THE MONSTER FILM METHOD

How I made a Monster ad.How I made
a Monster ad.

How I built my Monster spec ad, step by step: locking the assets, the three blocks I pasted into every scene, how I chained the scenes together, and the full prompt that got a werewolf out of a monitor. Plus the sound.

Made on Higgsfield with GPT Image 2.5, Seedream 5.0 Pro and Seedance 2.5.

The breakout.

Every scene in the film is one Seedance 2.5 generation of 6 to 30 seconds, with the cuts written into the prompt. The same images, the same character block and the same room map go into every one, and each scene starts on the last frame of the one before. That is what keeps one girl, one werewolf and one bedroom the same for two and a half minutes.

01

Lock it before you shoot.

Before any video, I locked every asset as an image. I cast her first, one photo, then turned that photo into a character sheet: front, back and a close-up. Then his sheet, the room from four angles with the grade already in it, and every prop, from the can to the lantern. Each one went into a register with its ID, and only those IDs ever went into a shot.

Her sheet
Her sheet
His sheet
His sheet
The room plate
The room plate
  • Images: GPT Image 2.5 for the sheets, Seedream 5.0 Pro for the room. Edit in small passes, one change per prompt.
  • Continuity: write the rules down once (his height against hers, where the can is, which side the door is) and keep them next to the register.

AMake her character sheet

Seedream 5.0 Pro · attach one photo of your character as image 1 · 16:9 · 2K

A soft, low-contrast film character sheet with quiet, delicate texture: the entire image lives in a broad, gentle middle register: highlights roll off early into soft cream while keeping full detail, shadows sit lifted, open and airy, micro-contrast stays low and tender, skin and fabric resolve with smooth, quiet, organic detail, edges stay rounded and creamy under a gentle veil of optical diffusion, as through an old uncoated vintage cinema prime, with a soft halation bloom breathing from the brightest points: the whole sheet reading like a delicately printed film photograph with a modern neutral colour grade.

Within that look, and using the attached image strictly as the identity reference for the person: their face, age, build, hair and costume are canon and must match the reference 100%: generate the character sheet: three stills of this same person arranged side by side on one sheet: FULL-BODY FRONT view on the left, FULL-BODY BACK view in the middle, and a THREE-QUARTER portrait on the right. The same person, identical outfit and proportions across all three panels, standing on a seamless PURE WHITE studio background that keeps its own even, luminous brightness, reading as clean paper; full figure in frame head to shoes with clear space above the head and below the feet, a soft contact shadow grounding them. Carry over only the person themselves: they stand alone in clean, empty white studio space, lifted out of the reference's scene and simply relit.

The character, anchored to the reference: exactly the person in the attached image, at their exact age and build, with their exact face carried over one-to-one as photographed: the precise hairline and haircut, the natural asymmetry between the left and right halves of the face, the true set of the eyes, nose and mouth, every crease, mark, shadow and skin texture kept as in the reference, real skin softly resolved: the same lived-in cinematic face, simply relit. They wear their reference costume exactly: every garment, its colours, closures, layers, accessories, folds and wear marks carried over consistent with the reference. Below the reference's frame edge the costume is completed faithfully in the same style, period and quality as the visible garments, down to period-true footwear. They stand NEUTRALLY with empty, relaxed hands: weight even on both feet, arms at their sides, shoulders level, head straight; expression neutral and at rest, mouth closed, face calm, eyes open and readable with soft catchlights.

FRONT panel: facing camera square, the costume's front reading clearly: its closures, collar, accessories and colours exactly as in the reference. BACK panel: turned fully away: the haircut from behind, the costume's back completed faithfully in the same period and construction, the sleeves at their sides, the trouser or hem line, the heels of the footwear. THREE-QUARTER portrait panel: a CHEST-LEVEL close portrait: the frame runs from just above the top of the head down to the upper chest, the bottom edge cutting across the chest, so the head and face fill most of the panel; face turned three-quarters to camera, calm and still, with the reference's exact asymmetry and every facial feature carried over unchanged: hairline, creases, skin and all.

The light is neutral and even across all three panels: a soft ambient wash, as if the whole space sits under a broad overcast sky: directionless, calm, arriving frontal and even, wrapping the figure gently, open transparent shadows, every fabric reading true, the exposure holding the low-contrast filmic register declared above. The figure's outline meets the pure white background softly along every edge. The colour story comes from the reference itself: the costume's dominant colour as the leading mass against the pure white field, its secondary tones as the supporting masses, and the smallest bright detail of the reference costume as the single accent.

Approved wardrobe completion: preserve the selected oversized off-white printed T-shirt and black over-ear headset exactly as visible in the reference. Below the shirt, she wears relaxed black cotton-jersey shorts with their hems visibly extending below the tee to mid-thigh, off-white crew socks and lightly worn off-white low-top canvas trainers with pale gray rubber soles. Shirt back is plain off-white. No hoodie or armour on this base outfit, no bag.

Then three blocks of words go into every video prompt, unchanged. Who she is, word for word (below). How she acts: a short profile of how she sits, moves and reacts, so the model plays a person instead of posing one. And where everything is: one paragraph that pins every wall, door and prop, so the room never rearranges itself. You can see all three working inside the breakout prompt further down.

BHer block, word for word

Paste this into every video prompt, unchanged. Same words, same girl, every scene.

a 24-year-old East Asian woman, slim, long straight black hair past her chest with a centre part, soft oval face, warm brown eyes, black over-ear gaming headset, oversized off-white tee with a black thorny graphic print, black shorts, off-white crew socks

02

Shoot it scene by scene.

Each scene is one Seedance 2.5 generation in omni reference, 6 to 30 seconds, with its cuts written in: SHOT 1 (0:00 to 0:03), SHOT 2, and so on. The scene prompts below show the full shape.

  • Drafts: run four at 480p, pick the best, and finalise only that one at 1080p.
  • The chain: attach the last frame of the scene before as the first image, so the next scene starts exactly where the last one ended. For big scenes, attach the exact last frame you need too, so it ends where the next one starts.
  • Fixes: change one thing per new draft. If it still fails after a few, change the plan, not the wording.
  • Big action: for the fight I drew a top-down plan of the room with both their marks and the camera positions, and attached it as an image.
03

Getting him out of the monitor.

Every time I asked for his whole body to come through the screen, he came out like a ghost, or climbed out from behind the monitor. It took ten prompt versions to find the fix, and the fix was a change of plan, not of wording: only his head and arms ever break the glass. Cut to her backing into the corner. Cut back and he is already out on the floor. Your brain fills in the rest.

The claw tears the M

CThe breakout prompt

Seedance 2.5 · omni reference · 5 images: the room at the start, her, him, the can, a height lineup · 16:9 · 20 seconds · draft first

SCENE CONTEXT: 3am in New York. MAYA is leaning in at her desk pressing the power button of her small desktop monitor; the screen has gone dark. Something inside hits the glass. One clawed hand tears the Monster claw M through the screen, and the werewolf forces its way out of the small monitor, drops to the floor and corners her against the wall. One scene, full real-time speed, never slow motion.

ACTIVE REFERENCES: @image1 is exactly how the room, the desk, the small monitor and MAYA are at the start. This exact slim black monitor on its stand is in every shot, same size and place; never copy this framing. @image2 is MAYA: a 24-year-old East Asian woman, long straight black hair with a centre part, black over-ear gaming headset, oversized off-white tee with a black thorny graphic print, black shorts, off-white crew socks. @image3 is the WEREWOLF: heavyweight black werewolf, huge shoulders, small wolf head with tall ears, dense black fur, grey cloth wraps on both forearms, a dark waist wrap, a long bushy tail, digitigrade legs, thick black claws, dark crimson eyes. He has exactly two arms and two hands. @image4 is the CAN: its green M is three jagged claw slashes, the exact shape his one hand tears through the screen. @image5 is the height lineup: her head comes level with his stomach.

FORMAT MODE: seven shots, hard cuts at 3s, 5.5s, 7.5s, 11s, 13.5s and 16.5s; every cut changes the angle and the shot size.

SHOT 1 (0:00 to 0:03): medium from beside the desk, her and the monitor in frame. She is holding the power button. A thud from inside: a white crack spiders across the screen. She takes her hand off and steps back from the desk, and goes still, watching it.

SHOT 2 (0:03 to 0:05.5): close, straight on the small monitor, the whole solid black screen and bezel in frame: ONE black clawed hand punches through the glass, its three claws out, and rips straight down, tearing three jagged slashes that form the Monster claw M exactly like @image4. One hand, three claws. Glass splinters onto the desk; green light comes out of the slashes only.

SHOT 3 (0:05.5 to 0:07.5): medium on MAYA from the desk side, green light from the M across her: she backs away a few steps, eyes on the screen, her breathing getting quicker.

SHOT 4 (0:07.5 to 0:11): close, straight on the small monitor, the torn M glowing, only darkness and green light inside it. His wet black snout punches out through the middle slash, solid and physical, glass bursting around it; his head forces out to the neck, the strips of glass between the slashes snapping off around it, and his two clawed arms punch out through the outer slashes and slam down on the desk, claws gouging the wood. He snarls. The monitor stays whole on its stand around him. The shot ends here, with only his head and arms out.

SHOT 5 (0:11 to 0:13.5): close on MAYA backing fast into the corner beside the pink lantern until her back hits the wall; off screen, a huge crash of the desk splintering and something enormous landing on the floor, the floor shaking, green light flickering across her; she flinches hard at the crash, breathing fast and hard, chest heaving, her eyes following it down to the floor.

SHOT 6 (0:13.5 to 0:16.5): wide over MAYA's shoulder, low: MAYA is in the near foreground pressed against the wall beside the pink lantern, breathing hard, her shoulder and head in frame; across the room he is already out, crouched on the floor in front of the desk, his back to the desk; the small monitor still stands on the splintered desk behind him, intact, the torn M glowing in its screen. He rises out of the crouch to his full height, hunching under the ceiling, and turns toward her.

SHOT 7 (0:16.5 to 0:20): close on MAYA standing against the wall beside the pink lantern: his head comes into the frame from the right, close to her face, and he gives one short deep snarl, lips pulled back; her hair stirs; she turns her face away and shuts her eyes, still and silent. Then he pulls his head back up and out of the frame, leaving her pressed to the wall, shaking, breathing hard and fast.

PHYSICS: only the monitor glass and plastic break, only where his claws and hands push. Shards fall at real speed. He comes only out of the front of the small screen. She moves only by her own feet and hands.

LIGHTING: dark room; the pink lantern from the back-left corner; after the glass breaks, green light comes out of the M and the hole and spills across the desk and room. No neon outlines, no sparks.

AUDIO: the computer fan, a thud inside the monitor, glass cracking, glass ripping as the claw tears down, shards tinkling, plastic cracking, a wet growl, the keyboard crashing, a heavy thud as he lands, one short deep snarl, his breathing, her fast hard breathing getting louder through the scene. She never screams. No long roar, no screaming. No music, no subtitles.

CHARACTER DOSSIER, MAYA: twenty-four. When something impossible happens a real person does not perform: at first she goes quiet and still; then as it comes out the fear takes her body over: her breathing gets faster and harder, she shakes, she presses herself into the wall. By the end she is breathing hard, as if she has been running.

LET IT SHOW: the fear is in her breathing and her body, rising through the scene; her face stays real, never posed. Lips may part as her breathing gets hard. A scream, a gasp for the camera, a wide-eyed horror face or a grimace = failed take.

STYLE: photoreal live-action, 35mm film, organic grain, practical-effects look, real glass, real fur, full real-time speed, never slow motion, never CG gloss.

POSITIVE CONSTRAINTS: one small monitor, the same in every shot, solid and opaque. Exactly one person and one werewolf with exactly two hands. The M is torn by one hand. The monitor's bezel and frame never break; only its glass tears. Nothing of him is visible inside the screen before he breaks out; we never see his body pass through the monitor, only his head and arms, then him already out on the floor. The camera never shows a door. The wall behind the monitor never breaks. She ends standing against the wall. No words or text on screen.

04

The power-up.

The drink that woke him is the drink that arms her. She can't fight him, so she slips under his reach, slides into the fridge and chugs while he punches the top in. The armour builds out of green pixels into real metal. The note that made it work: she has three seconds, so she moves and drinks like it.

The chug and the armour
05

The knock.

The joke is in what she doesn't do. Tight on her face, still panting from the fight, her mum knocks and she answers a bit too casually. Then her eyes drop to the sword and she bends out of frame, and the next shot picks her up mid-bend, tearing the armour off. That match on action is what makes the cut feel smooth.

"Yeah, one minute!"
06

Sound, then the finish.

The sound sells it more than the picture. Keep every real sound the takes give you, add very little, and plan where the silence goes. Then grade the whole film once, upscale it, and lay one layer of film grain over the top so every shot sits in the same world.

  • Grade: pull back the shots that jump warm or cool, deepen the blacks only where they look milky, then a soft bloom on the highlights.
  • 4K: Topaz on Higgsfield, 2160p.
  • Grain: add it after the upscale and after the grade, never before.

SThe sound recipe

ElevenLabs for the voices · the filters work in any editor

THE MUFFLED VOICES THROUGH THE DOOR
1. Write the argument as a two-voice dialogue (ElevenLabs text-to-dialogue, eleven_v3) with short, natural lines. Mum: "What was all that banging?" Her: "Nothing! I just... knocked something over." Mum: "At three in the morning?"
2. While the door is open, keep the first line clear: cut below 300 Hz and above 3,500 Hz and add a short slap echo (35 ms) so it sounds like the hallway.
3. The moment the door slams, switch to a heavy low-pass at 450 Hz, twice. Now it sounds like it's coming through the wood. Start her reply right on the slam so it goes muffled mid-line.
4. End it before the next beat lands. Here the argument fades out just before his gurgle, so the twist plays in silence.

THE MUSIC
1. Let the scene play with no music until it has earned it. Here the song comes in straight after the snarl.
2. Bring it in muffled (low-pass around 450 Hz) like she's still in shock, and open the filter across her drink so it's fully open on the green burst.
3. Stop it dead just before a beat, so the knock lands in silence.

THE RISER
One rising drone from the claw tearing through to the snarl, cut dead at its peak 0.06 seconds before the snarl. Silence, then the snarl.

TAKE IT WITH YOU

Every prompt, one file.

The sheet prompt, her block, the breakout prompt and the sound recipe, ready for your own monster.

THE 30-DAY CHALLENGE

One bedroom was
the warm-up.

The 30-day challenge goes way past spec ads. You build the work brands actually pay for, and the creative studio that sells it.

  • Ads
  • Product shots
  • Brand campaigns
  • UGC
  • Short films
  • Content for clients
  • 30 days, in a dated round with limited places
  • A weekly Q&A video and a review of your work, watched any time
  • A certificate with your name on it when you finish
  • The waitlist gets in 48 hours before everyone else
Join the waitlist Free to join. You hear first.
Waitlist gets in first
Waviboy W signature

A FREE WAVIBOY RESOURCE

How I made a Monster ad.

Enter your email to open the method behind the film, the prompts and the sound recipe.

Back to free resources ↗