five hats.

Vertical 9:16, photoreal, 20 seconds, seven shots. One trick: a hat swipes across the lens in one natural move, the outfit changes behind it while it fills the frame, and the hat lands on his head. The swaps get quicker as the take goes on.


FORMAT:
20.0 s, 9:16, seven shots. Shot A 0.0-2.4 s, shot B 2.4-6.7 s, shot C 6.7-8.8 s, shot D 8.8-10.6 s, shot E 10.6-12.2 s, shot F 12.2-13.7 s, shot G 13.7-20.0 s. A to B is a clean hard cut to a closer camera angle. Every later cut happens inside the moment a hat fills the whole frame, so it is never seen. The take ends while the yellow hard hat fills the whole frame. His voice comes from @Audio 1. No music.


ONE PERSON:
There is only one person in this entire piece, @you-in-grey. His face comes only from @selfie, the same face in all seven shots. Every hand and every arm in the piece is his. This rule overrides every reference.


THE TWO MOST IMPORTANT RULES IN THIS SHOT:
1. When a hat passes the lens, it fills 100% of the frame edge to edge for at least 4 frames: no wall, no clothing, no skin and no light visible in any corner. The hat passes about 2 cm from the glass, so it is many times wider than the frame, soft and motion-blurred.
2. Each hat in shots B to F is one continuous, quick move: grab, swing past the lens, land on his head. In shot G the hard hat makes the same move the other way: off his head and straight at the lens. He never stops, never holds a hat still in front of the camera, never shows it to the lens.
IF ANY CORNER SHOWS THE WALL OR HIS CLOTHES WHILE A HAT IS AT THE LENS, THE SHOT IS WRONG.
IF HE HOLDS A HAT STILL IN FRONT OF THE LENS, THE SHOT IS WRONG.
IF HIS FACE CHANGES BETWEEN THE SHOTS, THE SHOT IS WRONG.
IF THE WALL TURNS LIGHT GREY OR WHITE, THE SHOT IS WRONG.
IF THE ROBOT ARM APPEARS BEFORE SHOT G, THE SHOT IS WRONG.


REFERENCES (take only what is named, nothing else):
@selfie: the selfie, 100% as shown. Take ONLY his face and his hair. Not the clothes, not the background.
@you-in-grey: Raham, 100% as shown in the reference image. Take ONLY his grey oversized tee over the white long sleeve, his black cargo pants, his cream sneakers and his build. Take no face and no hair from it.
@hat-tower: hat tower, 100% as shown in the reference image. Take ONLY the five stacked hats, for shot A.
@grad-outfit: 100% as shown in the reference image. Take ONLY the navy gown with gold trim, the gold stole, the white collar and the black graduation cap for shot C. Take no face from it.
@artist-outfit: 100% as shown in the reference image. Take ONLY the open paint-splattered white shirt over the white ribbed tank and the brown beret for shot D. Take no face from it.
@coder-outfit: 100% as shown in the reference image. Take ONLY the black hoodie with the hood up and the black cap worn forward under it for shot E. Take no face from it.
@sales-outfit: 100% as shown in the reference image. Take ONLY the dark three-piece suit, the white shirt open at the collar with no tie, and the black fedora for shot F. Take no face from it.
@cyborg-outfit: 100% as shown in the reference image. Take ONLY the yellow hard hat, the hi-vis vest over the grey tee, and the robotic left arm for shot G. Take no face from it.
@studio: studio and chair, 100% as shown in the reference image. Take ONLY the cool blue-grey wall, the soft frontal light, the floor and the black chair.
@Audio 1: his voice. Sync his lips to it exactly.
From no reference take its background, its lighting, its colour, its sheet layout, its panel borders or any other person.


BODY AND SCALE:
Real adult proportions, his own build as in @selfie, never slimmed. Head about one eighth of his height. Every hat is normal real-life size for an adult head: the fedora fits his head snugly, and the tower in shot A is about one and a half times the height of his head, not bigger. The graduation cap's flat square top is wider than his face and as wide as his two hands side by side. The beret is a little wider than his head. The robot arm is as thick and as long as his real arm.


RAHAM - FACE LOCK, EXACT FEATURES:
@selfie controls the frontal face and the profile 100%. Zero drift. The same man in every shot and every frame. If anything below contradicts what the generator wants to produce, the reference and this description win.
Every feature exactly as in @selfie: face shape, eyes, brows, nose, mouth, facial hair, hairline and skin tone, with real texture and natural pores. Never narrower, longer, slimmer, younger or beautified.


SPATIAL LAYOUT:
Shot A: the black chair turned at an angle to the camera, the blue-grey wall about 2 m behind it. Shots B to G: he stands about 70 cm from the lens, the wall about 2 m behind him. The graduation cap, the black cap and the hard hat wait just outside the left edge of the frame; the beret and the fedora wait just outside the right edge, all within easy reach of his hands.


ACTION TIMING:
Shot A, 0.0-2.4 s, his grey outfit, the five-hat tower on his head.
He sits back in the chair, casual and relaxed, shoulders loose, knees apart, feet flat on the floor, hands resting loosely on his legs. He says: "Nothing I'm wearing in this video is real." The hat tower moves gently with his body, all five hats together as one, a split second late.
Shot B, 2.4-6.7 s, closer angle, his grey outfit, no hat.
2.4-4.5 s: he says: "Okay, I'm real. Mostly." A slight shrug on "Mostly".
4.5-5.8 s: he leans a little toward the lens and says: "Let me explain." On "explain" his right hand reaches out of the left edge of the frame.
5.8-6.7 s: his right hand comes back already gripping the graduation cap by its board and, without stopping, swings it in one arc up and across toward his head. From 6.4 s the board passes 2 cm in front of the lens and fills the whole frame, black and motion-blurred.
Shot C, 6.7-8.8 s, the gown and stole from @grad-outfit.
6.7-7.0 s: the board slides off the lens and lands on his head in the same motion, sitting level, the tassel hanging at one side.
7.1-8.5 s: he says: "I studied computer science."
8.3-8.8 s: on "science" his left hand reaches out of the right edge of the frame, grabs the brown beret and swings it toward his head. From 8.6 s the beret fills the whole frame, brown and motion-blurred.
Shot D, 8.8-10.6 s, the paint-splattered shirt from @artist-outfit.
8.8-9.2 s: the beret slides off the lens and lands on his head, tilted a little to his left.
9.45-10.3 s: he says: "Learned design."
10.1-10.6 s: on "design" his right hand reaches out of the left edge of the frame, grabs the black cap and swings it toward his head. From 10.4 s the cap fills the whole frame, black and motion-blurred.
Shot E, 10.6-12.2 s, the black hoodie from @coder-outfit.
10.6-11.0 s: the cap slides off the lens and lands on his head brim forward, and in the same motion his hand flips the hood up over it.
11.1-11.8 s: he says: "Wrote the code."
11.7-12.2 s: on "code" his left hand reaches out of the right edge of the frame, grabs the black fedora and swings it toward his head. From 12.0 s the fedora fills the whole frame, black and motion-blurred.
Shot F, 12.2-13.7 s, the dark three-piece suit from @sales-outfit.
12.2-12.45 s: the fedora slides off the lens and lands on his head, tilted a little forward.
12.55-13.15 s: he says: "Sold it."
13.0-13.7 s: on "it" his right hand reaches out of the left edge of the frame, grabs the yellow hard hat and swings it toward his head. From 13.5 s the hard hat fills the whole frame, yellow and motion-blurred.
Shot G, 13.7-20.0 s, the hi-vis vest and the robotic left arm from @cyborg-outfit.
13.7-13.9 s: the hard hat slides off the lens and lands on his head, sitting level.
14.0-15.1 s: he says: "Co-founded a studio."
15.2-19.4 s: he says: "Now I build AI agents that do real work for real businesses." On "AI agents" he raises his robotic left hand to chest height, flexes the metal fingers once and glances at it, then looks back into the lens for "real work for real businesses" and lowers the hand.
18.9-20.0 s: on "businesses" his right hand lifts the hard hat off his head and, without stopping, throws it straight at the lens. From 19.7 s the hard hat fills the whole frame, yellow and motion-blurred, to the end. No hat is left on his head.


ACTING:
Deadpan in shot A, a small grin breaking in shot B, proud on "computer science", a little playful on "design", cool and focused on "code", a confident small grin and a short nod on "Sold it", proud on "Co-founded a studio". Relaxed and a little amused, never posed. He looks into the lens while he speaks, except the short glance at the robot hand. During each swing his eyes follow the hat. Natural blinks, small micro-expressions, never a fixed stare.


CAMERA:
Shot A: full shot, head to shoes, camera slightly below chest height and a little to his left, the chair's legs, seat edge and backrest clearly visible, the whole hat tower inside the frame with a little space above the top hat. Shots B to G: camera about 70 cm in front of him at chest height, straight on, waist up, centred, the exact same framing in all six. Every shot: one fixed position, level, no zoom, no push, no pan, a very faint organic drift of 2 to 3 pixels, never handheld.


OPTICS:
Normal lens, roughly 45-50 degrees horizontal field of view, natural perspective, no wide-angle stretch, no distortion. Sharp on his face with a natural falloff to the wall. Very fine organic grain, mild vignetting.


PHYSICS AND ANATOMY:
Each hat has real weight: it travels in a natural arc, lands with a tiny settle, and its brim or board keeps its shape. The tassel swings and settles. The gown, the open shirt and the hoodie move with his arms. Every human hand has five correct fingers and holds the hat with a real grip. The robot hand has five metal fingers that bend only at the joints. No melting, no extra fingers.


LIGHTING:
The soft frontal studio light of @studio, from the same direction in all seven shots, a soft fill, no hard rim.


INTEGRATION:
He is physically in this room. His soft shadow falls on the wall and the floor behind him and moves when he moves. The chair gives a little under his weight in shot A. Each hat brim casts a soft shadow on his forehead. The robot arm catches the same light as his skin and clothes, with soft highlights on the steel.


MATERIALS:
The gown is smooth satin with a slight sheen, the stole matte gold. The beret is soft felted wool. The hoodie is heavy matte cotton. The fedora is brushed wool felt with a grosgrain band. The hard hat is glossy plastic. The vest is fluorescent mesh with silver reflective stripes. The robot arm is matte graphite carbon panels over brushed steel.


GRADE + GRADE-LOCK:
Natural colour, identical in all seven shots. The wall stays cool blue-grey, about #6C727E, from the first frame to the last. The navy gown, the gold stole, the brown beret, the paint colours, the black hoodie, the dark suit, the yellow hard hat and the hi-vis vest keep their colour. No shift in white balance between shots.


NOBODY FREEZES: his breathing, hands and shoulders move in every second.
COUNT: exactly one person. Shot A: exactly five hats on his head. Shot B: no hat on his head. Shots C to G: exactly one hat on his head (C the graduation cap, D the beret, E the black cap under the hood, F the fedora, G the hard hat until he throws it). Never more than one hat in his hand. Both arms human in shots A to F; only in shot G is his left arm robotic.


SOUND: his voice from @Audio 1, a soft cloth rustle on each reach, a short whoosh as each hat passes the lens, a faint servo whir when the robot fingers flex, quiet studio room tone. No music.


DO NOT: he does not hold a hat up to show it. A hat does not stop in front of the lens. A hat does not cover only part of the frame. A hat does not tilt into a diamond that leaves the corners open. He does not put a hat on as a separate second move. The camera does not follow the hat. The graduation cap, the black cap and the hard hat do not come from the right; the beret and the fedora do not come from the left. The outfit does not change before a hat fills the frame. Nobody touches the hat tower in shot A. No old hat stays on his head when the new one lands. The wall does not change colour between the shots. His face does not change between the shots. No tie. The black cap is never worn backwards. No glowing lights on the robot arm.
No text, no letters, no logos, no watermarks, no extra people, no second face, no other hands entering frame, no warped hands, no morphing face, no glitch. No slow motion, no fast motion, no freeze frame, no fisheye. No illustration, no 3D or CG look, no plastic or waxy skin, no beauty filter, no AI-smooth face, no glow overlay, no bloom, no HDR, no teal-and-orange grade, no oversaturation, no crushed blacks, no style drift.

Vertical 9:16, photoreal, 20 seconds, ten shots. One trick carries the first half: a hat swipes across the lens, the outfit changes behind it while it fills the frame. Then the camera moves for the first time, pushes into his palm to black, and the take ends on four quick close-ups of the work.


FORMAT:
20.0 s, 9:16, ten shots. Shot A 0.0-5.0 s, shot B 5.0-6.05 s, shot C 6.05-6.95 s, shot D 6.95-9.1 s, shot E 9.1-12.1 s, shot F 12.1-12.95 s, shot G 12.95-14.45 s, shot H 14.45-15.95 s, shot I 15.95-18.3 s, shot J 18.3-20.0 s. The cuts from A to E happen inside the moment a hat fills the whole frame, so they are never seen. E ends in pitch black and F comes out of the black. F to G, G to H, H to I and I to J are clean hard cuts. His voice comes from @Audio 1. No music.


ONE PERSON:
There is only one person in this entire piece, @you-in-grey. His face comes only from @selfie, the same face in every shot where it is seen. Every hand in the piece is his. This rule overrides every reference.


THE TWO MOST IMPORTANT RULES IN THIS SHOT:
1. When a hat passes the lens, it fills 100% of the frame edge to edge for at least 3 frames: no wall, no clothing, no skin and no light visible in any corner. The hat passes about 2 cm from the glass, so it is many times wider than the frame, soft and motion-blurred.
2. Each hat in shots A to D is one continuous, quick move: grab, swing past the lens, and the hat is on his head. He never stops, never holds a hat still in front of the camera, never shows it to the lens.
IF ANY CORNER SHOWS THE WALL OR HIS CLOTHES WHILE A HAT IS AT THE LENS, THE SHOT IS WRONG.
IF HIS FACE CHANGES BETWEEN THE SHOTS, THE SHOT IS WRONG.
IF HIS FACE APPEARS IN SHOTS G TO J, THE SHOT IS WRONG.
IF THE CAMERA MOVES IN ANY SHOT EXCEPT E, THE SHOT IS WRONG.


REFERENCES (take only what is named, nothing else):
@selfie: the selfie, 100% as shown. Take ONLY his face and his hair. Not the clothes, not the background.
@you-in-grey: Raham, 100% as shown in the reference image. Take ONLY his grey oversized tee over the white long sleeve, his black cargo pants and his build, for shots A, E and F. Take no face and no hair from it.
@coder-outfit: 100% as shown in the reference image. Take ONLY the black hoodie with the hood up and the black cap worn forward under it, for shot B, and the black cap alone for shot G. Take no face from it.
@artist-outfit: 100% as shown in the reference image. Take ONLY the open paint-splattered white shirt over the white ribbed tank and the brown beret, for shot C, and the beret alone for shot H. Take no face from it.
@sales-outfit: 100% as shown in the reference image. Take ONLY the dark three-piece suit, the white shirt open at the collar with no tie, and the black fedora, for shot D, and the fedora alone for shot I. Take no face from it.
@hard-hat: the yellow hard hat, 100% as shown in the reference image, for shot J. Take ONLY the hard hat.
@studio: studio, 100% as shown in the reference image. Take ONLY the cool blue-grey wall and the soft frontal light, for shots A to F. No chair in this take.
@Audio 1: his voice. Sync his lips to it exactly where his mouth is seen.
From no reference take its background, its lighting, its colour, its sheet layout, its panel borders or any other person.


BODY AND SCALE:
Real adult proportions, his own build as in @selfie, never slimmed. Every hat is normal real-life size for an adult head, fitting his head. In the close-ups G to J his hands are large, broad adult hands with warm medium-brown skin, and each hat lying on the desk is real-life size next to the keyboard, tablet, phone or monitor.


RAHAM - FACE LOCK, EXACT FEATURES:
@selfie controls the frontal face 100%. Zero drift. The same man in every shot where his face is seen. If anything below contradicts what the generator wants to produce, the reference and this description win.
Every feature exactly as in @selfie: face shape, eyes, brows, nose, mouth, facial hair, hairline and skin tone, with real texture and natural pores. Never narrower, longer, slimmer, younger or beautified.


SPATIAL LAYOUT:
Shots A to E: he stands about 70 cm from the lens, the blue-grey wall about 2 m behind him. The black cap and the fedora wait just outside the left edge of the frame; the beret waits just outside the right edge. Shot F: he stands about 1 m from the lens, the studio wall behind him in warm shadow. Shots G to J: close-ups on a dark wooden desk, the camera about 40 cm from the work.


ACTION TIMING:
Shot A, 0.0-5.0 s, his grey outfit, no hat.
0.1-1.7 s: deadpan, he says: "Five hats. One head." On "Five hats" his right hand comes up to chest height and shows five open fingers to the lens; on "One head" he taps the top of his head once with one finger, then the hand drops.
1.7-4.9 s: he says: "And every room I walked into only wanted one of them." A small shrug on "one of them".
4.6-5.0 s: on "them" his right hand reaches out of the left edge of the frame, grabs the black cap and swings it toward his head. From 4.85 s the cap fills the whole frame, black and motion-blurred.
Shot B, 5.0-6.05 s, the black hoodie, hood up over the black cap worn forward, from @coder-outfit.
5.0-5.15 s: the cap slides off the lens and it is already on his head under the hood.
5.2-5.9 s: he says: "So I swapped."
5.8-6.05 s: his left hand reaches out of the right edge of the frame, grabs the brown beret and swings it toward his head. From 5.95 s the beret fills the whole frame.
Shot C, 6.05-6.95 s, the paint-splattered shirt and the brown beret from @artist-outfit.
6.05-6.2 s: the beret slides off the lens and it is already on his head.
6.3-6.85 s: he says: "All day."
6.75-6.95 s: his right hand reaches out of the left edge of the frame, grabs the black fedora and swings it toward his head. From 6.85 s the fedora fills the whole frame.
Shot D, 6.95-9.1 s, the dark suit and the black fedora from @sales-outfit.
6.95-7.05 s: the fedora slides off the lens and it is already on his head.
7.0-9.0 s: a little tired now, he says: "A different me in every meeting." A small, weary shrug on "meeting".
8.85-9.1 s: his left hand lifts the fedora off his head and sweeps it across the lens. From 8.95 s the fedora fills the whole frame.
Shot E, 9.1-12.1 s, his grey outfit, no hat.
9.1-9.2 s: the fedora slides off the lens; he is in his grey tee, no hat, calm.
9.2-11.8 s: quietly and sincerely, straight into the lens, he says: "Until I found the one job that needs all five." On "all five" he raises his open right hand, palm toward the lens.
11.8-12.1 s: his palm fills the whole frame and the frame goes pitch black.
Shot F, 12.1-12.95 s, his grey outfit, warm gold light.
Out of the black, chest-up, he holds a plain black-and-white clapperboard beside his face. At 12.25 s he snaps it shut, and with the clap he says: "Films." A small confident grin.
Shot G, 12.95-14.45 s, close-up, his face not in frame: his two hands typing on a dark keyboard, a monitor behind with soft out-of-focus lines of code glowing blue. The black cap lies on the desk beside the keyboard. His voice says: "Code to build the tools."
Shot H, 14.45-15.95 s, close-up, his face not in frame: his right hand sketching a bold shape with a stylus on a drawing tablet, colour swatches on the screen. The brown beret lies on the desk beside the tablet. His voice says: "Design to make it beautiful."
Shot I, 15.95-18.3 s, close-up, his face not in frame: his hand holds a phone upright and his thumb scrolls a feed of short vertical videos. The black fedora lies on the desk next to the phone. His voice says: "Marketing so people actually watch."
Shot J, 18.3-20.0 s, close-up, his face not in frame: a monitor shows a video of the same man in a hat appearing frame by frame, as if being generated. The yellow hard hat sits on the desk beside the monitor. His voice says: "AI to make it possible."


ACTING:
Deadpan in shot A, a little faster and busier through B and C, tired by D, calm and sincere in E, a small confident grin in F. He looks into the lens while he speaks in A to F. During each swing his eyes follow the hat. Natural blinks, small micro-expressions, never a fixed stare. In G to J the hands move with purpose, at a natural working pace.


CAMERA:
Shots A to D: camera about 70 cm in front of him at chest height, straight on, waist up, centred, one fixed position, the exact same framing in all four. Shot E: the only moving shot. The camera starts in the same waist-up framing and pushes in slowly and steadily to a close-up of his face, then keeps moving into his raised palm until the palm fills the frame. Shot F: fixed, chest-up, straight on. Shots G to J: fixed close-ups from about 45 degrees above the desk. No handheld shake anywhere, no zoom, no pan; a very faint organic drift of 2 to 3 pixels in the fixed shots.


OPTICS:
Shots A to F: normal lens, roughly 45-50 degrees horizontal field of view, natural perspective, no distortion, sharp on his face with a natural falloff to the wall. Shots G to J: a macro-style close lens, shallow depth of field, the hands and the hat sharp, the background soft. Very fine organic grain, mild vignetting.


PHYSICS AND ANATOMY:
Each hat has real weight and travels in a natural arc; its brim keeps its shape. Clothes move with his arms. Every hand has five correct fingers, holds the hat with a real grip, and types, sketches and scrolls with real finger contact. The clapperboard's striped arm snaps down onto the slate with one clean clap. No melting, no extra fingers.


LIGHTING:
Shots A to E: the soft frontal studio light of @studio, from the same direction in all five. Shot F: one warm golden key light from his left, like a film-set lamp, the wall falling into warm shadow. Shot G: the cool blue glow of the monitor. Shot H: soft daylight on the desk. Shot I: a warm desk lamp. Shot J: the glow of the monitor.


INTEGRATION:
He is physically in the room. His soft shadow falls on the wall behind him and moves with him. Each hat brim casts a soft shadow on his forehead. In G to J his hands rest on the desk with real contact and soft shadows, and each hat sits flat on the desk with a soft contact shadow.


MATERIALS:
The hoodie is heavy matte cotton. The beret is soft felted wool. The paint shirt is worn cotton with dried paint. The suit is fine wool. The fedora is brushed wool felt with a grosgrain band. The hard hat is glossy plastic. The clapperboard is matte black with white diagonal stripes on the arm and a blank slate. The desk is dark oiled wood.


GRADE + GRADE-LOCK:
Natural colour. In shots A to E the wall stays cool blue-grey, about #6C727E, identical in all five, with no shift in white balance. Shot F is the only warm shot, golden. The black hoodie, the brown beret, the paint colours, the dark suit, the black fedora and the yellow hard hat keep their true colour everywhere.


NOBODY FREEZES: in A to F his breathing, hands and shoulders move in every second; in G to J the hands are always moving.
COUNT: exactly one person. Shot A: no hat on his head. Shot B: exactly one hat, the black cap under the hood. Shot C: exactly one hat, the beret. Shot D: exactly one hat, the fedora, until he sweeps it off. Shots E and F: no hat. Shots G to J: exactly one hat in each, lying on the desk, never on a head. Never more than one hat in his hand. Both arms human, no robot arm in this take.


SOUND: his voice from @Audio 1, a short whoosh as each hat passes the lens, a soft cloth rustle on each reach, a low rising hum during the push-in, one sharp clapperboard clap on "Films", soft keyboard clicks, a stylus scratch, a phone tap, a faint computer hum, quiet room tone. No music.


DO NOT: he does not hold a hat up to show it. A hat does not stop in front of the lens. A hat does not cover only part of the frame. He does not put a hat on as a separate second move. The camera does not move in shots A to D, F, G, H, I or J. The outfit does not change before a hat fills the frame. No old hat stays on his head when the new one lands. The wall does not change colour in shots A to E. His face does not change between the shots. No tie. The black cap is never worn backwards. No five-hat tower anywhere. No robot arm. No writing on the clapperboard.
No text, no letters, no readable writing on any screen, no logos, no watermarks, no extra people, no second face, no other hands entering frame, no warped hands, no morphing face, no glitch. No slow motion, no fast motion, no freeze frame, no fisheye. No illustration, no 3D or CG look, no plastic or waxy skin, no beauty filter, no AI-smooth face, no glow overlay, no bloom, no HDR, no teal-and-orange grade, no oversaturation, no crushed blacks, no style drift.

Vertical 9:16, photoreal, 20 seconds, two shots. A director on his own film set talks straight to us, then throws his cap at the lens.


FORMAT:
20.0 s, 9:16, two shots. Shot A 0.0-9.0 s, shot B 9.0-20.0 s. A to B is a clean hard cut to a closer camera angle. The take ends while the black cap fills the whole frame. His voice comes from @Audio 1. No music.


ONE PERSON IN FOCUS:
The only person in focus is @selfie, Raham, the same face in both shots. The actor @actor and the crew stay far behind him, small and out of focus, and never share his face. Every hand near the lens is his. This rule overrides every reference.


THE TWO MOST IMPORTANT RULES IN THIS SHOT:
1. His face is the face of the selfie, sharp and steady in both shots, never the actor's face, never a crew member's face.
2. His voice is @Audio 1 exactly as recorded: every word, in order, at its own time. Never drop, add, repeat, stretch or re-time a word, and leave no pause the audio does not have. The last line is three words: "Grab a hat." On "hat" the cap leaves his hand and flies straight at the lens; it fills 100% of the frame edge to edge from 16.6 s to the end.
IF HIS FACE CHANGES, NARROWS OR LOOKS LIKE THE ACTOR, THE SHOT IS WRONG.
IF THE ACTOR OR THE CREW COME INTO SHARP FOCUS, THE SHOT IS WRONG.
IF HE SAYS "GRAB HAT" WITHOUT "A", OR PAUSES BETWEEN THE WORDS, THE SHOT IS WRONG.
IF THE CAP IS THROWN LATER THAN "HAT", THE SHOT IS WRONG.
IF THE FIVE-HAT TOWER IS NOT BESIDE HIS CHAIR IN SHOT A, THE SHOT IS WRONG.
IF THE CAP COVERS ONLY PART OF THE FRAME AT THE END, THE SHOT IS WRONG.


REFERENCES (take only what is named, nothing else):
@selfie: the selfie, 100% as shown. Take ONLY his face and his hair. Not the clothes, not the background.
@director-outfit: 100% as shown in the reference image. Take ONLY the black knit sweater, the black cargo pants, the cream sneakers and the black cap worn forward with DIRECTOR in white on the front. Take no face from it.
@film-set: the film set, 100% as shown in the reference image: the warm golden light, the drapes, the chandelier, the stage lights in haze, the softbox, the cinema camera on its tripod, the monitor, the desk under the lights in the back. Take no face from it.
@chair: the black metal chair he sits on, 100% as shown.
@hat-tower: the five stacked hats, 100% as shown, standing on a tall black stand right beside his chair, clearly visible in shot A. Nobody touches it.
@actor: the actor being filmed, a man with a brown bob haircut with a fringe, a moustache and a white and yellow track jacket, sitting at the desk under the stage lights far in the back, reading a script. Take his look only for that small, soft figure.
@Audio 1: his voice. Sync his lips to it exactly.
From no reference take its white background, its sheet layout or its panel borders.


BODY AND SCALE:
Real adult proportions, his own build as in @selfie, never slimmed. The cap is normal real-life size, fitting his head. The hat tower on its stand reaches about his shoulder height when he sits. The actor and the crew are small in the frame, far behind him.


RAHAM - FACE LOCK, EXACT FEATURES:
@selfie controls the frontal face 100%. Zero drift. If anything below contradicts what the generator wants to produce, the reference and this description win.
Every feature exactly as in @selfie: face shape, eyes, brows, nose, mouth, facial hair, hairline and skin tone, with real texture and natural pores. Never narrower, longer, slimmer, younger or beautified.


SPATIAL LAYOUT:
Foreground: Raham on the black chair, the chair turned at an angle to the camera, the hat tower on its stand just beside him. Middle: the cinema camera on its tripod and the monitor, pointing toward the back. Back: the desk under the stage lights with the actor sitting at it, two crew members moving quietly, one holding a boom mic.


ACTION TIMING:
Shot A, 0.0-9.0 s, full shot.
His sitting pose for the whole shot, like an editorial portrait: he sits back in the chair, his body turned three-quarter to the camera, his legs crossed with his right ankle resting on his left knee, his right forearm lying along the raised leg with the right hand hanging loose over the knee, his left hand resting loosely on his left thigh, shoulders relaxed, chin level, head turned to look into the lens. Calm, confident, at ease. His legs stay crossed the whole shot. The five-hat tower stands on its tall black stand right beside his chair, clearly in frame.
0.0-2.0 s: no words. He holds the pose, the DIRECTOR cap on, looking into the lens; behind him the crew work quietly and the actor reads at the desk.
2.0-4.35 s: staying in the pose, he says: "Every film I post comes with the full recipe." His left hand lifts a little off the thigh in a small open-palm gesture, then settles.
4.7-7.3 s: he says: "Prompts, workflows, automations, mistakes." His left hand comes up to chest height and counts them on its fingers, one finger per word, while his right hand stays loose over the knee.
7.3-8.75 s: he points at the lens with his left index finger and says: "So you can make yours." Then the hand settles back on the thigh.
Shot B, 9.0-20.0 s, closer, chest-up, the same set soft behind him.
Same relaxed seated pose, legs still crossed below the frame, body three-quarter, head to the lens.
9.0-9.75 s: he lifts the DIRECTOR cap off his head with his right hand; his own hair, exactly as in @selfie.
9.8-10.65 s: holding the cap out toward the lens in his right hand, he says: "Want to try one on?" His left hand rests loosely on his thigh.
10.85-14.55 s: the cap held loosely in his right hand low in the frame, with a small grin, he says: "Comment HATS. I'll send you how I made this, and what it cost me." On "how I made this" his left hand opens toward the lens in a small, easy gesture.
14.9-15.6 s: straight on, with no pause before it, he says: "I'm Raham."
15.85-16.4 s: straight after, he says all three words: "Grab a hat." On "Grab" his right hand brings the cap to chest height; on "hat" he flicks it straight at the lens in one quick, short throw from the chest, no wind-up above his head, no pause.
16.6-20.0 s: the cap fills the whole frame, black and motion-blurred, then stays pressed against the lens, dark, to the end.


ACTING:
Warm, confident and relaxed, like a director who loves his work, never posed. A small grin on "Comment HATS". He looks into the lens while he speaks. Natural blinks, small micro-expressions, never a fixed stare. The actor in the back reads a script at the desk; the crew work quietly.


CAMERA:
Shot A: full shot, head to shoes, camera slightly below chest height and a little to his left, fixed. Shot B: chest-up, straight on, fixed, about 80 cm from him. No zoom, no pan, no handheld shake; a very faint organic drift of 2 to 3 pixels.


OPTICS:
Normal lens, roughly 45-50 degrees horizontal field of view, natural perspective, no distortion. Sharp on his face; the set falls off softly behind him, the actor and the crew clearly out of focus. Very fine organic grain, mild vignetting.


PHYSICS AND ANATOMY:
The cap has real weight and travels in a straight, fast line to the lens. His hands have five correct fingers each; the counting gesture is clear. No melting, no extra fingers.


LIGHTING:
The warm golden light of @film-set: a warm key from his left, the stage lights glowing in haze behind, a soft warm rim on his shoulders.


INTEGRATION:
He is physically on the set. His chair stands on the floor with a soft shadow, the cables run past it, the haze sits in the light beams behind him, and his shadow moves when he leans.


MATERIALS:
The sweater is soft black knit. The cap is black cotton twill with the white embroidered word DIRECTOR. The chair is black metal. The drapes are heavy velvet.


GRADE + GRADE-LOCK:
Warm golden film-set grade, identical in both shots. The black sweater and cap stay black, the skin stays its natural tone, the hat colours stay true.


NOBODY FREEZES: his breathing, hands and shoulders move in every second; the crew in the back move quietly.
COUNT: exactly one person in focus. Exactly one actor at the desk in the back, exactly two crew members. Exactly one cap on his head until he takes it off; then exactly one cap in his hand. The hat tower keeps exactly five hats.


SOUND: his voice from @Audio 1, the soft hum of a film set, a light rustle as he lifts the cap, a quick whoosh as the cap flies at the lens, quiet room tone. No music.


DO NOT: he does not stand up. He does not uncross his legs in shot A. He does not put the cap back on. Nobody touches the hat tower. The actor does not walk forward. No one else comes near the camera. The camera does not follow the cap. The cap does not cover only part of the frame at the end. The word on the cap stays DIRECTOR and stays readable on the cap only.
No other text, no letters, no logos, no watermarks, no second sharp face, no other hands near the lens, no warped hands, no morphing face, no glitch. No slow motion, no fast motion, no freeze frame, no fisheye. No illustration, no 3D or CG look, no plastic or waxy skin, no beauty filter, no AI-smooth face, no glow overlay, no bloom, no HDR, no teal-and-orange grade, no oversaturation, no crushed blacks, no style drift.

Vertical 9:16, photoreal, 4 seconds, one shot. A director on his own film set says one word, "Films.", and snaps a clapperboard on it, cool and easy.


FORMAT:
4.0 s, 9:16, one continuous shot. His voice comes from @Audio 1: the single word "Films.", nothing else. No music.


ONE PERSON IN FOCUS:
The only person in focus is @selfie, Raham, the selfie. Any crew in the back stays small, soft and out of focus. This rule overrides every reference.


THE TWO MOST IMPORTANT RULES IN THIS SHOT:
1. He is relaxed and natural, never stiff, never posing: weight on one leg, loose shoulders, the easy confidence of a director on his own set.
2. He says only "Films." at about 1.5 s, and on that word the clapperboard arm snaps shut with one clean clap.
IF HE STANDS STIFF AND SQUARE TO THE CAMERA HOLDING THE CLAPPERBOARD UP LIKE A SIGN, THE SHOT IS WRONG.
IF HIS FACE CHANGES, NARROWS OR LOOKS LIKE SOMEONE ELSE, THE SHOT IS WRONG.
IF ANY WRITING APPEARS ON THE CLAPPERBOARD, THE SHOT IS WRONG.


REFERENCES (take only what is named, nothing else):
@selfie: the selfie, 100% as shown. Take ONLY his face and his hair. Not the clothes, not the background.
@director-outfit: 100% as shown in the reference image. Take ONLY the black knit sweater, the black cargo pants and the black cap worn forward with DIRECTOR in white on the front. Take no face from it.
@film-set: the film set, 100% as shown in the reference image: the warm golden light, the drapes, the chandelier, the stage lights in haze, the cinema camera on its tripod, the monitor. Take no face from it.
@Audio 1: his voice. Sync his lips to it exactly.


RAHAM - FACE LOCK, EXACT FEATURES:
@selfie controls the frontal face 100%. Zero drift. Every feature exactly as in @selfie: face shape, eyes, brows, nose, mouth, facial hair, hairline and skin tone, with real texture and natural pores. Never narrower, longer, slimmer, younger or beautified. His own build, never slimmed.


SPATIAL LAYOUT:
He stands beside the cinema camera on its tripod, the monitor near him; behind him the chandelier and the stage lights glowing in haze, the drapes at the back, crew far behind and out of focus.


ACTION TIMING:
Medium-wide shot from the thighs up, standing, the whole shot.
0.0-1.0 s: he stands loose beside the cinema camera, his weight on one leg, his left forearm resting casually on top of the camera, the clapperboard hanging easily in his right hand at his side, looking into the lens.
1.0-1.5 s: he lifts the clapperboard to chest height in one easy move, the arm open.
1.5-2.1 s: he says "Films." and snaps the clapperboard shut toward the lens, one clean clap, a half-smile.
2.1-4.0 s: he lowers it again and leans back on the camera, relaxed, a small nod, still looking into the lens.


ACTING:
Cool, warm and easy, like a director who has done this a thousand times. Small, real movements; never a pose. Natural blinks, never a fixed stare.


CAMERA:
Fixed camera at chest height, no zoom, no pan, no handheld shake; a very faint organic drift of 2 to 3 pixels.


OPTICS:
Normal lens, roughly 45-50 degrees horizontal field of view, natural perspective. Sharp on his face; the set falls off softly behind him. Very fine organic grain, mild vignetting.


PHYSICS AND ANATOMY:
The clapperboard arm snaps down onto the slate with one clean clap. His hands have five correct fingers each. His forearm really rests on the camera. No melting, no extra fingers.


LIGHTING:
The warm golden light of @film-set: a warm key from his left, the stage lights glowing in haze behind, a soft warm rim on his shoulders.


MATERIALS:
The sweater is soft black knit. The cap is black cotton twill with the white embroidered word DIRECTOR. The clapperboard is matte black with white diagonal stripes on the arm and a completely blank slate.


GRADE + GRADE-LOCK:
Warm golden film-set grade. Black stays black, skin stays natural warm brown.


NOBODY FREEZES: his breathing, hands and shoulders move in every second.
COUNT: exactly one person in focus, exactly one clapperboard, exactly one cap on his head.


SOUND: his voice from @Audio 1, one sharp clapperboard clap on "Films.", the soft hum of a film set, quiet room tone. No music.


DO NOT: he does not stand stiff and square to the camera. He does not hold the clapperboard up in front of his chest like a sign. He does not say any other word. No writing, letters or labels on the clapperboard. He does not take the cap off.
No other text, no logos, no watermarks, no second sharp face, no warped hands, no morphing face, no glitch. No slow motion, no fast motion, no freeze frame, no fisheye. No illustration, no 3D or CG look, no plastic or waxy skin, no beauty filter, no AI-smooth face, no glow overlay, no bloom, no HDR, no oversaturation, no crushed blacks.