Studio
Every clip and still here is a real recorded result from the course's own production rig, not a live model call. On a wide screen the canvas builds the rig up as boards — one per week, plus the Cut and the reference episode — with an edge from each input to the take it produced, and a desk that replays what the rig recorded for whatever you ask it for. On a phone the desk is a form and the results are the list below. Recorded on the convener's rig with MiniMax H3 (local weights), keyframes with NoobAI-XL, reference stills for The Ninth Head with ChatGPT's image generator. Nothing is generated in the browser.
Desk
Pick a recorded run and the rig replays it. The prompt below is the one the rig was actually given; change it and the rig will say so rather than make something up.
Every recorded result, in order
Week 2 — The Loop
t1 — seed 66740
SLOP8760/W02/t1/66740
Seed
Seed: 66740
For the target video, at 0.00 seconds into the target video, <Picture 1> (from [Shot 1]) is fully referenced. integrated_multimodal_description: [Shot 1] 2D-animated, cinematic, the young woman shown in <Picture 1> stays where she stands on the wet night street, keeping her face, her long dark hair, her dark buttoned coat with the raised collar, and the warm amber lights behind her. Heavy rain keeps falling in diagonal streaks while loose strands of her hair lift and settle and her shoulders rise and fall with slow breathing. The camera holds a static shot for the whole duration. No other person enters the frame and the shot never cuts. overall_soundscape: Heavy rain falls on wet asphalt with a low distant traffic hum underneath; small drops tick against her coat. non_diegetic_music: N/A
One prompt, five seeds: watch how much of the shot the seed alone decides.
t2 — seed 66741
SLOP8760/W02/t2/66741
Seed
Seed: 66741
For the target video, at 0.00 seconds into the target video, <Picture 1> (from [Shot 1]) is fully referenced. integrated_multimodal_description: [Shot 1] 2D-animated, cinematic, the young woman shown in <Picture 1> stays where she stands on the wet night street, keeping her face, her long dark hair, her dark buttoned coat with the raised collar, and the warm amber lights behind her. Heavy rain keeps falling in diagonal streaks while loose strands of her hair lift and settle and her shoulders rise and fall with slow breathing. The camera holds a static shot for the whole duration. No other person enters the frame and the shot never cuts. overall_soundscape: Heavy rain falls on wet asphalt with a low distant traffic hum underneath; small drops tick against her coat. non_diegetic_music: N/A
Compare the rain direction and the hair drift against t1 - same words, different roll.
t3 — seed 66742
SLOP8760/W02/t3/66742
Seed
Seed: 66742
For the target video, at 0.00 seconds into the target video, <Picture 1> (from [Shot 1]) is fully referenced. integrated_multimodal_description: [Shot 1] 2D-animated, cinematic, the young woman shown in <Picture 1> stays where she stands on the wet night street, keeping her face, her long dark hair, her dark buttoned coat with the raised collar, and the warm amber lights behind her. Heavy rain keeps falling in diagonal streaks while loose strands of her hair lift and settle and her shoulders rise and fall with slow breathing. The camera holds a static shot for the whole duration. No other person enters the frame and the shot never cuts. overall_soundscape: Heavy rain falls on wet asphalt with a low distant traffic hum underneath; small drops tick against her coat. non_diegetic_music: N/A
Watch whether the coat collar and the amber lights stay where the still put them.
t4 — seed 66743
SLOP8760/W02/t4/66743
Seed
Seed: 66743
For the target video, at 0.00 seconds into the target video, <Picture 1> (from [Shot 1]) is fully referenced. integrated_multimodal_description: [Shot 1] 2D-animated, cinematic, the young woman shown in <Picture 1> stays where she stands on the wet night street, keeping her face, her long dark hair, her dark buttoned coat with the raised collar, and the warm amber lights behind her. Heavy rain keeps falling in diagonal streaks while loose strands of her hair lift and settle and her shoulders rise and fall with slow breathing. The camera holds a static shot for the whole duration. No other person enters the frame and the shot never cuts. overall_soundscape: Heavy rain falls on wet asphalt with a low distant traffic hum underneath; small drops tick against her coat. non_diegetic_music: N/A
Look for the seeds that drift the face: identity is the first thing a bad roll loses.
t5 — seed 66744
SLOP8760/W02/t5/66744
Seed
Seed: 66744
For the target video, at 0.00 seconds into the target video, <Picture 1> (from [Shot 1]) is fully referenced. integrated_multimodal_description: [Shot 1] 2D-animated, cinematic, the young woman shown in <Picture 1> stays where she stands on the wet night street, keeping her face, her long dark hair, her dark buttoned coat with the raised collar, and the warm amber lights behind her. Heavy rain keeps falling in diagonal streaks while loose strands of her hair lift and settle and her shoulders rise and fall with slow breathing. The camera holds a static shot for the whole duration. No other person enters the frame and the shot never cuts. overall_soundscape: Heavy rain falls on wet asphalt with a low distant traffic hum underneath; small drops tick against her coat. non_diegetic_music: N/A
Pick the take you would keep; that choice is the whole loop.
Week 3 — Text to Image
t1 — one line
SLOP8760/W03/t1/noseed
Prompt
a young woman standing in the rain at night

One line names the subject and nothing else: the model picks its own wide, distant framing.
t2 — setting and mood
SLOP8760/W03/t2/noseed
Prompt
a young woman with long black hair in a dark buttoned coat, standing on a wet city street at night in heavy rain under warm street lamps, calm and melancholy, looking back over her shoulder

Setting and mood fill the street with neon and a stray figure the prompt never asked for.
t3 — four-part grammar
SLOP8760/W03/t3/noseed
Prompt
subject: 1girl, solo, young woman, long black hair, wet hair, dark long coat, closed coat, high collar, calm expression, melancholy composition: cowboy shot, upper body, looking back, looking at viewer, over the shoulder, turning head light: night, heavy rain, rain streaks, wet street, reflective ground, warm street lamp glow, amber light, warm backlight, rim light style: anime screencap, movie still, cinematic lighting, depth of field, blurry background

Subject, framing, light and style anchor land the over-the-shoulder shot in one cool blue pass.
t4 — + tag vocabulary + negative prompt
SLOP8760/W03/t4/noseed
Prompt
masterpiece, best quality, amazing quality, very aesthetic, absurdres, newest, official art, anime screencap, movie still, 1girl, solo, young woman, cowboy shot, upper body, looking back, looking at viewer, over the shoulder, turning head, long black hair, wet hair, hair between eyes, floating hair, strands of hair on face, fair skin, delicate features, natural makeup, calm expression, restrained emotion, slightly parted lips, quiet gaze, melancholy, dark long coat, closed coat, buttoned coat, high collar, turtleneck, long sleeves, covered shoulders, fully clothed, damp clothes, wet collar, night, rain, heavy rain, rain streaks, falling raindrops, wet street, wet asphalt, puddle, reflective ground, reflection, city street at night, warm street lamp glow, amber light, warm backlight, rim light, warm light on skin, depth of field, shallow focus, blurry background, bokeh, cinematic lighting, cinematic composition, dramatic lighting, film grain
Negative prompt
worst quality, low quality, normal quality, lowres, bad anatomy, bad hands, bad proportions, extra digits, fewer digits, extra limbs, missing limbs, mutated hands, deformed, disfigured, ugly, blurry face, jpeg artifacts, signature, watermark, username, artist name, text, english text, chinese text, japanese text, korean text, letters, alphabet, numbers, logo, emblem, sign, signboard, storefront sign, neon sign, billboard, poster, banner, subtitles, speech bubble, dated, stamp, border, frame, cropped head, blue theme, cyan, teal, cold color, cool lighting, monochrome, greyscale, sepia, washed out, overexposed, bare shoulders, off shoulder, off-shoulder, bare arms, sleeveless, bare back, cleavage, undressing, exposed skin, swimsuit, lingerie, tank top, camisole, multiple girls, 2girls, crowd, group, male focus, chibi, 3d, photorealistic, sketch, unfinished, nsfw, nude, censored

The first negative prompt bans cold colour outright, and the whole frame turns warm and golden.
t5 — + shot size, camera angle, colour script
SLOP8760/W03/t5/noseed
Prompt
masterpiece, best quality, amazing quality, very aesthetic, absurdres, newest, official art, anime screencap, movie still, 1girl, solo, young woman, cowboy shot, upper body, looking back, looking at viewer, over the shoulder, turning head, long black hair, wet hair, hair between eyes, floating hair, strands of hair on face, fair skin, delicate features, natural makeup, calm expression, restrained emotion, slightly parted lips, quiet gaze, melancholy, dark long coat, closed coat, buttoned coat, high collar, turtleneck, long sleeves, covered shoulders, fully clothed, damp clothes, wet collar, night, rain, heavy rain, rain streaks, falling raindrops, wet street, wet asphalt, puddle, reflective ground, reflection, city street at night, warm street lamp glow, amber light, golden light, orange light, bronze tone, warm backlight, rim light, warm light on skin, abstract glowing light bands, blurred colored light spots, colorful bokeh, glowing haze, streaks of light on wet pavement, depth of field, shallow focus, blurry background, bokeh, cinematic lighting, cinematic composition, dramatic lighting, chiaroscuro, high contrast, dark theme, film grain, rule of thirds, head in upper third, empty space above head, headroom
Negative prompt
worst quality, low quality, normal quality, lowres, bad anatomy, bad hands, bad proportions, extra digits, fewer digits, extra limbs, missing limbs, mutated hands, deformed, disfigured, ugly, blurry face, jpeg artifacts, signature, watermark, username, artist name, text, english text, chinese text, japanese text, korean text, letters, alphabet, numbers, logo, emblem, sign, signboard, storefront sign, neon sign, billboard, poster, banner, subtitles, speech bubble, dated, stamp, border, frame, cropped head, blue theme, cyan, teal, cold color, cool lighting, monochrome, greyscale, sepia, washed out, overexposed, bare shoulders, off shoulder, off-shoulder, bare arms, sleeveless, bare back, cleavage, undressing, exposed skin, swimsuit, lingerie, tank top, camisole, multiple girls, 2girls, crowd, group, male focus, chibi, 3d, photorealistic, sketch, unfinished, nsfw, nude, censored

Rule-of-thirds headroom and a named colour script finish the keyframe this course reuses from here on.
Week 4 — Image to Image
t1 — rough sketch
SLOP8760/W04/t1/noseed
Image

masterpiece, best quality, amazing quality, very aesthetic, absurdres, newest, official art, anime screencap, movie still, 1girl, solo, young woman, cowboy shot, upper body, looking back, looking at viewer, over the shoulder, turning head, long black hair, wet hair, hair between eyes, floating hair, strands of hair on face, fair skin, delicate features, natural makeup, calm expression, restrained emotion, slightly parted lips, quiet gaze, melancholy, dark long coat, closed coat, buttoned coat, high collar, turtleneck, long sleeves, covered shoulders, fully clothed, damp clothes, wet collar, night, rain, heavy rain, rain streaks, falling raindrops, wet street, wet asphalt, puddle, reflective ground, reflection, city street at night, warm street lamp glow, amber light, golden light, orange light, bronze tone, warm backlight, rim light, warm light on skin, abstract glowing light bands, blurred colored light spots, colorful bokeh, glowing haze, streaks of light on wet pavement, depth of field, shallow focus, blurry background, bokeh, cinematic lighting, cinematic composition, dramatic lighting, chiaroscuro, high contrast, dark theme, film grain, rule of thirds, head in upper third, empty space above head, headroom
Negative prompt
worst quality, low quality, normal quality, lowres, bad anatomy, bad hands, bad proportions, extra digits, fewer digits, extra limbs, missing limbs, mutated hands, deformed, disfigured, ugly, blurry face, jpeg artifacts, signature, watermark, username, artist name, text, english text, chinese text, japanese text, korean text, letters, alphabet, numbers, logo, emblem, sign, signboard, storefront sign, neon sign, billboard, poster, banner, subtitles, speech bubble, dated, stamp, border, frame, cropped head, blue theme, cyan, teal, cold color, cool lighting, monochrome, greyscale, sepia, washed out, overexposed, bare shoulders, off shoulder, off-shoulder, bare arms, sleeveless, bare back, cleavage, undressing, exposed skin, swimsuit, lingerie, tank top, camisole, multiple girls, 2girls, crowd, group, male focus, chibi, 3d, photorealistic, sketch, unfinished, nsfw, nude, censored

The prompt is fixed from here to t5 and only a rough sketch is guiding this pass. The denoising strength was held constant too, but its value was never written down - this manifest carries no strength and no seed, in the one week whose whole point is that dial. Read it as the argument for the log rather than as a number to copy.
t2 — clean lineart
SLOP8760/W04/t2/noseed
Image

masterpiece, best quality, amazing quality, very aesthetic, absurdres, newest, official art, anime screencap, movie still, 1girl, solo, young woman, cowboy shot, upper body, looking back, looking at viewer, over the shoulder, turning head, long black hair, wet hair, hair between eyes, floating hair, strands of hair on face, fair skin, delicate features, natural makeup, calm expression, restrained emotion, slightly parted lips, quiet gaze, melancholy, dark long coat, closed coat, buttoned coat, high collar, turtleneck, long sleeves, covered shoulders, fully clothed, damp clothes, wet collar, night, rain, heavy rain, rain streaks, falling raindrops, wet street, wet asphalt, puddle, reflective ground, reflection, city street at night, warm street lamp glow, amber light, golden light, orange light, bronze tone, warm backlight, rim light, warm light on skin, abstract glowing light bands, blurred colored light spots, colorful bokeh, glowing haze, streaks of light on wet pavement, depth of field, shallow focus, blurry background, bokeh, cinematic lighting, cinematic composition, dramatic lighting, chiaroscuro, high contrast, dark theme, film grain, rule of thirds, head in upper third, empty space above head, headroom
Negative prompt
worst quality, low quality, normal quality, lowres, bad anatomy, bad hands, bad proportions, extra digits, fewer digits, extra limbs, missing limbs, mutated hands, deformed, disfigured, ugly, blurry face, jpeg artifacts, signature, watermark, username, artist name, text, english text, chinese text, japanese text, korean text, letters, alphabet, numbers, logo, emblem, sign, signboard, storefront sign, neon sign, billboard, poster, banner, subtitles, speech bubble, dated, stamp, border, frame, cropped head, blue theme, cyan, teal, cold color, cool lighting, monochrome, greyscale, sepia, washed out, overexposed, bare shoulders, off shoulder, off-shoulder, bare arms, sleeveless, bare back, cleavage, undressing, exposed skin, swimsuit, lingerie, tank top, camisole, multiple girls, 2girls, crowd, group, male focus, chibi, 3d, photorealistic, sketch, unfinished, nsfw, nude, censored

Cleaning the lines before the pass steadies the proportions the rough sketch left uncertain.
t3 — flat-coloured lineart
SLOP8760/W04/t3/noseed
Image

masterpiece, best quality, amazing quality, very aesthetic, absurdres, newest, official art, anime screencap, movie still, 1girl, solo, young woman, cowboy shot, upper body, looking back, looking at viewer, over the shoulder, turning head, long black hair, wet hair, hair between eyes, floating hair, strands of hair on face, fair skin, delicate features, natural makeup, calm expression, restrained emotion, slightly parted lips, quiet gaze, melancholy, dark long coat, closed coat, buttoned coat, high collar, turtleneck, long sleeves, covered shoulders, fully clothed, damp clothes, wet collar, night, rain, heavy rain, rain streaks, falling raindrops, wet street, wet asphalt, puddle, reflective ground, reflection, city street at night, warm street lamp glow, amber light, golden light, orange light, bronze tone, warm backlight, rim light, warm light on skin, abstract glowing light bands, blurred colored light spots, colorful bokeh, glowing haze, streaks of light on wet pavement, depth of field, shallow focus, blurry background, bokeh, cinematic lighting, cinematic composition, dramatic lighting, chiaroscuro, high contrast, dark theme, film grain, rule of thirds, head in upper third, empty space above head, headroom
Negative prompt
worst quality, low quality, normal quality, lowres, bad anatomy, bad hands, bad proportions, extra digits, fewer digits, extra limbs, missing limbs, mutated hands, deformed, disfigured, ugly, blurry face, jpeg artifacts, signature, watermark, username, artist name, text, english text, chinese text, japanese text, korean text, letters, alphabet, numbers, logo, emblem, sign, signboard, storefront sign, neon sign, billboard, poster, banner, subtitles, speech bubble, dated, stamp, border, frame, cropped head, blue theme, cyan, teal, cold color, cool lighting, monochrome, greyscale, sepia, washed out, overexposed, bare shoulders, off shoulder, off-shoulder, bare arms, sleeveless, bare back, cleavage, undressing, exposed skin, swimsuit, lingerie, tank top, camisole, multiple girls, 2girls, crowd, group, male focus, chibi, 3d, photorealistic, sketch, unfinished, nsfw, nude, censored

A flat colour guide pins the palette and the silhouette before a single gradient exists.
t4 — finished keyframe, refined
SLOP8760/W04/t4/noseed
Image

masterpiece, best quality, amazing quality, very aesthetic, absurdres, newest, official art, anime screencap, movie still, 1girl, solo, young woman, cowboy shot, upper body, looking back, looking at viewer, over the shoulder, turning head, long black hair, wet hair, hair between eyes, floating hair, strands of hair on face, fair skin, delicate features, natural makeup, calm expression, restrained emotion, slightly parted lips, quiet gaze, melancholy, dark long coat, closed coat, buttoned coat, high collar, turtleneck, long sleeves, covered shoulders, fully clothed, damp clothes, wet collar, night, rain, heavy rain, rain streaks, falling raindrops, wet street, wet asphalt, puddle, reflective ground, reflection, city street at night, warm street lamp glow, amber light, golden light, orange light, bronze tone, warm backlight, rim light, warm light on skin, abstract glowing light bands, blurred colored light spots, colorful bokeh, glowing haze, streaks of light on wet pavement, depth of field, shallow focus, blurry background, bokeh, cinematic lighting, cinematic composition, dramatic lighting, chiaroscuro, high contrast, dark theme, film grain, rule of thirds, head in upper third, empty space above head, headroom
Negative prompt
worst quality, low quality, normal quality, lowres, bad anatomy, bad hands, bad proportions, extra digits, fewer digits, extra limbs, missing limbs, mutated hands, deformed, disfigured, ugly, blurry face, jpeg artifacts, signature, watermark, username, artist name, text, english text, chinese text, japanese text, korean text, letters, alphabet, numbers, logo, emblem, sign, signboard, storefront sign, neon sign, billboard, poster, banner, subtitles, speech bubble, dated, stamp, border, frame, cropped head, blue theme, cyan, teal, cold color, cool lighting, monochrome, greyscale, sepia, washed out, overexposed, bare shoulders, off shoulder, off-shoulder, bare arms, sleeveless, bare back, cleavage, undressing, exposed skin, swimsuit, lingerie, tank top, camisole, multiple girls, 2girls, crowd, group, male focus, chibi, 3d, photorealistic, sketch, unfinished, nsfw, nude, censored

The input here is week03's own top-tier keyframe, refined rather than built again from a sketch.
t5 — finished keyframe + ControlNet pose lock
SLOP8760/W04/t5/noseed
Image

masterpiece, best quality, amazing quality, very aesthetic, absurdres, newest, official art, anime screencap, movie still, 1girl, solo, young woman, cowboy shot, upper body, looking back, looking at viewer, over the shoulder, turning head, long black hair, wet hair, hair between eyes, floating hair, strands of hair on face, fair skin, delicate features, natural makeup, calm expression, restrained emotion, slightly parted lips, quiet gaze, melancholy, dark long coat, closed coat, buttoned coat, high collar, turtleneck, long sleeves, covered shoulders, fully clothed, damp clothes, wet collar, night, rain, heavy rain, rain streaks, falling raindrops, wet street, wet asphalt, puddle, reflective ground, reflection, city street at night, warm street lamp glow, amber light, golden light, orange light, bronze tone, warm backlight, rim light, warm light on skin, abstract glowing light bands, blurred colored light spots, colorful bokeh, glowing haze, streaks of light on wet pavement, depth of field, shallow focus, blurry background, bokeh, cinematic lighting, cinematic composition, dramatic lighting, chiaroscuro, high contrast, dark theme, film grain, rule of thirds, head in upper third, empty space above head, headroom
Negative prompt
worst quality, low quality, normal quality, lowres, bad anatomy, bad hands, bad proportions, extra digits, fewer digits, extra limbs, missing limbs, mutated hands, deformed, disfigured, ugly, blurry face, jpeg artifacts, signature, watermark, username, artist name, text, english text, chinese text, japanese text, korean text, letters, alphabet, numbers, logo, emblem, sign, signboard, storefront sign, neon sign, billboard, poster, banner, subtitles, speech bubble, dated, stamp, border, frame, cropped head, blue theme, cyan, teal, cold color, cool lighting, monochrome, greyscale, sepia, washed out, overexposed, bare shoulders, off shoulder, off-shoulder, bare arms, sleeveless, bare back, cleavage, undressing, exposed skin, swimsuit, lingerie, tank top, camisole, multiple girls, 2girls, crowd, group, male focus, chibi, 3d, photorealistic, sketch, unfinished, nsfw, nude, censored

A bare pose skeleton, with no colour or line detail at all, still holds this composition in place.
Week 5 — Text to video
t1 — one line
SLOP8760/W05/t1/66740
Prompt
integrated_multimodal_description: [Shot 1] a young woman standing in the rain at night. overall_soundscape: Heavy rain falls on wet asphalt with a low distant traffic hum underneath; small drops tick against her coat. non_diegetic_music: N/A
One line gives the model everything to invent: watch the camera wander and the framing change.
t2 — subject + action
SLOP8760/W05/t2/66740
Prompt
integrated_multimodal_description: [Shot 1] 2D-animated, a young woman with long black hair in a dark buttoned coat stands on a wet city street at night in heavy rain under warm street lamps and looks back over her shoulder. overall_soundscape: Heavy rain falls on wet asphalt with a low distant traffic hum underneath; small drops tick against her coat. non_diegetic_music: N/A
Naming the subject and the action locks the pose; the light and the camera are still free.
t3 — four sentences
SLOP8760/W05/t3/66740
Prompt
integrated_multimodal_description: [Shot 1] 2D-animated, cinematic, a young woman with long dark hair and a dark buttoned coat with a raised collar stands on a wet night street looking back over her shoulder at the camera, with warm amber light bands and blurred colored light spots behind her. Heavy rain keeps falling in diagonal streaks while loose strands of her hair lift and settle and her shoulders rise and fall with slow breathing. The camera holds a static shot for the whole duration. No other person enters the frame and the shot never cuts. overall_soundscape: Heavy rain falls on wet asphalt with a low distant traffic hum underneath; small drops tick against her coat. non_diegetic_music: N/A
The four-sentence shape pins style, motion, camera and continuity - watch the shot stop cutting.
t4 — four sentences + beats + negatives
SLOP8760/W05/t4/66740
Prompt
integrated_multimodal_description: [Shot 1] 2D-animated, cinematic, a young woman with long dark hair and a dark buttoned coat with a raised collar stands on a wet night street looking back over her shoulder at the camera, with warm amber light bands and blurred colored light spots behind her. Heavy rain keeps falling in diagonal streaks while loose strands of her hair lift and settle and her shoulders rise and fall with slow breathing. The camera holds a static shot for the whole duration. No other person enters the frame and the shot never cuts. From 00:00.000 to 00:02.000 she holds completely still except for breathing; at 00:02.000 she blinks once; at 00:03.500 a gust lifts the hair on the left side of her face; by 00:05.000 the hair has settled back. There is no lettering, sign, or logo anywhere in the frame; she holds no umbrella; the coat stays buttoned to the collar; her feet do not move; the camera does not move. overall_soundscape: Heavy rain falls on wet asphalt with a low distant traffic hum underneath; small drops tick against her coat. non_diegetic_music: N/A
Timecoded beats put the blink and the gust where you asked; check 00:02 and 00:03.5.
t5 — contradictory prompt (counter-example)
SLOP8760/W05/t5/66740
Prompt
integrated_multimodal_description: [Shot 1] Live-action photoreal, 2D-animated, watercolor, claymation, vintage film; the camera pushes in with large amplitude at fast speed, the camera holds a static shot, the camera arcs around her at fast speed, a POV shot from directly above; a young woman with long dark hair and a dark buttoned coat with a raised collar stands on a wet night street looking back over her shoulder at the camera, with warm amber light bands and blurred colored light spots behind her. Heavy rain keeps falling in diagonal streaks while loose strands of her hair lift and settle and her shoulders rise and fall with slow breathing. The camera holds a static shot for the whole duration. No other person enters the frame and the shot never cuts. Bright noon daylight, deep night, golden hour, a neon cyberpunk blue palette, a black-and-white silent-film look. overall_soundscape: Heavy rain falls on wet asphalt with a low distant traffic hum underneath; small drops tick against her coat. non_diegetic_music: N/A
Every camera move and every look contradicts the last one - watch the model pick for you.
Deliberate counter-example.
Week 6 — Image to video
t1 — sketch as first frame
SLOP8760/W06/t1/66740
Image

For the target video, at 0.00 seconds into the target video, <Picture 1> (from [Shot 1]) is fully referenced. integrated_multimodal_description: [Shot 1] 2D-animated, the young woman in <Picture 1> stands in the rain at night while the rain falls and her hair moves; the camera holds a static shot. overall_soundscape: Heavy rain falls on wet asphalt with a low distant traffic hum underneath; small drops tick against her coat. non_diegetic_music: N/A
A line sketch as the first frame: watch the model paint the whole look from nothing.
t2 — flat colour as first frame
SLOP8760/W06/t2/66740
Image

For the target video, at 0.00 seconds into the target video, <Picture 1> (from [Shot 1]) is fully referenced. integrated_multimodal_description: [Shot 1] 2D-animated, the young woman in <Picture 1> stands in the rain at night while the rain falls and her hair moves; the camera holds a static shot. overall_soundscape: Heavy rain falls on wet asphalt with a low distant traffic hum underneath; small drops tick against her coat. non_diegetic_music: N/A
Flat colour fixes the palette and the silhouette; the face is still up for grabs.
t3 — finished key frame
SLOP8760/W06/t3/66740
Image

For the target video, at 0.00 seconds into the target video, <Picture 1> (from [Shot 1]) is fully referenced. integrated_multimodal_description: [Shot 1] 2D-animated, the young woman in <Picture 1> stands in the rain at night while the rain falls and her hair moves; the camera holds a static shot. overall_soundscape: Heavy rain falls on wet asphalt with a low distant traffic hum underneath; small drops tick against her coat. non_diegetic_music: N/A
A finished frame holds the identity - compare the face here with t1 and t2.
t4 — finished key frame + four sentences
SLOP8760/W06/t4/66740
Image

For the target video, at 0.00 seconds into the target video, <Picture 1> (from [Shot 1]) is fully referenced. integrated_multimodal_description: [Shot 1] 2D-animated, cinematic, the young woman shown in <Picture 1> stays where she stands on the wet night street, keeping her face, her long dark hair, her dark buttoned coat with the raised collar, and the warm amber lights behind her. Heavy rain keeps falling in diagonal streaks while loose strands of her hair lift and settle and her shoulders rise and fall with slow breathing. The camera holds a static shot for the whole duration. No other person enters the frame and the shot never cuts. overall_soundscape: Heavy rain falls on wet asphalt with a low distant traffic hum underneath; small drops tick against her coat. non_diegetic_music: N/A
Same first frame as t3, richer prompt: the picture holds the look, the words hold the motion.
Same recorded result as week02-t1.
t5 — first frame + last frame (FL2VA)
SLOP8760/W06/t5/66740
Image pair


How the reference pictures align with the target video — Picture 1 (from Shot 1) aligns with the 0.00-second mark of the target video; Picture 2 (from Shot 1) aligns with the 5.17-second mark of the target video. integrated_multimodal_description: [Shot 1] 2D-animated, cinematic, the young woman begins in the position and framing established by Picture 1, looking back over her shoulder on the wet night street, keeping her face, her long dark hair, her dark buttoned coat with the raised collar, and the warm amber lights behind her. Heavy rain keeps falling in diagonal streaks while she turns her whole body around toward the camera, her hair swinging across her shoulder and settling, the raised collar shifting and falling back, her feet staying in place. The camera holds a static shot for the whole duration. By the end of the shot she faces the camera squarely with both shoulders level and her eyes on the lens, settling into the pose, spacing, and composition established by Picture 2. No other person enters the frame and the shot never cuts. overall_soundscape: Heavy rain falls on wet asphalt with a low distant traffic hum underneath; small drops tick against her coat. non_diegetic_music: N/A
Two anchors instead of one: watch the turn land exactly on the tail frame at 5.17 s.
Week 7 — Your own workflow
t1 — official T2V template
SLOP8760/W07/t1/66741
Workflow graph
integrated_multimodal_description: [Shot 1] 2D-animated, cinematic, a young woman with long dark hair and a dark buttoned coat with a raised collar stands on a wet night street looking back over her shoulder at the camera, with warm amber light bands and blurred colored light spots behind her. Heavy rain keeps falling in diagonal streaks while loose strands of her hair lift and settle and her shoulders rise and fall with slow breathing. The camera holds a static shot for the whole duration. No other person enters the frame and the shot never cuts. overall_soundscape: Heavy rain falls on wet asphalt with a low distant traffic hum underneath; small drops tick against her coat. non_diegetic_music: N/A
The stock text-to-video graph, untouched: this is your baseline shot.
t2 — Ref2VA, two reference slots
SLOP8760/W07/t2/66741
Workflow graph
subject_definitions: <Subject 1> is the young woman whose face, long dark hair, and dark buttoned coat with the raised collar come from <Picture 1>, and whose full-length proportions, coat length, and boots come from <Picture 2>. summary: [reference generation] The target video is a single static shot of <Subject 1> standing on a wet night street in heavy rain, looking back over her shoulder at the camera, with warm amber light bands and blurred colored light spots behind her. retention_analysis: <Subject 1> (appears in [Shot 1]): fully_preserved - the face, hair, coat, collar, and boots are retained. detailed_description: The target video is a 2D-animated cinematic style with a warm amber palette on a dark rainy street and no on-screen text. [Shot 1] 2D-animated, cinematic, <Subject 1> stands on a wet night street looking back over her shoulder at the camera, with warm amber light bands and blurred colored light spots behind her. Heavy rain keeps falling in diagonal streaks while loose strands of her hair lift and settle and her shoulders rise and fall with slow breathing. The camera holds a static shot for the whole duration. No other person enters the frame and the shot never cuts. overall_soundscape: Heavy rain falls on wet asphalt with a low distant traffic hum underneath; small drops tick against her coat. non_diegetic_music: N/A
Same shot through the reference graph: the identity comes from the pictures, not the words.
t3 — t1 upscaled x2 to 1080x1920
SLOP8760/W07/t3/66741
Upscale graph
No new generation - the t1 clip through Real-ESRGAN, run after the render and outside any graph, so this tier ships command-line steps rather than a workflow: watch the line quality, not the content.
t4 — I2VA from the key frame
SLOP8760/W07/t4/66741
Workflow graph
For the target video, at 0.00 seconds into the target video, <Picture 1> (from [Shot 1]) is fully referenced. integrated_multimodal_description: [Shot 1] 2D-animated, cinematic, the young woman shown in <Picture 1> stays where she stands on the wet night street, keeping her face, her long dark hair, her dark buttoned coat with the raised collar, and the warm amber lights behind her. Heavy rain keeps falling in diagonal streaks while loose strands of her hair lift and settle and her shoulders rise and fall with slow breathing. The camera holds a static shot for the whole duration. No other person enters the frame and the shot never cuts. overall_soundscape: Heavy rain falls on wet asphalt with a low distant traffic hum underneath; small drops tick against her coat. non_diegetic_music: N/A
The same shot again from a first frame: compare which of the three holds the face best.
Same recorded result as week02-t2.
Week 8 — Continuity
t1 — 1 reference: face
SLOP8760/W08/t1/66740
Reference set

subject_definitions: <Subject 1> is the young woman whose face comes from <Picture 1>. summary: [reference generation] The target video is a two-shot scene: <Subject 1> waits dry under a shop awning at the edge of a wet night street, then steps out into the rain and looks back over her shoulder at the camera. retention_analysis: <Subject 1> (appears in [Shot 1], [Shot 2]): fully_preserved - the face is retained. detailed_description: The target video is a 2D-animated cinematic style with a warm amber palette on a dark rainy street and no on-screen text. [Shot 1] A medium-full shot from chest height. <Subject 1> stands dry under a shop awning at the edge of the street, her hair dry and loose, her dark coat buttoned to the collar, both hands in the coat pockets, her feet fully visible on the dry paving under the awning. Rain falls beyond the awning's edge at frame-right, and warm amber light bands and blurred colored light spots fill the background. She looks frame-left down the street, then takes one step forward to the edge of the awning and stops with her toes at the line where the rain begins. The camera holds a static shot. [Shot 2] At 00:07.000, the shot cuts to a close-up at eye level. <Subject 1> has stepped out into the rain: her hair is wet and plastered to her cheeks, drops run down her face, and the raised collar is dark with water. She turns her head to look back over her shoulder at the camera and holds the look, blinking once, while the rain keeps falling in diagonal streaks across the frame. The camera holds a static shot to the end. overall_soundscape: Rain falls steadily on the awning fabric and the wet street with a low distant traffic hum underneath; one footstep lands on the paving, followed by rain ticking against her coat. non_diegetic_music: N/A
One face reference across a two-shot cut: watch the coat and the street change at 00:07.
t2 — 2 references: face + full body
SLOP8760/W08/t2/66740
Reference set


subject_definitions: <Subject 1> is the young woman whose face, long dark hair, and dark buttoned coat with the raised collar come from <Picture 1>, and whose full-length proportions, coat length, and boots come from <Picture 2>. summary: [reference generation] The target video is a two-shot scene: <Subject 1> waits dry under a shop awning at the edge of a wet night street, then steps out into the rain and looks back over her shoulder at the camera. retention_analysis: <Subject 1> (appears in [Shot 1], [Shot 2]): fully_preserved - the face, the long dark hair, the dark buttoned coat with the raised collar, the full-length proportions, the coat length, and the boots are retained. detailed_description: The target video is a 2D-animated cinematic style with a warm amber palette on a dark rainy street and no on-screen text. [Shot 1] A medium-full shot from chest height. <Subject 1> stands dry under a shop awning at the edge of the street, her hair dry and loose, her dark coat buttoned to the collar, both hands in the coat pockets, her feet fully visible on the dry paving under the awning. Rain falls beyond the awning's edge at frame-right, and warm amber light bands and blurred colored light spots fill the background. She looks frame-left down the street, then takes one step forward to the edge of the awning and stops with her toes at the line where the rain begins. The camera holds a static shot. [Shot 2] At 00:07.000, the shot cuts to a close-up at eye level. <Subject 1> has stepped out into the rain: her hair is wet and plastered to her cheeks, drops run down her face, and the raised collar is dark with water. She turns her head to look back over her shoulder at the camera and holds the look, blinking once, while the rain keeps falling in diagonal streaks across the frame. The camera holds a static shot to the end. overall_soundscape: Rain falls steadily on the awning fabric and the wet street with a low distant traffic hum underneath; one footstep lands on the paving, followed by rain ticking against her coat. non_diegetic_music: N/A
Adding the full-length picture fixes the coat length and the boots across the cut.
t3 — 4 references: + dry state
SLOP8760/W08/t3/66740
Reference set




subject_definitions: <Subject 1> is the young woman whose face, long dark hair, and dark buttoned coat with the raised collar come from <Picture 1>, and whose full-length proportions, coat length, and boots come from <Picture 2>. Her dry state — dry loose hair and a dry coat — comes from <Picture 3> and <Picture 4>. summary: [reference generation] The target video is a two-shot scene: <Subject 1> waits dry under a shop awning at the edge of a wet night street, then steps out into the rain and looks back over her shoulder at the camera. retention_analysis: <Subject 1> (appears in [Shot 1], [Shot 2]): fully_preserved - the face, the long dark hair, the dark buttoned coat with the raised collar, the full-length proportions, the coat length, the boots, and the dry state are retained. detailed_description: The target video is a 2D-animated cinematic style with a warm amber palette on a dark rainy street and no on-screen text. [Shot 1] A medium-full shot from chest height. <Subject 1> stands dry under a shop awning at the edge of the street, her hair dry and loose, her dark coat buttoned to the collar, both hands in the coat pockets, her feet fully visible on the dry paving under the awning. Rain falls beyond the awning's edge at frame-right, and warm amber light bands and blurred colored light spots fill the background. She looks frame-left down the street, then takes one step forward to the edge of the awning and stops with her toes at the line where the rain begins. The camera holds a static shot. [Shot 2] At 00:07.000, the shot cuts to a close-up at eye level. <Subject 1> has stepped out into the rain: her hair is wet and plastered to her cheeks, drops run down her face, and the raised collar is dark with water. She turns her head to look back over her shoulder at the camera and holds the look, blinking once, while the rain keeps falling in diagonal streaks across the frame. The camera holds a static shot to the end. overall_soundscape: Rain falls steadily on the awning fabric and the wet street with a low distant traffic hum underneath; one footstep lands on the paving, followed by rain ticking against her coat. non_diegetic_music: N/A
The dry pair is what makes Shot 1 read as dry - check the hair before she steps out.
t4 — 5 references: + scene
SLOP8760/W08/t4/66740
Reference set





subject_definitions: <Subject 1> is the young woman whose face, long dark hair, and dark buttoned coat with the raised collar come from <Picture 1>, and whose full-length proportions, coat length, and boots come from <Picture 2>. Her dry state — dry loose hair and a dry coat — comes from <Picture 3> and <Picture 4>. <Subject 2> is the wet night street in <Picture 5>, with warm amber light bands, blurred colored light spots, and reflections on the wet asphalt. summary: [reference generation] The target video is a two-shot scene: <Subject 1> waits dry under a shop awning at the edge of a wet night street, then steps out into the rain and looks back over her shoulder at the camera. retention_analysis: <Subject 1> (appears in [Shot 1], [Shot 2]): fully_preserved - the face, the long dark hair, the dark buttoned coat with the raised collar, the full-length proportions, the coat length, the boots, and the dry state are retained. <Subject 2> (appears in [Shot 1], [Shot 2]): fully_preserved - the wet night street, the warm amber light bands, the blurred colored light spots, and the reflections on the wet asphalt are retained. detailed_description: The target video is a 2D-animated cinematic style with a warm amber palette on a dark rainy street and no on-screen text. [Shot 1] A medium-full shot from chest height. <Subject 1> stands dry under a shop awning at the edge of the street of <Subject 2>, her hair dry and loose, her dark coat buttoned to the collar, both hands in the coat pockets, her feet fully visible on the dry paving under the awning. Rain falls beyond the awning's edge at frame-right, and warm amber light bands and blurred colored light spots fill the background. She looks frame-left down the street, then takes one step forward to the edge of the awning and stops with her toes at the line where the rain begins. The camera holds a static shot. [Shot 2] At 00:07.000, the shot cuts to a close-up at eye level. <Subject 1> has stepped out into the rain: her hair is wet and plastered to her cheeks, drops run down her face, and the raised collar is dark with water. She turns her head to look back over her shoulder at the camera and holds the look, blinking once, while the rain keeps falling in diagonal streaks across the frame. The camera holds a static shot to the end. overall_soundscape: Rain falls steadily on the awning fabric and the wet street with a low distant traffic hum underneath; one footstep lands on the paving, followed by rain ticking against her coat. non_diegetic_music: N/A
With the street pinned too, both shots should look like the same place at the same hour.
Week 9 — Script and storyboard
t1 — prose
SLOP8760/W09/t1/66740
Script
subject_definitions: <Subject 1> is Heracles in <Picture 1>: a bearded man in his early thirties with short curling dark hair, a scar through the left eyebrow, a lion-pelt cloak with the lion's head hood pushed back, a linen tunic, leather wrist wraps, and a short bronze sword; soaked, with black hydra blood on his face and pelt. <Subject 2> is the Lernaean hydra in <Picture 2>: a serpentine body rising from black water with eight viper-like necks of dark green scales, and a ninth head held highest at the center, a third larger, with gold eyes and a crown of bone spikes. <Subject 3> is the Lerna swamp at night in <Picture 3>: a muddy shore with a row of burning torch stakes, a chest-high boulder at the water's edge, black still water, reeds, dead trees, and thin mist. summary: [reference generation] The target video is a three-shot sequence in which the ninth head of <Subject 2> knocks <Subject 1> down into the water of <Subject 3>, he pulls a burning torch stake from the water, and he presses the fire to a fresh neck stump so that it chars and stops regrowing. retention_analysis: <Subject 1> (appears in [Shot 1], [Shot 2], [Shot 3]): fully_preserved - face, beard, scar, lion pelt, tunic, wrist wraps, sword, and the wet blood-spattered state are retained. <Subject 2> (appears in [Shot 1], [Shot 3]): fully_preserved - the body, the dark green scales, and the crowned ninth head are retained. <Subject 3> (appears in [Shot 1], [Shot 2], [Shot 3]): fully_preserved - the torch stakes, black water, reeds, and mist are retained. detailed_description: The target video is a painterly realistic CG action style with thick-paint texture, night, lit only by warm amber torchlight over black swamp water, with no on-screen text. Night in the Lerna swamp. Heracles is waist-deep in black water, ringed by the hydra's heads, the shore torches burning behind him. The largest head — gold-eyed, crowned with bone — shoulders the others aside and slams into him; he goes down, one hand catching a torch stake. He drags the burning stake up out of the water, firelight sweeping across his face. His blade takes the next neck and he presses the fire to the stump at once. The flesh chars black and smokes. Nothing grows back. overall_soundscape: Black swamp water sloshes and drips, hydra hisses overlap, one heavy blow lands, a torch flame flares, and burning flesh hisses; a low wind moves through the reeds. non_diegetic_music: N/A
Prose leaves the order of events to the model: check whether the fire really comes after the fall.
t2 — beat sheet
SLOP8760/W09/t2/66740
Script
subject_definitions: <Subject 1> is Heracles in <Picture 1>: a bearded man in his early thirties with short curling dark hair, a scar through the left eyebrow, a lion-pelt cloak with the lion's head hood pushed back, a linen tunic, leather wrist wraps, and a short bronze sword; soaked, with black hydra blood on his face and pelt. <Subject 2> is the Lernaean hydra in <Picture 2>: a serpentine body rising from black water with eight viper-like necks of dark green scales, and a ninth head held highest at the center, a third larger, with gold eyes and a crown of bone spikes. <Subject 3> is the Lerna swamp at night in <Picture 3>: a muddy shore with a row of burning torch stakes, a chest-high boulder at the water's edge, black still water, reeds, dead trees, and thin mist. summary: [reference generation] The target video is a three-shot sequence in which the ninth head of <Subject 2> knocks <Subject 1> down into the water of <Subject 3>, he pulls a burning torch stake from the water, and he presses the fire to a fresh neck stump so that it chars and stops regrowing. retention_analysis: <Subject 1> (appears in [Shot 1], [Shot 2], [Shot 3]): fully_preserved - face, beard, scar, lion pelt, tunic, wrist wraps, sword, and the wet blood-spattered state are retained. <Subject 2> (appears in [Shot 1], [Shot 3]): fully_preserved - the body, the dark green scales, and the crowned ninth head are retained. <Subject 3> (appears in [Shot 1], [Shot 2], [Shot 3]): fully_preserved - the torch stakes, black water, reeds, and mist are retained. detailed_description: The target video is a painterly realistic CG action style with thick-paint texture, night, lit only by warm amber torchlight over black swamp water, with no on-screen text. BEAT SHEET — "Fire" (12 s, 3 beats, no dialogue) Setting: Lerna swamp, night. Torch stakes burning on the shore. Heracles waist-deep in black water, hydra heads closing from three sides. Beat 1 (0–4 s) KNOCKED DOWN — The ninth head (a third larger, gold eyes, bone crown) shoves the other heads aside and strikes his chest. Result state: Heracles is down in the water, his right hand on a torch stake at the shore. Beat 2 (4–7 s) THE FIRE — His hand closes on the burning stake and pulls it up out of the water. Firelight crosses his face. The decision is made. Beat 3 (7–12 s) SEALED — The blade cuts a neck; the torch is pressed to the stump in the same second. It chars black and smokes. The stump stops swelling: no new heads. Sound: swamp water, hydra hisses, one blow, the hiss of burning flesh. No music change. overall_soundscape: Black swamp water sloshes and drips, hydra hisses overlap, one heavy blow lands, a torch flame flares, and burning flesh hisses; a low wind moves through the reeds. non_diegetic_music: N/A
Beats give each moment a result state: watch 0-4 s for 'down in the water, hand on the stake'.
t3 — beat sheet + shot list
SLOP8760/W09/t3/66740
Script
subject_definitions: <Subject 1> is Heracles in <Picture 1>: a bearded man in his early thirties with short curling dark hair, a scar through the left eyebrow, a lion-pelt cloak with the lion's head hood pushed back, a linen tunic, leather wrist wraps, and a short bronze sword; soaked, with black hydra blood on his face and pelt. <Subject 2> is the Lernaean hydra in <Picture 2>: a serpentine body rising from black water with eight viper-like necks of dark green scales, and a ninth head held highest at the center, a third larger, with gold eyes and a crown of bone spikes. <Subject 3> is the Lerna swamp at night in <Picture 3>: a muddy shore with a row of burning torch stakes, a chest-high boulder at the water's edge, black still water, reeds, dead trees, and thin mist. summary: [reference generation] The target video is a three-shot sequence in which the ninth head of <Subject 2> knocks <Subject 1> down into the water of <Subject 3>, he pulls a burning torch stake from the water, and he presses the fire to a fresh neck stump so that it chars and stops regrowing. retention_analysis: <Subject 1> (appears in [Shot 1], [Shot 2], [Shot 3]): fully_preserved - face, beard, scar, lion pelt, tunic, wrist wraps, sword, and the wet blood-spattered state are retained. <Subject 2> (appears in [Shot 1], [Shot 3]): fully_preserved - the body, the dark green scales, and the crowned ninth head are retained. <Subject 3> (appears in [Shot 1], [Shot 2], [Shot 3]): fully_preserved - the torch stakes, black water, reeds, and mist are retained. detailed_description: The target video is a painterly realistic CG action style with thick-paint texture, night, lit only by warm amber torchlight over black swamp water, with no on-screen text. BEAT SHEET — "Fire" (12 s, 3 beats, no dialogue) Setting: Lerna swamp, night. Torch stakes burning on the shore. Heracles waist-deep in black water, hydra heads closing from three sides. Beat 1 (0–4 s) KNOCKED DOWN — The ninth head (a third larger, gold eyes, bone crown) shoves the other heads aside and strikes his chest. Result state: Heracles is down in the water, his right hand on a torch stake at the shore. Beat 2 (4–7 s) THE FIRE — His hand closes on the burning stake and pulls it up out of the water. Firelight crosses his face. The decision is made. Beat 3 (7–12 s) SEALED — The blade cuts a neck; the torch is pressed to the stump in the same second. It chars black and smokes. The stump stops swelling: no new heads. SHOT LIST S8 0.0–4.0 s MEDIUM, facing him, camera at water level, slow push-in. The ninth head enters from top-left, shoves two smaller heads out of frame and hits him in the chest. He falls back into the water; his right hand lands on a torch stake at the shore. Hold on the result: down, hand on the stake, the ninth head looming above. S9 4.0–7.0 s CLOSE-UP on the hand and the stake, slight low angle, static camera. The hand tightens and pulls the burning stake head up out of the water surface; the flame flares; firelight sweeps across his wet face in the top of frame. S10 7.0–12.0 s MEDIUM, side profile, waist up, static camera. The blade crosses frame left to right — cut — a neck stump in front of him; the torch is pressed onto it within the same second. Flesh chars black, smoke rises, the stump stays sealed. He does not look at the camera. Cuts: hard cuts only, on the beat. No slow motion. No new characters. No text. Sound: swamp water, hydra hisses, one blow, the hiss of burning flesh. No music change. overall_soundscape: Black swamp water sloshes and drips, hydra hisses overlap, one heavy blow lands, a torch flame flares, and burning flesh hisses; a low wind moves through the reeds. non_diegetic_music: N/A
The shot list fixes size and angle: 4-7 s should be a close-up on the hand, and the cuts should land on 4.0 and 7.0.
t4 — storyboard (+ 3 frames)
SLOP8760/W09/t4/66740
Script
subject_definitions: <Subject 1> is Heracles in <Picture 1>: a bearded man in his early thirties with short curling dark hair, a scar through the left eyebrow, a lion-pelt cloak with the lion's head hood pushed back, a linen tunic, leather wrist wraps, and a short bronze sword; soaked, with black hydra blood on his face and pelt. <Subject 2> is the Lernaean hydra in <Picture 2>: a serpentine body rising from black water with eight viper-like necks of dark green scales, and a ninth head held highest at the center, a third larger, with gold eyes and a crown of bone spikes. <Subject 3> is the Lerna swamp at night in <Picture 3>: a muddy shore with a row of burning torch stakes, a chest-high boulder at the water's edge, black still water, reeds, dead trees, and thin mist. <Picture 4> is a storyboard reference for [Shot 1], defining its viewpoint and the result state: <Subject 1> down in the water with his right hand on a torch stake and the crowned head above him. <Picture 5> is a storyboard reference for [Shot 2], defining its viewpoint: the hand and the burning stake pulled from the water, the wet face lit from below. <Picture 6> is a storyboard reference for [Shot 3], defining its viewpoint: side profile, the torch pressed to the charred stump. summary: [reference generation] The target video is a three-shot sequence in which the ninth head of <Subject 2> knocks <Subject 1> down into the water of <Subject 3>, he pulls a burning torch stake from the water, and he presses the fire to a fresh neck stump so that it chars and stops regrowing. retention_analysis: <Subject 1> (appears in [Shot 1], [Shot 2], [Shot 3]): fully_preserved - face, beard, scar, lion pelt, tunic, wrist wraps, sword, and the wet blood-spattered state are retained. <Subject 2> (appears in [Shot 1], [Shot 3]): fully_preserved - the body, the dark green scales, and the crowned ninth head are retained. <Subject 3> (appears in [Shot 1], [Shot 2], [Shot 3]): fully_preserved - the torch stakes, black water, reeds, and mist are retained. <Picture 4> ([Shot 1] storyboard): fully_preserved - viewpoint and subject placement are followed. <Picture 5> ([Shot 2] storyboard): fully_preserved - viewpoint and subject placement are followed. <Picture 6> ([Shot 3] storyboard): fully_preserved - viewpoint and subject placement are followed. detailed_description: The target video is a painterly realistic CG action style with thick-paint texture, night, lit only by warm amber torchlight over black swamp water, with no on-screen text. BEAT SHEET — "Fire" (12 s, 3 beats, no dialogue) Setting: Lerna swamp, night. Torch stakes burning on the shore. Heracles waist-deep in black water, hydra heads closing from three sides. Beat 1 (0–4 s) KNOCKED DOWN — The ninth head (a third larger, gold eyes, bone crown) shoves the other heads aside and strikes his chest. Result state: Heracles is down in the water, his right hand on a torch stake at the shore. Beat 2 (4–7 s) THE FIRE — His hand closes on the burning stake and pulls it up out of the water. Firelight crosses his face. The decision is made. Beat 3 (7–12 s) SEALED — The blade cuts a neck; the torch is pressed to the stump in the same second. It chars black and smokes. The stump stops swelling: no new heads. SHOT LIST S8 0.0–4.0 s MEDIUM, facing him, camera at water level, slow push-in. The ninth head enters from top-left, shoves two smaller heads out of frame and hits him in the chest. He falls back into the water; his right hand lands on a torch stake at the shore. Hold on the result: down, hand on the stake, the ninth head looming above. S9 4.0–7.0 s CLOSE-UP on the hand and the stake, slight low angle, static camera. The hand tightens and pulls the burning stake head up out of the water surface; the flame flares; firelight sweeps across his wet face in the top of frame. S10 7.0–12.0 s MEDIUM, side profile, waist up, static camera. The blade crosses frame left to right — cut — a neck stump in front of him; the torch is pressed onto it within the same second. Flesh chars black, smoke rises, the stump stays sealed. He does not look at the camera. STORYBOARD FRAMES (attached) Frame 1 = S8 result state (<Picture 4>): start S8 on the strike and end exactly on this composition — him down in the water, right hand on the stake, the crowned head above. Frame 2 = S9 (<Picture 5>): match this framing for the whole shot — the hand and the burning stake, firelight on the face at the top of frame. Frame 3 = S10 (<Picture 6>): match this framing — side profile, torch pressed to the charred stump. Keep the character, the pelt, the sword and the torch exactly as in the frames. Cuts: hard cuts only, on the beat. No slow motion. No new characters. No text. Sound: swamp water, hydra hisses, one blow, the hiss of burning flesh. No music change. overall_soundscape: Black swamp water sloshes and drips, hydra hisses overlap, one heavy blow lands, a torch flame flares, and burning flesh hisses; a low wind moves through the reeds. non_diegetic_music: N/A
Three attached frames on top of t3: check the torch in S9 and the profile in S10 against the boards.
The Cut — four windows on one master
Master · 1080x1920
Every clip below is a window cut out of this one file at 24 fps; no clip was generated separately.
S3
SLOP8760/W11/establishing/noseed
S3 · 00:07.000-00:11.000 · 4s
S1
SLOP8760/W11/hook/noseed
S1 · 00:00.000-00:03.000 · 3s
S6
SLOP8760/W11/reaction/noseed
S6 · 00:19.000-00:22.000 · 3s
S10
SLOP8760/W11/turn/noseed
S10 · 00:33.000-00:38.000 · 5s
Reference episode — The Ninth Head
Segment A · S1-S4 · seed 66740
subject_definitions: <Subject 1> is Heracles, whose face, beard, scar through the left eyebrow, lion-pelt cloak with the hood pushed back, linen tunic, leather wrist wraps, and short bronze sword come from <Picture 1>, and whose full-length proportions and sandals come from <Picture 2>. <Subject 2> is the Lernaean hydra in <Picture 3>: a serpentine body as thick as a fallen oak rising from black water, eight viper-like necks with dark green scales that glint gold in firelight, pale ridged throats and open jaws. <Subject 3> is the ninth head of the hydra in <Picture 3>: held highest at the center, a third larger than the others, with gold slit-pupil eyes and a crown of pale bone spikes. <Subject 4> is the Lerna swamp at night in <Picture 4>: a muddy shore in the foreground with a row of six burning torch stakes running from frame-left into the distance, a chest-high boulder at the water's edge at frame-right, black still water, reeds, dead trees, and thin mist. <Picture 5> is the first frame of [Shot 1], showing a smaller hydra head lunging at the camera with a bronze blade entering from frame-right. <Picture 6> is a storyboard reference for [Shot 3], defining its viewpoint from the shore, the torch stakes in the foreground, <Subject 1> small in the midground, and <Subject 2> fanned out in the background. summary: [keyframe completion + reference generation] The target video is a four-shot opening: a hydra head lunges at the camera and is cut off, <Subject 1> stands knee-deep in the water of <Subject 4> with <Subject 2> fanned out above him, a wide shot from the shore shows the whole standoff, and a fresh neck stump splits and grows two new heads. retention_analysis: <Subject 1> (appears in [Shot 2], [Shot 3]): fully_preserved - face, beard, scar, lion pelt, tunic, wrist wraps, and sword are retained; he is dry above the waist. <Subject 2> (appears in [Shot 1], [Shot 2], [Shot 3], [Shot 4]): fully_preserved - the body, the dark green scales, the pale throats, and the fanned necks are retained. <Subject 3> (appears in [Shot 2], [Shot 3]): fully_preserved - the larger crowned head with gold eyes stays highest at the center. <Subject 4> (appears in [Shot 2], [Shot 3], [Shot 4]): fully_preserved - the torch stakes, boulder, black water, reeds, and mist are retained. <Picture 5> ([Shot 1] first frame): fully_preserved - the lunging head and the entering blade start the shot exactly as shown. <Picture 6> ([Shot 3] storyboard): fully_preserved - the viewpoint, layering, and placement are followed. detailed_description: The target video is a painterly realistic CG action style with thick-paint texture on skin and scales, night, lit only by warm amber torchlight from a row of stakes on the shore over black swamp water, with thin mist and no on-screen text. [Shot 1] An extreme close-up at water level, static shot; the shot begins from <Picture 5>. One of the smaller heads of <Subject 2> lunges straight at the camera with its jaws wide, fangs and pale throat lit by the firelight, filling the upper two-thirds of the frame. The bronze blade of <Subject 1> enters from frame-right mid-swing, passes through the neck in one stroke, and the severed head tumbles down and out of the bottom of the frame into the black water with a heavy splash, leaving the cut neck swaying in the top of the frame. [Shot 2] At 00:03.000, the shot cuts to a medium shot from a low angle. <Subject 1>, dry above the waist, stands knee-deep in the black water of <Subject 4> facing frame-right, his right arm raised with the dripping bronze sword, the lion-pelt hood pushed back behind his neck. Behind and above him the eight necks of <Subject 2> fan out over the reeds against the dark sky, and <Subject 3>, the larger crowned head, is held highest at the center directly above him. The torch stakes burn on the shore at frame-left. His feet stay planted; his chest rises once with a breath; the heads sway slowly and the jaws open and close. The camera holds a static shot. [Shot 3] At 00:07.000, the shot cuts to a wide shot from the shore, following <Picture 6>. In the foreground the row of burning torch stakes of <Subject 4> runs from frame-left into the distance; in the midground <Subject 1> stands small in the frame, knee-deep in the water, the sword raised; in the background <Subject 2> rises from the water with its necks spread in a half circle around him, <Subject 3> at the center of the arc. The necks lean inward toward him and ripples spread outward from the body across the water. The camera holds a static shot. [Shot 4] At 00:11.000, the shot cuts to a close-up of a fresh neck stump of <Subject 2> just above the water, the cut surface facing the camera in the firelight. The surface bubbles, swells, and splits down the middle; two new heads push out of it side by side, their scales still wet and pale, and both open their jaws toward the camera as the shot ends. The camera holds a static shot. overall_soundscape: Black swamp water sloshes and drips against reeds, torch flames crackle on the shore, hydra hisses rise and overlap, and the bronze blade rings on each cut; a low wind moves through dead trees, and heavy splashes follow each fall into the water. non_diegetic_music: N/A
S1-S4, 00:00.000-00:15.000 on the finished timeline.
Segment B · S5-S8 · seed 66740
subject_definitions: <Subject 1> is Heracles, whose face, beard, scar through the left eyebrow, lion-pelt cloak, linen tunic, leather wrist wraps, and short bronze sword come from <Picture 1> and <Picture 2>; he is soaked, with black hydra blood spattered on his face and the chest of the pelt. <Subject 2> is the Lernaean hydra in <Picture 3>: a serpentine body rising from black water with many viper-like necks of dark green scales, pale ridged throats and open jaws. <Subject 3> is the ninth head of the hydra in <Picture 3>: a third larger than the others, with gold slit-pupil eyes and a crown of pale bone spikes. <Subject 4> is the Lerna swamp at night in <Picture 4>: a muddy shore with a row of burning torch stakes, black still water, reeds, and thin mist. <Picture 5> is a storyboard reference for [Shot 4], defining its result state: <Subject 1> fallen back into waist-deep water with his right hand on a torch stake at frame-right and <Subject 3> looming above him at the top-left. summary: [reference generation] The target video is a four-shot sequence: <Subject 1> cuts two heads, a close-up shows his face as the stumps swell again, the heads multiply and drive him back toward the shore of <Subject 4>, and <Subject 3> knocks him down into the water beside a torch stake. retention_analysis: <Subject 1> (appears in [Shot 1], [Shot 2], [Shot 3], [Shot 4]): fully_preserved - face, beard, scar, lion pelt, tunic, wrist wraps, sword, and the wet blood-spattered state are retained. <Subject 2> (appears in [Shot 1], [Shot 3], [Shot 4]): fully_preserved - the body and the dark green scaled necks are retained; the number of heads increases within the sequence. <Subject 3> (appears in [Shot 4]): fully_preserved - the larger crowned head with gold eyes is retained. <Subject 4> (appears in [Shot 1], [Shot 3], [Shot 4]): fully_preserved - the torch stakes, black water, reeds, and mist are retained. <Picture 5> ([Shot 4] storyboard): fully_preserved - the result state and placement are followed. detailed_description: The target video is a painterly realistic CG action style with thick-paint texture on skin and scales, night, lit only by warm amber torchlight from a row of stakes on the shore over black swamp water, with thin mist and no on-screen text. [Shot 1] A medium shot from the side, static shot. <Subject 1> stands knee-deep in the black water of <Subject 4> facing frame-right, and two necks of <Subject 2> hang in the frame in front of him. He swings the bronze sword twice; after the first stroke one head drops into the water with a burst of spray, and after the second stroke the other head follows, leaving two cut necks swaying above the water at the end of the shot. [Shot 2] At 00:04.000, the shot cuts to a close-up of the face of <Subject 1>, facing frame-right, black blood spattered across his left cheek and beard, the wet sword tip in the lower corner of the frame with water dripping from it. His eyes are locked on something at frame-right out of the shot; his jaw is set and his breathing is heavy. The camera holds a static shot; his eyes do not leave that point. [Shot 3] At 00:07.000, the shot cuts to a wide shot from a high angle looking down. <Subject 2> now has more than a dozen necks, closing in on <Subject 1> from three sides; he is waist-deep in the black water and has been driven back until the torch stakes of <Subject 4> on the shore are directly behind him at frame-left. The necks press inward and the water churns around him. The camera holds a static shot. [Shot 4] At 00:11.000, the shot cuts to a medium shot at water level facing him. <Subject 3>, the larger crowned head with gold eyes, enters from the top-left, shoves two smaller heads out of the frame, and strikes him in the chest. He falls backward into the waist-deep water with a heavy splash; his right hand lands on a burning torch stake at the shore at frame-right and grips it. The shot ends on the result state of <Picture 5>: <Subject 1> down in the water, his hand on the stake, <Subject 3> looming above him with its jaws open. The camera holds a static shot. overall_soundscape: Black swamp water sloshes and drips against reeds, torch flames crackle on the shore, hydra hisses rise and overlap, and the bronze blade rings on each cut; a low wind moves through dead trees, and heavy splashes follow each fall into the water. non_diegetic_music: N/A
S5-S8, 00:15.000-00:30.000 on the finished timeline.
Segment C · S9-S12 · seed 66740
subject_definitions: <Subject 1> is Heracles in <Picture 1>: a bearded man with short curling dark hair, a scar through the left eyebrow, a lion-pelt cloak, a linen tunic, leather wrist wraps, and a short bronze sword; soaked, with black hydra blood on his face and pelt. <Subject 2> is the Lernaean hydra in <Picture 2>: a serpentine body of dark green scales collapsing into black water, its necks ending in charred black smoking stumps. <Subject 3> is the ninth head of the hydra, whose crowned form on the body comes from <Picture 2> and whose severed form lying on the mud comes from <Picture 3>: gold slit-pupil eyes, a crown of pale bone spikes, and a clean dry cut at the neck with no blood. <Subject 4> is the Lerna swamp at night in <Picture 4>: a muddy shore with a row of burning torch stakes, black still water, reeds, and thin mist. <Picture 5> is a storyboard reference for [Shot 1], defining its framing: the right hand of <Subject 1> gripping a burning torch stake pulled from the water, his wet face in the upper part of the frame lit from below. <Picture 6> is a storyboard reference for [Shot 2], defining its framing: <Subject 1> in side profile pressing the burning torch onto a fresh neck stump. summary: [reference generation] The target video is a four-shot sequence: <Subject 1> pulls a burning torch stake from the water of <Subject 4>, presses the fire to a fresh neck stump of <Subject 2> so that it chars and stops regrowing, the remaining necks are sealed one after another until only <Subject 3> stands, and <Subject 3> is cut down onto the mud where it keeps biting. retention_analysis: <Subject 1> (appears in [Shot 1], [Shot 2], [Shot 3]): fully_preserved - face, beard, scar, lion pelt, wrist wraps, sword, and the wet blood-spattered state are retained. <Subject 2> (appears in [Shot 2], [Shot 3]): fully_preserved - the body and the charred smoking stumps are retained. <Subject 3> (appears in [Shot 3], [Shot 4]): fully_preserved - the crowned head with gold eyes is retained on the body and then severed on the mud with a dry cut. <Subject 4> (appears in [Shot 1], [Shot 3], [Shot 4]): fully_preserved - the torch stakes, black water, mud, reeds, and mist are retained. <Picture 5> ([Shot 1] storyboard): fully_preserved - framing and hand placement are followed. <Picture 6> ([Shot 2] storyboard): fully_preserved - the side profile and the torch on the stump are followed. detailed_description: The target video is a painterly realistic CG action style with thick-paint texture on skin and scales, night, lit only by warm amber torchlight from a row of stakes on the shore over black swamp water, with thin mist and no on-screen text. [Shot 1] A close-up from a slight low angle, following <Picture 5>. The right hand of <Subject 1>, wrapped in leather, is closed around the burning head of a torch stake and pulls it up out of the water surface of <Subject 4>; water streams off the wood and the flame flares as it clears the surface, and the firelight sweeps upward across his wet face in the upper third of the frame. The camera holds a static shot. [Shot 2] At 00:03.000, the shot cuts to a medium shot in side profile, following <Picture 6>. <Subject 1>, waist up, faces frame-right with the burning torch in one hand and the sword in the other. The sword crosses the frame from left to right in one stroke and a fresh neck stump of <Subject 2> is left in front of him; in the same second he presses the burning torch onto the stump. The flesh chars black, smoke rises past his face, and the stump stays sealed without swelling. His face is lit from below by the torch; he does not look at the camera. The camera holds a static shot. [Shot 3] At 00:08.000, the shot cuts to a wide shot from the shore, the torch stakes of <Subject 4> in the foreground. The necks of <Subject 2> are charred black and smoking one after another, and the body slides sideways and collapses into the black water; only <Subject 3>, the crowned head, is still raised at the center above the smoke, with <Subject 1> small in the midground holding the torch and the sword. The camera holds a static shot. [Shot 4] At 00:12.000, the shot cuts to a close-up on the mud at the water's edge. <Subject 3> lies severed on the black mud, a third larger than a man's torso, its gold slit-pupil eyes open and its crown of bone spikes catching the firelight; the jaws open and close twice, biting at nothing, and the cut at the neck is clean and dry with no blood. The camera holds a static shot to the end. overall_soundscape: Black swamp water sloshes and drips against reeds, torch flames crackle on the shore, hydra hisses rise and overlap, and the bronze blade rings on each cut; a low wind moves through dead trees, and heavy splashes follow each fall into the water. non_diegetic_music: N/A
S9-S12, 00:30.000-00:45.000 on the finished timeline.
Segment D · S13-S16 · seed 66742
subject_definitions: <Subject 1> is Heracles in <Picture 1>: a bearded man with short curling dark hair, a scar through the left eyebrow, a lion-pelt cloak, a linen tunic, leather wrist wraps, and sandals, holding a burning torch stake in his left hand and a short bronze sword in his right; soaked, with black hydra blood on his face and pelt. <Subject 2> is the Lerna swamp shore, whose night state with burning torch stakes and a chest-high boulder standing at the water's edge comes from <Picture 2>, and whose pre-dawn state with the torches burnt to embers, a thin grey-blue band of sky at the top, and the boulder tipped over in the mud with a thin line of gold light under it comes from <Picture 3>. <Subject 3> is the body of the Lernaean hydra in <Picture 4>: dark green scales, necks ending in charred black smoking stumps, sinking into black water. <Picture 5> is a storyboard reference for [Shot 2], defining its wide viewpoint: the boulder lying in the mud with the gold light under it, <Subject 1> small beside it, and <Subject 3> sinking behind. summary: [reference generation] The target video is a four-shot ending: <Subject 1> plants the torch in the mud and pushes the boulder of <Subject 2> over onto the ground at frame-bottom, a wide shot shows the boulder settled with a line of gold light under it as dawn begins and <Subject 3> sinks, he walks away toward the torch stakes, and the boulder shifts a hand's width before the picture cuts to black. retention_analysis: <Subject 1> (appears in [Shot 1], [Shot 2], [Shot 3]): fully_preserved - face, beard, scar, lion pelt, tunic, wrist wraps, sandals, torch, sword, and the wet blood-spattered state are retained. <Subject 2> (appears in [Shot 1], [Shot 2], [Shot 3], [Shot 4]): fully_preserved - the shore, the torch stakes, and the boulder are retained; the state changes from night to pre-dawn across the shots. <Subject 3> (appears in [Shot 2]): fully_preserved - the charred stumps and the sinking body are retained. <Picture 5> ([Shot 2] storyboard): fully_preserved - the wide viewpoint and placement are followed. detailed_description: The target video is a painterly realistic CG action style with thick-paint texture on skin and scales, night, lit only by warm amber torchlight from a row of stakes on the shore over black swamp water, with thin mist and no on-screen text. [Shot 1] A medium shot from behind and over the shoulder of <Subject 1>, night state of <Subject 2>. He pushes the burning torch stake into the mud at frame-left so that it stands upright, then sets both hands against the chest-high boulder at the water's edge in front of him, lowers his shoulders, and drives forward with his back and legs; the boulder tips, rolls forward, and falls out of the bottom of the frame with a deep wet thud, mud sliding after it. The camera holds a static shot. [Shot 2] At 00:05.000, the shot cuts to a wide shot from the shore, following <Picture 5>. The boulder lies settled in the mud at the water's edge with a thin line of warm gold light showing under it; <Subject 1> stands beside it, small in the frame, the torch in his hand burning low; behind him the body of <Subject 3> sinks slowly into the black water, its stumps smoking; the torch stakes of <Subject 2> have burnt down to embers and a thin grey-blue band of pre-dawn shows at the top of the frame. Mist drifts over the water. The camera holds a static shot. [Shot 3] At 00:09.000, the shot cuts to a medium shot from behind. <Subject 1> turns away from the boulder and walks toward the row of ember-red torch stakes along the shore, his back to the camera, each step splashing through the shallow water at the edge of the mud, the low torch swinging at his side. The camera holds a static shot as he moves away and grows smaller. [Shot 4] At 00:13.000, the shot cuts to a close-up of the boulder lying in the mud, the pre-dawn state of <Subject 2>, the thin line of gold light under its edge. The boulder shifts a hand's width toward the camera, wet mud slides down from its base, and the gold light under it brightens for an instant; then the picture cuts to black for the final frames. The camera holds a static shot. overall_soundscape: Black swamp water sloshes and drips against reeds, torch flames crackle on the shore, hydra hisses rise and overlap, and the bronze blade rings on each cut; a low wind moves through dead trees, and heavy splashes follow each fall into the water. non_diegetic_music: N/A
S13-S16, 00:45.000-01:00.000 on the finished timeline.
episode
SLOP8760/W10/episode/noseed
60s, 1440 frames, four segments concatenated and upscaled to 1080x1920.