning.</d> All on-screen lips stay closed during the narration.
[Shot 2] Inside <Subject 2>, <Subject 1> and two other ringers in the same brown robes haul on the thick ropes with all their weight; the three big bronze bells swing hard on the beam while rain slants past the arches. The camera circles slowly around them. The unseen narrator, in the warm, slightly raspy voice of a middle-aged man, speaking English with a neutral American accent, a brisk, fast storytelling pace and playful, dramatic energy (S1), says in an off-screen voiceover: <d>[English] When a storm came, the bell ringers climbed the church tower</d> All on-screen lips stay closed during the narration.
[Shot 3] A close view of the biggest glossy bronze bell swinging toward the camera with motion lines, then <Subject 1>'s face with puffed cheeks and squeezed eyes as he pulls, rain beyond. The unseen narrator, in the warm, slightly raspy voice of a middle-aged man, speaking English with a neutral American accent, a brisk, fast storytelling pace and playful, dramatic energy (S1), says in an off-screen voiceover: <d>[English] and rang the bells as loud as they could.</d> All on-screen lips stay closed during the narration.
The narration flows continuously from shot to shot without long pauses and finishes before the final half second.
overall_soundscape:
Rolling thunder, rain on stone, ropes creaking, and loud deep bronze bell clangs ringing over and over, kept below the narration. Effects stay below the narration; no on-screen character speaks.
non_diegetic_music:
N/A
And that same tower was where the city kept its gunpowder. What could be safer than a church?
Video prompt
subject_definitions:
<Subject 1> is the head bell ringer, a stocky 2D cartoon man with a large peach-colored bald head with a ring of dark hair and a plain brown hooded robe tied with a rope belt, shown in <Picture 1>. This image supplies his identity and costume only.
<Subject 2> is the bell chamber at the top of the church tower shown in <Picture 2>: open stone arches over the town roofs, three big bronze bells on heavy wooden beams with thick ropes, and an open wooden floor hatch showing gunpowder barrels stacked in the room below.
summary:
[reference generation] A 7.292-second vertical sequence of 2 shots with continuous off-screen narration by one narrator. Beneath the swinging bells, the tower stores the city's gunpowder, and the ringer proudly pats a barrel.
retention_analysis:
<Subject 1> (appears in [Shot 2]): fully_preserved - Keep his bald head, ring of dark hair and brown robe with rope belt; the plain reference background is not part of his role.
<Subject 2> (appears in [Shot 1]): fully_preserved - Keep the arches, the three bells on the beam, the ropes and the floor hatch with barrels below.
detailed_description:
Hand-drawn 2D cartoon characters with large peach-colored heads, thick black outlines, tiny dot eyes and big expressive eyebrows, short rounded bodies and simple stick-like arms and legs, placed in richly painted storybook backgrounds of an 18th-century Italian hill town with terracotta roofs, warm stone and volumetric light. Key props such as the iron spike, the copper wire, the bells and the barrels are rendered as glossy 3D objects with metallic highlights and shallow depth of field. Comic emotion marks such as sweat drops, motion lines and laughing tears. Vertical 9:16 framing with the main action in the upper two thirds. Energetic, constantly moving camera. No readable text, letters, subtitles, captions, logos or watermarks anywhere. The only voice is one unseen off-screen narrator; on-screen characters never speak and never lip-sync, their mouths stay closed except for silent open-mouthed laughter or gasps. Warm tower interior, golden stone and dark wooden barrels.
[Shot 1] In <Subject 2>, the camera descends from the swinging bronze bells straight down through the open floor hatch into the room below, revealing dozens of stacked wooden gunpowder barrels. The unseen narrator, in the warm, slightly raspy voice of a middle-aged man, speaking English with a neutral American accent, a brisk, fast storytelling pace and playful, dramatic energy (S1), says in an off-screen voiceover: <d>[English] And that same tower was where the city kept its gunpowder.</d> All on-screen lips stay closed during the narration.
[Shot 2] A medium view in the barrel room: <Subject 1> pats the top of a barrel proudly, gives a confident thumbs up and smiles with closed lips, while the bells keep swinging in the hatch above him. The unseen narrator, in the warm, slightly raspy voice of a middle-aged man, speaking English with a neutral American accent, a brisk, fast storytelling pace and playful, dramatic energy (S1), says in an off-screen voiceover: <d>[English] What could be safer than a church?</d> All on-screen lips stay closed during the narration.
The narration flows continuously from shot to shot without long pauses and finishes before the final half second.
overall_soundscape:
Bells ringing above, muffled by stone, wooden barrel thumps. Effects stay below the narration; no on-screen character speaks.
non_diegetic_music:
N/A
The first storm, lightning hit his spike. Sparks ran down a wire into the mud, and his house didn't even get warm.
Video prompt
subject_definitions:
<Subject 1> is the inventor, a thin 2D cartoon man with a large peach-colored head, round wire spectacles, messy bright orange hair, a long bottle-green coat with brass buttons and brown boots, shown in <Picture 1>. This image supplies his identity and costume only.
<Subject 2> is the inventor's small stone cottage on a grassy hill shown in <Picture 2>, with a terracotta roof, a tall black iron spike on the roof ridge, a copper wire running down the side wall into the ground, a vegetable patch in front and the walled town with its church tower in the distance below.
summary:
[reference generation] A 8.0-second vertical sequence of 3 shots with continuous off-screen narration by one narrator. In the first storm, lightning strikes the inventor's spike, sparks run down the wire into the mud, and his house is unharmed.
retention_analysis:
<Subject 1> (appears in [Shot 3]): fully_preserved - Keep his peach head, round spectacles, messy orange hair and long green coat; the plain reference background is not part of his role.
<Subject 2> (appears in [Shot 1]): fully_preserved - Keep the cottage, the spike on the ridge, the copper wire down the wall and the distant town and tower in the same layout.
detailed_description:
Hand-drawn 2D cartoon characters with large peach-colored heads, thick black outlines, tiny dot eyes and big expressive eyebrows, short rounded bodies and simple stick-like arms and legs, placed in richly painted storybook backgrounds of an 18th-century Italian hill town with terracotta roofs, warm stone and volumetric light. Key props such as the iron spike, the copper wire, the bells and the barrels are rendered as glossy 3D objects with metallic highlights and shallow depth of field. Comic emotion marks such as sweat drops, motion lines and laughing tears. Vertical 9:16 framing with the main action in the upper two thirds. Energetic, constantly moving camera. No readable text, letters, subtitles, captions, logos or watermarks anywhere. The only voice is one unseen off-screen narrator; on-screen characters never speak and never lip-sync, their mouths stay closed except for silent open-mouthed laughter or gasps. Grey-green storm light, white-blue electric sparks.
[Shot 1] A wide view of <Subject 2> under a dark grey-green storm sky with rain. A brilliant white lightning bolt strikes the tip of the iron spike on the roof ridge and the whole frame flashes. The unseen narrator, in the warm, slightly raspy voice of a middle-aged man, speaking English with a neutral American accent, a brisk, fast storytelling pace and playful, dramatic energy (S1), says in an off-screen voiceover: <d>[English] The first storm, lightning hit his spike.</d> All on-screen lips stay closed during the narration.
[Shot 2] A close view along the side wall: bright white-blue sparks race down the glossy copper wire like a fuse and disappear into the muddy ground, which hisses and steams briefly. The camera follows the sparks downward. The unseen narrator, in the warm, slightly raspy voice of a middle-aged man, speaking English with a neutral American accent, a brisk, fast storytelling pace and playful, dramatic energy (S1), says in an off-screen voiceover: <d>[English] Sparks ran down a wire into the mud,</d> All on-screen lips stay closed during the narration.
[Shot 3] A medium view of the cottage door: <Subject 1> opens it and peeks out in the rain, touches the wall with one hand, finds it perfectly fine, and gives a proud closed-mouth smile with raised eyebrows. The camera holds still. The unseen narrator, in the warm, slightly raspy voice of a middle-aged man, speaking English with a neutral American accent, a brisk, fast storytelling pace and playful, dramatic energy (S1), says in an off-screen voiceover: <d>[English] and his house didn't even get warm.</d> All on-screen lips stay closed during the narration.
The narration flows continuously from shot to shot without long pauses and finishes before the final half second.
overall_soundscape:
A huge thunder crack, electric crackling sparks running down metal, a hiss of steam in mud, steady rain, a creaking door. Effects stay below the narration; no on-screen character speaks.
non_diegetic_music:
N/A
"See? It pulls lightning in!" Now they laughed even harder. They called him the lightning fisherman and banned his spike.
Video prompt
subject_definitions:
<Subject 1> is the mayor, a very round 2D cartoon man with a large peach-colored egg-shaped head, smug half-closed eyes, a huge curled white mustache, a black tricorn hat and a crimson coat with gold trim, shown in <Picture 1>. This image supplies his identity and costume only.
<Subject 2> is the town square shown in <Picture 2>: sunlit cobbles, a stone fountain in the center, market stalls with red and white striped awnings and a tall stone church tower with an open bell chamber rising at the back.
summary:
[reference generation] A 8.0-second vertical sequence of 3 shots with continuous off-screen narration by one narrator. In the market the mayor points at the spike on the hill, the town laughs even harder, mocks the inventor as a lightning fisherman and the mayor bans the spike.
retention_analysis:
<Subject 1> (appears in [Shot 1], [Shot 3]): fully_preserved - Keep his round body, tricorn hat, white curled mustache and crimson coat; the plain reference background is not part of his role.
<Subject 2> (appears in [Shot 1]): fully_preserved - Keep the fountain, striped stalls and church tower in the same layout and painted storybook look.
detailed_description:
Hand-drawn 2D cartoon characters with large peach-colored heads, thick black outlines, tiny dot eyes and big expressive eyebrows, short rounded bodies and simple stick-like arms and legs, placed in richly painted storybook backgrounds of an 18th-century Italian hill town with terracotta roofs, warm stone and volumetric light. Key props such as the iron spike, the copper wire, the bells and the barrels are rendered as glossy 3D objects with metallic highlights and shallow depth of field. Comic emotion marks such as sweat drops, motion lines and laughing tears. Vertical 9:16 framing with the main action in the upper two thirds. Energetic, constantly moving camera. No readable text, letters, subtitles, captions, logos or watermarks anywhere. The only voice is one unseen off-screen narrator; on-screen characters never speak and never lip-sync, their mouths stay closed except for silent open-mouthed laughter or gasps. Wet cobbles after rain, pale sun breaking through clouds.
[Shot 1] In <Subject 2> after the rain, <Subject 1> points up at the distant hill where the little spike stands against the clearing sky, eyes wide, mustache bristling, mouth closed. The unseen narrator, in the warm, slightly raspy voice of a middle-aged man, speaking English with a neutral American accent, a brisk, fast storytelling pace and playful, dramatic energy (S1), says in an off-screen voiceover: <d>[English] "See? It pulls lightning in!"</d> All on-screen lips stay closed during the narration.
[Shot 2] The crowd behind him laughs even harder with laughing tears; several townspeople mime fishing, casting imaginary lines with sticks up toward the sky. The camera pans across them. The unseen narrator, in the warm, slightly raspy voice of a middle-aged man, speaking English with a neutral American accent, a brisk, fast storytelling pace and playful, dramatic energy (S1), says in an off-screen voiceover: <d>[English] Now they laughed even harder. They called him the lightning fisherman</d> All on-screen lips stay closed during the narration.
[Shot 3] A close view of <Subject 1> slamming a big red wax seal onto a blank scroll with a thump, then pointing straight at the camera with a smug half-closed look. The unseen narrator, in the warm, slightly raspy voice of a middle-aged man, speaking English with a neutral American accent, a brisk, fast storytelling pace and playful, dramatic energy (S1), says in an off-screen voiceover: <d>[English] and banned his spike.</d> All on-screen lips stay closed during the narration.
The narration flows continuously from shot to shot without long pauses and finishes before the final half second.
overall_soundscape:
Dripping water, crowd laughter rising to a roar, whooshing sticks, a heavy wooden thump of the seal. Effects stay below the narration; no on-screen character speaks.
non_diegetic_music:
N/A
Then one night, the storm of the century rolled in. He ran to the church, begging the ringers to come down. "Go home, fisherman. The bells have got this."
Video prompt
subject_definitions:
<Subject 1> is the inventor, a thin 2D cartoon man with a large peach-colored head, round wire spectacles, messy bright orange hair, a long bottle-green coat with brass buttons and brown boots, shown in <Picture 1>. This image supplies his identity and costume only.
<Subject 2> is the head bell ringer, a stocky 2D cartoon man with a large peach-colored bald head with a ring of dark hair and a plain brown hooded robe tied with a rope belt, shown in <Picture 2>. This image supplies his identity and costume only.
<Subject 3> is the bell chamber at the top of the church tower shown in <Picture 3>: open stone arches over the town roofs, three big bronze bells on heavy wooden beams with thick ropes, and an open wooden floor hatch showing gunpowder barrels stacked in the room below.
<Subject 4> is the town square shown in <Picture 4>: sunlit cobbles, a stone fountain in the center, market stalls with red and white striped awnings and a tall stone church tower with an open bell chamber rising at the back.
summary:
[reference generation] A 11.542-second vertical sequence of 3 shots with continuous off-screen narration by one narrator. On the night of the storm of the century the inventor runs to the church tower and begs the ringers to come down, but the head ringer waves him away.
retention_analysis:
<Subject 1> (appears in [Shot 2], [Shot 3]): fully_preserved - Keep his peach head, round spectacles, messy orange hair and long green coat; the plain reference background is not part of his role.
<Subject 2> (appears in [Shot 3]): fully_preserved - Keep his bald head, ring of dark hair and brown robe with rope belt; the plain reference background is not part of his role.
<Subject 3> (appears in [Shot 3]): fully_preserved - Keep the arches, the three bells on the beam, the ropes and the floor hatch with barrels below.
<Subject 4> (appears in [Shot 1]): fully_preserved - Keep the fountain, striped stalls and church tower in the same layout and painted storybook look.
detailed_description:
Hand-drawn 2D cartoon characters with large peach-colored heads, thick black outlines, tiny dot eyes and big expressive eyebrows, short rounded bodies and simple stick-like arms and legs, placed in richly painted storybook backgrounds of an 18th-century Italian hill town with terracotta roofs, warm stone and volumetric light. Key props such as the iron spike, the copper wire, the bells and the barrels are rendered as glossy 3D objects with metallic highlights and shallow depth of field. Comic emotion marks such as sweat drops, motion lines and laughing tears. Vertical 9:16 framing with the main action in the upper two thirds. Energetic, constantly moving camera. No readable text, letters, subtitles, captions, logos or watermarks anywhere. The only voice is one unseen off-screen narrator; on-screen characters never speak and never lip-sync, their mouths stay closed except for silent open-mouthed laughter or gasps. Deep indigo and violet night, lightning flashes, warm lantern points.
[Shot 1] A huge wide view over the town at night: enormous violet storm clouds tower over the church tower in <Subject 4>, branching lightning flickers across the sky and trees bend in the wind. The camera slowly rises. The unseen narrator, in the warm, slightly raspy voice of a middle-aged man, speaking English with a neutral American accent, a brisk, fast storytelling pace and playful, dramatic energy (S1), says in an off-screen voiceover: <d>[English] Then one night, the storm of the century rolled in.</d> All on-screen lips stay closed during the narration.
[Shot 2] <Subject 1> runs across the rain-soaked square toward the tower holding a swinging lantern, coat flapping, a sweat drop flying off. Fast tracking camera. The unseen narrator, in the warm, slightly raspy voice of a middle-aged man, speaking English with a neutral American accent, a brisk, fast storytelling pace and playful, dramatic energy (S1), says in an off-screen voiceover: <d>[English] He ran to the church, begging the ringers to come down.</d> All on-screen lips stay closed during the narration.
[Shot 3] Inside <Subject 3> at night: <Subject 1> reaches the top of the stairs, panting, and waves both arms, pleading, pointing at the sky. <Subject 2>, holding his rope, gives a lazy dismissive wave of his hand with a smug closed-mouth smile, turns away and grips the rope to ring. Lightning flashes through the arches. The unseen narrator, in the warm, slightly raspy voice of a middle-aged man, speaking English with a neutral American accent, a brisk, fast storytelling pace and playful, dramatic energy (S1), says in an off-screen voiceover: <d>[English] "Go home, fisherman. The bells have got this."</d> All on-screen lips stay closed during the narration.
The narration flows continuously from shot to shot without long pauses and finishes before the final half second.
overall_soundscape:
Howling wind, heavy rain, distant thunder, running footsteps splashing, a lantern creaking, panting breath. Effects stay below the narration; no on-screen character speaks.
non_diegetic_music:
N/A
The bells rang. Lightning hit the tower. By morning, a sixth of the city was gone. His little house still stood, spike and all.
Video prompt
subject_definitions:
<Subject 1> is the inventor, a thin 2D cartoon man with a large peach-colored head, round wire spectacles, messy bright orange hair, a long bottle-green coat with brass buttons and brown boots, shown in <Picture 1>. This image supplies his identity and costume only.
<Subject 2> is the inventor's small stone cottage on a grassy hill shown in <Picture 2>, with a terracotta roof, a tall black iron spike on the roof ridge, a copper wire running down the side wall into the ground, a vegetable patch in front and the walled town with its church tower in the distance below.
<Subject 3> is the bell chamber at the top of the church tower shown in <Picture 3>: open stone arches over the town roofs, three big bronze bells on heavy wooden beams with thick ropes, and an open wooden floor hatch showing gunpowder barrels stacked in the room below.
summary:
[reference generation] A 9.417-second vertical sequence of 3 shots with continuous off-screen narration by one narrator. The bells ring, lightning strikes the tower and the gunpowder explodes; by morning part of the city lies in ruins, but the little house on the hill still stands.
retention_analysis:
<Subject 1> (appears in [Shot 2], [Shot 3]): fully_preserved - Keep his peach head, round spectacles, messy orange hair and long green coat; the plain reference background is not part of his role.
<Subject 2> (appears in [Shot 2], [Shot 3]): fully_preserved - Keep the cottage, the spike on the ridge, the copper wire down the wall and the distant town and tower in the same layout.
<Subject 3> (appears in [Shot 1]): fully_preserved - Keep the arches, the three bells on the beam, the ropes and the floor hatch with barrels below.
detailed_description:
Hand-drawn 2D cartoon characters with large peach-colored heads, thick black outlines, tiny dot eyes and big expressive eyebrows, short rounded bodies and simple stick-like arms and legs, placed in richly painted storybook backgrounds of an 18th-century Italian hill town with terracotta roofs, warm stone and volumetric light. Key props such as the iron spike, the copper wire, the bells and the barrels are rendered as glossy 3D objects with metallic highlights and shallow depth of field. Comic emotion marks such as sweat drops, motion lines and laughing tears. Vertical 9:16 framing with the main action in the upper two thirds. Energetic, constantly moving camera. No readable text, letters, subtitles, captions, logos or watermarks anywhere. The only voice is one unseen off-screen narrator; on-screen characters never speak and never lip-sync, their mouths stay closed except for silent open-mouthed laughter or gasps. Night indigo with an orange fireball, then hazy pale gold dawn.
[Shot 1] Inside <Subject 3> at night, the three bells swing and clang; a blinding lightning bolt strikes the top of the tower and the frame flashes white. The unseen narrator, in the warm, slightly raspy voice of a middle-aged man, speaking English with a neutral American accent, a brisk, fast storytelling pace and playful, dramatic energy (S1), says in an off-screen voiceover: <d>[English] The bells rang. Lightning hit the tower.</d> All on-screen lips stay closed during the narration.
[Shot 2] From the hill beside <Subject 2>, far away, the church tower in the town below erupts in a huge orange fireball and a rising column of smoke; <Subject 1> in the foreground flinches with his lantern. The unseen narrator, in the warm, slightly raspy voice of a middle-aged man, speaking English with a neutral American accent, a brisk, fast storytelling pace and playful, dramatic energy (S1), says in an off-screen voiceover: <d>[English] By morning, a sixth of the city was gone.</d> All on-screen lips stay closed during the narration.
[Shot 3] Hazy pale gold dawn: a thick smoke plume rises over a flattened, ruined section of the distant town where the tower stood. In the foreground the little cottage of <Subject 2> stands untouched, the iron spike gleaming in the sunrise, and <Subject 1> stands beside it, quiet and solemn. The camera slowly pushes in. The unseen narrator, in the warm, slightly raspy voice of a middle-aged man, speaking English with a neutral American accent, a brisk, fast storytelling pace and playful, dramatic energy (S1), says in an off-screen voiceover: <d>[English] His little house still stood, spike and all.</d> All on-screen lips stay closed during the narration.
The narration flows continuously from shot to shot without long pauses and finishes before the final half second.
overall_soundscape:
Bell clangs, an enormous thunder crack, a deep rumbling distant explosion, then quiet morning birds and a low wind. Effects stay below the narration; no on-screen character speaks.
non_diegetic_music:
N/A
Lightning wants the easiest road into the ground. An iron spike with a wire into wet earth is a highway.
Video prompt
subject_definitions:
<Subject 1> is the inventor's small stone cottage on a grassy hill shown in <Picture 1>, with a terracotta roof, a tall black iron spike on the roof ridge, a copper wire running down the side wall into the ground, a vegetable patch in front and the walled town with its church tower in the distance below.
summary:
[reference generation] A 8.0-second vertical sequence of 2 shots with continuous off-screen narration by one narrator. A cross-section diagram shows lightning running down the spike and wire safely into wet earth like a highway. The diagram contains no people, no faces, no characters and no barrels.
retention_analysis:
<Subject 1> (appears in [Shot 1]): fully_preserved - Keep the cottage, the spike on the ridge, the copper wire down the wall and the distant town and tower in the same layout.
detailed_description:
A clean educational cutaway diagram in painted storybook game-art style: the cottage, its glossy 3D iron spike and copper wire, smooth layered earth and a deep blue storm sky. This is an empty technical illustration with no living beings at all: no people, no characters, no heads, no faces, no animals, no barrels. Vertical 9:16 framing, slow smooth camera. No readable text, letters, labels, arrows with words, subtitles, logos or watermarks. Clean cutaway diagram: dark brown earth layers, deep blue sky, bright cyan electricity.
[Shot 1] A stylized side cross-section of the cottage in <Subject 1> and the ground beneath it, with layered brown earth and dark wet soil. A lightning bolt hits the spike and a bright cyan glowing path runs down the spike, along the copper wire down the wall and into the wet earth, where it spreads out harmlessly like glowing roots. Only the cottage, the spike, the wire, the earth layers and the sky are visible: no people, no faces, no characters, no barrels. The unseen narrator, in the warm, slightly raspy voice of a middle-aged man, speaking English with a neutral American accent, a brisk, fast storytelling pace and playful, dramatic energy (S1), says in an off-screen voiceover: <d>[English] Lightning wants the easiest road into the ground.</d> All on-screen lips stay closed during the narration.
[Shot 2] The camera follows the path underground as the cyan line widens into a smooth glowing road of light, with bright pulses streaming along it into the deep earth like traffic on a highway. The unseen narrator, in the warm, slightly raspy voice of a middle-aged man, speaking English with a neutral American accent, a brisk, fast storytelling pace and playful, dramatic energy (S1), says in an off-screen voiceover: <d>[English] An iron spike with a wire into wet earth is a highway.</d> All on-screen lips stay closed during the narration.
The narration flows continuously from shot to shot without long pauses and finishes before the final half second.
overall_soundscape:
A thunder crack, a smooth rising electric hum flowing downward, a soft deep rumble fading into the earth. Effects stay below the narration; no on-screen character speaks.
non_diegetic_music:
N/A
That first strike wasn't a failure. It was the spike working. The bells never scared lightning away. They just rang it toward a tower full of gunpowder.
Video prompt
subject_definitions:
<Subject 1> is the head bell ringer, a stocky 2D cartoon man with a large peach-colored bald head with a ring of dark hair and a plain brown hooded robe tied with a rope belt, shown in <Picture 1>. This image supplies his identity and costume only.
<Subject 2> is the bell chamber at the top of the church tower shown in <Picture 2>: open stone arches over the town roofs, three big bronze bells on heavy wooden beams with thick ropes, and an open wooden floor hatch showing gunpowder barrels stacked in the room below.
<Subject 3> is the inventor's small stone cottage on a grassy hill shown in <Picture 3>, with a terracotta roof, a tall black iron spike on the roof ridge, a copper wire running down the side wall into the ground, a vegetable patch in front and the walled town with its church tower in the distance below.
summary:
[reference generation] A 11.542-second vertical sequence of 3 shots with continuous off-screen narration by one narrator. The first strike on the spike replays as success, then a cross-section of the church tower shows the ringing bells drawing lightning down into the gunpowder.
retention_analysis:
<Subject 1> (appears in [Shot 2]): fully_preserved - Keep his bald head, ring of dark hair and brown robe with rope belt; the plain reference background is not part of his role.
<Subject 2> (appears in [Shot 2]): fully_preserved - Keep the arches, the three bells on the beam, the ropes and the floor hatch with barrels below.
<Subject 3> (appears in [Shot 1]): fully_preserved - Keep the cottage, the spike on the ridge, the copper wire down the wall and the distant town and tower in the same layout.
detailed_description:
Hand-drawn 2D cartoon characters with large peach-colored heads, thick black outlines, tiny dot eyes and big expressive eyebrows, short rounded bodies and simple stick-like arms and legs, placed in richly painted storybook backgrounds of an 18th-century Italian hill town with terracotta roofs, warm stone and volumetric light. Key props such as the iron spike, the copper wire, the bells and the barrels are rendered as glossy 3D objects with metallic highlights and shallow depth of field. Comic emotion marks such as sweat drops, motion lines and laughing tears. Vertical 9:16 framing with the main action in the upper two thirds. Energetic, constantly moving camera. No readable text, letters, subtitles, captions, logos or watermarks anywhere. The only voice is one unseen off-screen narrator; on-screen characters never speak and never lip-sync, their mouths stay closed except for silent open-mouthed laughter or gasps. Warm sepia flashback, then indigo night cutaway with cyan electricity and an orange glow.
[Shot 1] A warm sepia-tinted replay: lightning strikes the spike of <Subject 3>, sparks run down the copper wire into the ground, and the cottage stays untouched. The unseen narrator, in the warm, slightly raspy voice of a middle-aged man, speaking English with a neutral American accent, a brisk, fast storytelling pace and playful, dramatic energy (S1), says in an off-screen voiceover: <d>[English] That first strike wasn't a failure. It was the spike working.</d> All on-screen lips stay closed during the narration.
[Shot 2] A side cross-section of the tall wet stone church tower at night: in <Subject 2> at the top, <Subject 1> and the ringers pull the ropes and the bells swing. A jagged lightning bolt from the clouds bends straight toward the top of the tower. The unseen narrator, in the warm, slightly raspy voice of a middle-aged man, speaking English with a neutral American accent, a brisk, fast storytelling pace and playful, dramatic energy (S1), says in an off-screen voiceover: <d>[English] The bells never scared lightning away.</d> All on-screen lips stay closed during the narration.
[Shot 3] The cyan lightning path runs down the wet ropes and stone walls of the cutaway tower into the room of stacked gunpowder barrels at its base; the barrels glow brighter and brighter orange until the frame flashes to white. The unseen narrator, in the warm, slightly raspy voice of a middle-aged man, speaking English with a neutral American accent, a brisk, fast storytelling pace and playful, dramatic energy (S1), says in an off-screen voiceover: <d>[English] They just rang it toward a tower full of gunpowder.</d> All on-screen lips stay closed during the narration.
The narration flows continuously from shot to shot without long pauses and finishes before the final half second.
overall_soundscape:
A thunder crack, crackling sparks, bells ringing, a rising electric whine, then a sharp bright boom cut short. Effects stay below the narration; no on-screen character speaks.
non_diegetic_music:
N/A
H3 Gallery
