Vertical 9:16 clip, 5 seconds, one continuous unbroken take, no cuts, shot from @cc's own iPhone 15 Pro front-facing selfie camera, phone held up and angled by @cc's own hand throughout — handheld micro-shake, imperfect UGC framing, mild autofocus hunting that recovers quickly, subtle wide-angle edge distortion, no beauty filter, no gimbal smoothness, no slow motion, no zoom, 24fps real-time speed, no polished commercial cinematography. @surrounding ref this is the plaza, building and surrounding references lock REFERENCE LOCK: @cc is the sole reference, matching the attached photo exactly — face shape, fair skin, blue-grey eyes, long wavy ginger-red hair, sunglasses pushed up on the head, and pearl drop earrings never change; wardrobe locked to the sleeveless dark-red ribbed polo top with a button placket, thin brown leather belt, beige wide-leg trousers with cream drawstring ties, a dark-brown crossbody bag with a small red heart charm, and dark-brown leather loafers — no redesign or restyling. Only @cc appears on camera. Busy plaza in front of a large modern museum-style building with an angular stone-and-glass facade, overcast daylight, soft even light, pedestrians and city traffic moving naturally in the background. [0–1.2s] Phone held up at arm's length in a normal selfie framing, @cc smiles warmly into the lens, a takeaway coffee cup in her other hand down at her side. Lip-synced: "Hey guys." [1.2–2.8s] She tilts the phone downward and back in one smooth sweeping motion, tracking down over her top, belt, and the full front of her beige trousers to at least mid-thigh, clean and unstained, before tilting back up to her face. Lip-synced: "Quick fit check." [2.8–4s] She turns her head to glance back toward the building so its angular facade sweeps into frame behind her, then turns back to the lens with a pleased, showing-off expression. Lip-synced: "Look where I am at." [4–5s] She lifts the coffee cup up into frame beside her face with a small proud tilt toward the lens, settling into a relaxed closing smile. Lip-synced: "Got my morning coffee." Framing lock: the phone-camera angle stays a selfie POV throughout, moved only by @cc's own arm — the downward tilt in the fit-check beat is the one moment her trousers fill more of the frame, shown plainly and briefly rather than lingered on, before the framing returns to her face and shoulders for the rest of the clip. Audio is diegetic: plaza ambience, distant traffic, overlapping pedestrian chatter, light wind, and her own voice close and clear from the phone mic. No music, no narrator, no on-screen text, captions, logos, or UI. The shot never freezes and never fades — it ends on her smile holding the coffee cup up.

Vertical 9:16 Reel, 20s total, two hard cuts, shot on an iPhone 15 Pro front-facing selfie camera held at arm's length throughout — handheld micro-shake, imperfect UGC framing, natural walking bounce, mild autofocus hunting that recovers quickly, subtle wide-angle edge distortion, no beauty filter, no gimbal smoothness, no slow motion, no zoom, 24fps real-time speed, no polished commercial cinematography. Screen direction and framing stay consistent throughout so it reads as one continuous filming session. REFERENCE LOCK: @cc is the sole reference, followed exactly throughout — face shape, fair skin, blue-grey eyes, long wavy ginger-red hair, sunglasses pushed up on the head, and pearl drop earrings never change; wardrobe locked to the sleeveless dark-red ribbed polo top with a button placket, thin brown leather belt, beige wide-leg trousers with cream drawstring ties and a raw frayed side seam, a dark-brown crossbody bag with a small red heart charm worn across the body, and dark-brown leather loafers — no redesign or restyling. SHOT 1 [0–10s], one continuous unbroken take, busy city street, overcast midday daylight, soft even sun as the only light source. [0–2.5s] @cc walks at a relaxed pace, phone in right hand, an open lidless takeaway coffee cup in left hand; @cc's gaze catches a striking piece of architecture ahead, eyes brightening. [2.5–4s] @cc turns and walks toward it, lifting the phone, deciding to film it, smiling: "Oh, this is such a good shot." [4–5.5s] Closing the last steps, a passerby's shoulder clips @cc's from frame-left; @cc's hand jolts, the cup tilts, a modest gravity-true splash of coffee hits the front thigh of the beige trousers, leaving one small believable stain. @cc's smile drops, breath catching: "Oh…" [5.5–7.5s] @cc stops directly in front of the architecture, facing it, the building filling the background at a fixed medium distance. A beat of stillness, then @cc's shoulders loosen into a confident smile as @cc looks back to the lens: "It's fine. I've got Creator City." [7.5–10s] @cc raises the phone into proper filming position, framing @cc with the architecture behind, gives one small presenting gesture toward the building, and settles into a held filming pose, stain still visible on the trousers. Position and camera lock here — no further walking. HARD CUT, occlusion-style, pushing tight on the stained fabric. SHOT 2 [10–14s], tight UGC crop on the coffee-stained beige trousers, matching Shot 1's lighting exactly, no laptop, no screen, no editing software or interface ever visible: a thin white lasso line draws itself in one continuous loop around the stain. The instant it closes, the stain lifts in under a second, the fabric resolving back to clean beige with no wipe or dissolve. Lip-synced: "With Creator City, little accidents like this don't mean you have to reshoot." Prop/continuity lock: the stain's size, shape, and position stay identical from the end of Shot 1 into Shot 2, appears once, is removed once, and the clean state holds fixed into Shot 3. HARD CUT, match cut back to the exact framing of the end of Shot 1. SHOT 3 [14–20s], the identical locked composition from the end of Shot 1 — same building, same distance and angle, same held filming pose — only the trousers are now fully clean, no trace of the stain. Hair, top, belt, bag, shoes, stance, and phone position all match exactly. @cc stays silent on-screen, continuing to film in front of the building — small natural posing shifts, a slight tilt of the phone, weight shifting lightly between her feet, an easy relaxed expression — with no mouth movement or dialogue. A voice-over plays over the shot, not lip-synced to on-screen movement, in the same voice heard earlier: "Edit like a pro with Creator City." Framing lock: the end of Shot 1 and Shot 3 share identical camera position, distance, angle, and background, so the only variable across the cut is the stain — a direct before/after callback. Audio is diegetic for Shots 1 and 2, no music, no narrator: Shot 1 carries street ambience, traffic, chatter, footsteps, cup-handling sounds, a soft foley at the bump, and a small spill sound, settling quieter once @cc stops; Shot 2 carries a subtle lasso-drawing sound and a soft transformation sound as the stain vanishes, with faint street ambience underneath. Shot 3 carries the same settled street ambience under a clear, natural voice-over line, spoken off-camera-style with no lip movement on screen. No on-screen text, captions, logos, or UI at any point. The video never freezes or fades — it ends on @cc's relaxed pose in Shot 3.

SCENE CONTEXT 15s UGC reveal: a female influencer asks if you can make UGC videos with one click, cuts to a fast montage of 5 different creators each showcasing one product, returns to react, then closes laughing. Every scene shot handheld selfie-style on an iPhone front camera — phone itself never visible on screen. No studio lighting, no beauty filter. FORMAT MODE Timed multishot, HARD CUTs at stated seconds, camera does not cut on its own. FIRST FRAME Western female influencer, late-20s, fitted knit top, gold layered necklace, hair loose and glossy, stands at a marble kitchen counter, iPhone held selfie-style, only hand/wrist edge visible, phone body out of frame. Soft window light camera-left. Behind her: open shelving with curated ceramics, brass fixtures, woven pendant light — elevated but lived-in. OPTICS/CAMERA All shots: iPhone front camera, ~80° wide FOV, arm's-length, held slightly below eye level looking up. Phone body/screen never enters frame. Constant handheld tremor, 1-2cm, natural micro-reframing, no gimbal smoothness. No drift mid-segment. ACTION 0-2s HOOK — She leans toward lens, eyebrows raised: "Wait, you can literally make U.G.C videos with one click now?" Skeptical half-smile: "Okay, watch this—" reframes toward next beat, motion-blur whip, settles by 2s. HARD CUT. 2-3.4s SHOT 1 BEAUTY — Different Western woman, brunette, hair tied back, sweatshirt, bright bathroom with brass fixtures, arched mirror, selfie POV. Holds a glass serum dropper bottle at chin height, label facing lens: "Okay, I've been trying this serum for a week…" HARD CUT. 3.4-4.8s SHOT 2 FASHION — Different Western woman, blonde, sunlit bedroom, gallery wall, velvet chair, selfie POV. Wears a cropped denim jacket, tugs the collar: "Wait… I actually love this jacket." HARD CUT. 4.8-6.2s SHOT 3 FOOD — Different Western man, casual fit, café table, marble top, rattan chairs, low selfie POV, a matcha croissant on a small plate in frame. Takes a bite, pauses, eyebrows lift: "Oh wow… this croissant is actually really good." HARD CUT. 6.2-7.6s SHOT 4 TECH — Different Western man, sofa near window, walnut coffee table, selfie POV. Holds wireless earbuds in their charging case toward lens: "Okay, these earbuds are way more useful than I thought." HARD CUT. 7.6-9s SHOT 5 GRWM — Different Western woman, morning light through sheer curtains, vanity, round mirror, selfie POV. Holds a tinted-moisturizer stick, dabs her cheek: "Come get ready with me — this is my new base." HARD CUT. 9-12s REACTION — Back to original kitchen influencer, same counter/light/wardrobe/decor, selfie POV, coffee mug in hand. Half-laugh, half-shock: "I literally made all of these with one click." Beat, smile fades to mock-confusion, glances off-frame: "What?" HARD CUT. 12-14s FINAL MONTAGE (3 fast cuts, ~0.6s each, selfie POV, phone never visible) — A: hands tearing open a skincare package on the styled marble counter, surprised half-smile. B: man mid-laugh, low selfie angle, product box in frame corner, soft sunlight flare. C: creator walking outdoors past a tree-lined street, product-branded tote on her shoulder, hair in breeze. HARD CUT. 14-15s CLOSE — Back to original kitchen influencer, same framing/wardrobe/decor, selfie POV. Smiles, shakes head, sips mug: "Yeah… I'm not filming ads anymore." Short real breathy laugh, eyes crinkling. WARDROBE Original influencer: fitted knit top, gold layered necklace, subtle earrings throughout — polished but casual. PERFORMANCE Matte, unfiltered skin — visible pores, natural T-zone sheen, no smoothing. Genuine micro-expressions: real blink timing, asymmetric smiles, small hesitations before lines. Restrained, conversational delivery. LIGHTING Natural daylight only — large window as sole key light, soft diffusion, realistic shadow falloff. True-to-life exposure. WB ~5600K consistent across all cuts. No rim light, no studio softbox character. PHYSICS Hand/arm tremor reacts naturally to movement and laughter. Hair/fabric respond to real air movement outdoors. Natural grip and finger tension on held products. STYLE Photoreal, unfiltered iPhone-camera look — compressed dynamic range, mild shadow noise, no beauty-filter smoothing, no added color grade, no text/logos/UI no brands POSITIVE LOCKS Every scene is selfie-POV on an iPhone front camera — phone body/screen never appears in any shot. Each montage character is distinct in a distinct location, each holding/using one clearly visible product. Original influencer and kitchen appear identically only in 0-2s, 9-12s, and 14-15s. Daylight-only lighting and matte skin hold throughout.

SCENE: Selfie-style influencer video, quiet NYC brownstone street (@image1: cobblestone, red-brick, fire escapes, trees, café tables, distant Art Deco skyline, light unbranded traffic). One continuous 15s handheld phone shot, no cuts, constant imperfect framing. REFS: @image2=Western blonde woman, mid-20s, fair skin, center-parted low bun w/ face-framing strands, gold hoops/bangles, oversized mint-teal knit sweater over white collared shirt, tan pleated mini skirt, white crew socks, brown penny loafers, brown leather bag w/ green tassel. Match exactly; natural American English. CAMERA: Selfie phone, 84° FOV, constant gait bounce/drift/tilt, autofocus hunting, never static. Phone locked in her right hand at arm's length for all 15s — never released, never swapped hands; that hand's motion drives all camera movement, including the final whip/blur. ACTION: 0–2s Hook, walking: "Wait — if you're not doing this yet, you're already behind, I need you to see this." Left hand pulls lipstick from bag to lens; right hand stays fixed on phone. Balloon vendor loses balloons behind her, zero reaction. 2–4s Lipstick held in left hand near lens: "What if you want to change the product after you've filmed it?" 4s "You just change it" — lipstick in left hand instantly becomes sunglasses, same hand/grip point, no hand swap, no morph/cut. Brief wrist twitch only; right hand still on phone. 4–6s Puts on sunglasses with left hand, walking, right hand still gripping phone: "And it's not just the product." 6–8.5s "You can change the lighting of the scene — without reshooting" — lighting instantly flips day→full NYC night, everything else (pose, outfit, phone grip) unchanged. 8.5–12s CTA, confident, night: "You can do all this in Creator City. Sign up for free trial now. Limited time only." Steps off curb, crosses street. 12–13.5s Off-frame taxi horn — she flinches, gasps "Whoa—!", stride breaks, phone whips hard in her right hand but is never dropped. 13.5–15s Camera judders into fast motion blur, no recovery, ends in blur — phone still gripped in her hand throughout. Taxi shown stopped, sound only. PERFORMANCE/PHYSICS: Natural UGC. Reflexive lipstick reach and swap-flinch confined entirely to the free (left) hand; phone hand stays steady, never changes grip, even through the final blur. Deadpan on balloon loss. Gait bounce; bun secure, strands react subtly. Fabric/bag/hair move naturally. Day→night changes only light/sky. LIGHTING/AUDIO: 0–6s natural flat daylight; at 6s instant full NYC night (streetlamps, window glow, cooler tone), rest unchanged. Diegetic only, no music: her dialogue, ambient NYC, faint balloon whoosh, soft sound at each swap, sharp taxi horn at 12s, gasp, motion whoosh to blur. VOICE: NYC accent, female, early-mid 30s. Warm, relaxed, confident, casual conversational delivery, natural breathing/pauses/pitch, slight vocal fry. Sounds like she just pulled out her phone and started talking naturally, not like she is reading a script STYLE/OUTPUT: Authentic handheld UGC footage, realistic exposure, natural skin texture, imperfect autofocus, no stabilization/polish. Vertical 9:16, real-time single take, 15s. No text/captions/watermark. LOCKS: One unbroken 15s shot, no cuts. Match @image2 exactly throughout. Street matches @image1. Balloons multi-colored, unprinted. No brands/logos/signage/music/on-screen text. Camera always handheld/moving, framing off-center. Phone stays in the same (right) hand the entire video — never set down, released, or switched, including during final blur. Lipstick→sunglasses swap occurs within the same free hand that pulled the lipstick — never jumps to phone hand. Balloon-loss and proposal-rejection clear within first 2s, zero reaction. Lipstick pull-out simultaneous with hook. Both swaps instant, zero transition/morph, only brief physical reaction. No shoe or food swap. Video ends honk→shock→jolt→blur, no recovery. Taxi shown stopped—sound only.

Single continuous take, no cuts, 28s, vertical 9:16, natural UGC vlog register, shot on an iPhone front selfie camera in her own hand the whole time, never stabilized, reads as ordinary self-shot footage. Constant low handheld drift from arm/breath, frame swinging with her shoulder when she turns, easing back after. Ultra-wide ~84° FOV, natural selfie-lens warp, focus locked on her face. @image1 sets the room, 100% matches reference: warm minimalist japandi living room — cream boucle curved sofa, round dark-wood coffee table with books and a small plant, grey-cream striped rug, herringbone wood floor, dark-wood dining table with black leather chairs behind, walnut sideboard under a woven tapestry artwork, twin dome pendant lights, a large monstera in the corner, warm sunlight through sheer curtains. Light later swaps instantly to soft daylight, then a mellow warm-accent shift, then locks permanently back to the same sunlit glow as @image1. A Gen Z woman, early-20s, relaxed and genuinely pleasant, not hyped-up, casual thrifted-retro outfit (graphic tee, wide-leg jeans, gold hoops), sits on the sofa, phone in her own hand at arm's length, filming herself throughout. Energy stays comfortable and mildly delighted — small smiles, a light laugh at most, calm delivery, never shouting or over-gesturing. [0–2.5s] She leans in slightly, easy smile, telling the lens calmly what she's about to show. [2.5–6s] Holding a soda can, mid-sentence it instantly becomes a water bottle in the same grip, condensation appearing, camera motion unbroken; small amused exhale, "wait, not this one," instantly reverts to the soda can — the one deliberate swap-back; it never changes again. [6–9.5s] She turns to the wall artwork; it instantly becomes a bolder, clearly different textile piece, believable not cartoonish, stays permanently. [9.5–14s] Camera pans to the sofa; the cream boucle instantly becomes a deep charcoal linen sofa, distinctly different silhouette and texture, holds for the rest of the video. [14–16s] She gestures at the rug; it instantly becomes a bold terracotta geometric rug, permanent. [16–18.5s] "Oh, and even what I'm wearing" — her outfit instantly becomes a fitted knit top and tailored trousers in a rich solid tone, same pose, stays through the ending. [18.5–21.5s] Light instantly shifts sunlit→soft morning, then a warmer accent tone, small laugh, "yeah, that's a lot," then permanently settles back to the same sunlit glow as the reference. [21.5–25s] She settles, calmer half-joking tone: "Okay... now I'm gonna address the elephant in the room." She extends her arm a little further, tilting the phone back slightly to keep herself fully in frame while opening the space behind her. Instantly a real elephant appears at a large but plausible scale just behind and beside the sofa, her face and its full body sharing the frame, zero transition on its appearance but the room reacts physically to its weight the instant it's there: floorboards creak faintly, the rug compresses under its feet, the monstera leaves tremble and sway, the pendant lights sway slightly, sideboard objects rattle, curtains stir, coffee-table items shift from the vibration. [25–26.5s] Her reaction stays grounded eyes widen, she goes still, a short surprised exhale, shoulders tense slightly restrained, not a scream, still sharing the frame with the elephant as the room keeps subtly settling. [26.5–28s] Her grip loosens naturally, her arm dips the phone down in a real, slightly uneven motion-blurred arc, not a violent throw; it tumbles briefly and lands screen-up, settling into a static level shot of the plain, empty ceiling. Every swap is instant and zero-frame — no morph, no glow, no particle effect each a clear, believable before/after in plain material terms, and once changed the wall art, sofa, rug, outfit, and final lighting permanently hold with no reversion; only the soda-can beat intentionally swaps twice before locking. The elephant appears instantly, but everything around it shows real secondary motion from its weight, never weightless or pasted-in. Her handheld motion, natural performance, and lip-sync continue uninterrupted through every swap; the phone drop is physical arm motion within the same shot, never a cut. Audio is diegetic only, no music, normal volume: "Guys, now with Seedance 2.5, you can object-swap with one click while keeping the original camera movement... see, it's water now, and the shot didn't skip a beat. Wait, not this one... there we go, lip-sync's still perfect too. I can even switch up the wall art... nice. And my couch... totally different now. Even the rug... yeah. Oh, and even what I'm wearing... completely different. And the lighting too... yeah, that's a lot... there we go, perfect. Okay... now I'm gonna address the elephant in the room" — short surprised exhale, "oh—", faint creak and rustle, soft phone-drop impact, her breathing and a low elephant rumble continuing off-frame over the closing ceiling shot

A 15-second vertical (9:16) UGC sneaker try-on video filmed on a smartphone. Noticeably lower-resolution, non-4K quality — soft compression artifacts, slight motion blur on quick movements, mild grain in the shadow areas, and the faint over-sharpened look typical of a phone's front camera, not a professional camera. A young tanned woman with brunette hair pulled back films herself in a bright bedroom with white bedding, dark built-in closets in the background, wood flooring, and strong natural sunlight streaming through blinds casting diagonal light stripes across the floor. Handheld selfie-style camera with visible natural shake, authentic casual fashion-influencer energy, warm natural tones, talking directly to camera throughout with unscripted, casual delivery. 0–2s: She sits on the edge of the bed in a fitted beige crop t-shirt and light blue pleated athletic skirt, holding up the attached sneaker — a green, black, and grey mesh running shoe with a chunky white midsole, black knit toe and heel panels, grey diagonal overlay stripe, and black laces — in both hands close to the camera, rotating the pair slightly to show the side profile and white sole. She says: "Okay wait, you guys need to see these before I put them on." Excited, slightly breathless tone, smiling with raised eyebrows. Preserve the exact colorway, panel layout, and sole shape of the reference shoe. 2–5s: Quick cut to a low close-up angle on her bare legs and feet — she pulls on white ankle socks, slides her foot into one green-and-grey sneaker, adjusts the heel, and tightens the black laces. She says, slightly off-frame: "So comfy already, like — right out the box comfy." Casual, half-laughing delivery. Diagonal sunlight stripes cross the wood floor, catching the texture of the mesh upper and white foam sole. 5–6s: Static low ankle-height shot of both feet now fully laced into the green, black, and grey sneakers, standing on the wood floor. She says off-camera: "Look how clean that is." Short, satisfied tone. 6–11s: Cut to a full-body static shot — she stands in front of the bed, does a small playful shimmy/spin, gesturing with her hands as if presenting the shoes. She says: "Okay but the fit though? Like this is giving exactly what I wanted." Laughing mid-sentence, natural pauses, slightly overlapping words like real unscripted speech. The green and grey sneakers are clearly visible against the light blue skirt. 11–15s: She settles into a relaxed standing pose, hands falling naturally to her sides, smiles directly at camera and adds: "Yeah... these are staying in rotation." Warm, confident, trailing off casually as the clip ends naturally, sneakers in full view at the bottom of frame. Style: Authentic UGC / TikTok sneaker-haul try-on aesthetic — looks like a real iPhone front-camera selfie video, not a polished production. Visibly lower bitrate/resolution, slight digital noise, minor motion blur during quick transitions, natural handheld wobble, occasional autofocus hunting/refocus blur when she moves close to camera. Quick clean jump cuts between each stage. Natural warm bedroom daylight, no ring light, no studio lighting, no color grading. Constant casual talking with natural pauses, laughs, and filler words — not a scripted voiceover. The woman remains visually identical throughout every shot, and the sneaker's green/black/grey colorway and construction stay identical to the attached reference in every shot. No music specified. No text overlays, no logos, no filters, no beauty smoothing, no CGI appearance.

SHOT 1 (00:00-00:01) — Macro Label Texture Reveal EFFECT: Diagonal light-streak wipe + crash zoom-out Extreme macro on the blue label's blueberry illustration and foil rim, catching hot specular streaks as light rakes across the surface Camera pulls back rapidly from texture to reveal the full jar Bold bubble-text crashes in top-frame: "GLOSSY & JUICY!" (red/yellow chrome, comic outline) Exits as the zoom-out completes into Shot 2 — no cut, continuous pull SHOT 2 (00:01-00:02) — Label Detail Hold EFFECT: Static macro hold + secondary text stamp Camera holds on the blueberry-and-leaf print with soft bokeh light flares drifting across Starburst badge pops in bottom-right: "FLAVOR BURST!!" (blue-outlined yellow burst shape) Hard cut to full jar reveal SHOT 3 (00:02-00:04) — Hero Product Reveal EFFECT: Vertical bloom flash entry + settle Jar reveal against a deep blue starfield gradient, rim-lit with a hot white specular flare along the lid Camera settles into a slight low-angle hero framing, jar centered Blueberries drift in softly at frame edges (foreground/background blur) Transition: hard cut on a beat SHOT 4 (00:04-00:06) — 3D Spin-Up EFFECT: Product rotation (360° yaw) + digital zoom (slow push-in) + rim glow Jar rotates smoothly in 3D space against a bright saturated blue background Blueberries fly into frame from top corners, arcing past camera Bold stacked bubble-text crashes in bottom third: "BLUEBERRY BLAST!" (gold/red comic lettering with motion lines) This is the SIGNATURE VISUAL EFFECT — the full spin is the most memorable beat Exits on a whip-pan blur as text finishes crashing in SHOT 5 (00:06-00:08) — Comic Sticker Explosion EFFECT: Multi-sticker pop-in (staggered) + camera micro-shake Jar sits diagonally on a saturated blue backdrop while 4-5 comic-style stickers pop in one after another around it: "SO GOOD!", "INSANELY TASTY", "90 GUMMIES", "CHEW TO THE LAST BEAR!" Each sticker enters with a small scale-bounce (overshoot then settle) Blueberries scattered around the frame edges, some in soft focus Camera has subtle continuous jitter to keep energy up Transition: fast cut SHOT 6 (00:08-00:10) — Flight / Power Shot EFFECT: Falling/flying product-and-fruit composite + speed lines Jar tumbles diagonally through frame as blueberries rain and orbit around it, trailing soft motion blur Comic speed-lines radiate from center Bold diagonal text crashes in: "FLY, GUMMY POWER!" (red/gold chrome, thick outline, slight perspective skew) Small sparkle particles trail behind the jar Transition: hard cut into the splash shot SHOT 7 (00:10-00:11) — Splash / Money Shot EFFECT: Composited burst/splash (thick gel simulation) + macro fruit Blueberries burst through a thick, glossy blueberry-gel splash frozen mid-motion, droplets and sparkle particles suspended in the air Deep blue background, high-gloss highlights on every droplet Chunky blue-chrome text crashes in from bottom: "REAL BLUEBERRIES" (thick 3D lettering, heavy drop shadow) This is a second SIGNATURE VISUAL EFFECT — the tactile "juicy" money shot Transition: fade/dissolve into darkness SHOT 8 (00:11-00:12) — Slow-Motion Pour Close-Up EFFECT: Slow-motion (deep, ~20-25% speed) + macro close-up + soft rim light Camera is tight on the jar's mouth as it tilts and pours — blue gummy bears tumble out one after another in extreme slow motion, catching soft highlights on their glossy, translucent surfaces as they fall Shallow depth of field: the foreground gummy bears are crisp, the jar label blurs softly behind A few gummies tumble and bounce gently past camera, trailing tiny motion-blur arcs Small sparkle particles drift through frame to keep the polish consistent with the rest of the edit Text crashes in bottom-frame, glowing purple/magenta neon: "SOFT & CHEWY, EVERY TIME" Transition: snap cut to full brightness for Shot 9 SHOT 9 (00:12-00:15) — Hero Finish / Badge Wall EFFECT: Headline crash-in + multi-badge pop-in + closing tagline Massive gold/red chrome headline crashes in from top: "BUY NOW!" Starburst badge pops in left: "No.1" ribbon below it: "LIMITED OFFERS" Second badge pops in right: "ORDER YOUR FIRST PACK NOW!" Jar sits centered and static, fully lit, blueberries piled at its base left and right Closing tagline slides up from bottom in bold white/purple lettering: "FLAVOR, PACKED IN EVERY BITE!" Hold on final frame — energy fully resolved, no further movement negative prompt: no voice over

High-energy cinematic product commercial. Molten liquid in glossy emerald green and polished chrome-silver flows and swirls elegantly in mid-air, the two colors twisting together like liquid metal, pouring into a sleek luxury wristwatch. The green and silver liquid ripples across the watch's surface, filling every link groove, dial texture, and strap stitch line like molten paint. Dynamic camera movements: extreme macro tracking shots of the fluid strands coiling around the crown and lugs, aggressive 360-degree orbits around the watch case, and fast whip pans between the dial face and the strap. Speed-ramping transitions: buttery slow-motion as the liquids collide and blend across the dial's sunburst texture, followed by snap-fast reveals of the fully coated timepiece, hour markers glinting through the metallic sheen. Deep charcoal-black studio background with professional rim lighting, crisp highlights, and subtle lens flares for contrast against the green/silver. The video concludes with a smooth match-cut transition to a final packshot: a matte black watch box with a deep green interior sitting centered on a polished, reflective black surface, a faint green and silver liquid sheen still gliding across the floor. High-end 3D liquid simulation, 8K resolution, hyper-realistic materials, polished commercial aesthetic.

HOOK (0–2 sec) POV handheld, slightly shaky. A mauve/dusty-rose shopping bag is tossed onto a white unmade bed — soft thud, subtle tissue rustle. Warm natural daylight. JUMP CUT 1 (2–4 sec) Close-up hands grab the handles of the matching mauve structured tote and pull it into frame. Focus shifts blur → sharp. Gold hardware catches the light. JUMP CUT 2 (4–7 sec) The tote is lifted from the bed, revealing its structured silhouette, rounded top handles, clean front panel, and polished gold fittings. The large gold interlocking keychain charm swings naturally with a soft metallic clink. JUMP CUT 3 (7–11 sec) Matching mauve shoulder/crossbody bag in the foreground is picked up and moved beside the tote. Fast handheld camera circle around both bags, emphasizing the smooth leather texture, coordinated color, sculpted shape, and gold hardware. JUMP CUT 4 (11–13 sec) Extreme close-up: fingers glide across the leather surface, then trace the gold clasp and hardware. The interlocking gold charm swings and catches the light. OUTRO (13–15 sec) Both bags are lifted toward the camera for a quick full reveal — large tote behind, smaller crossbody in front. Natural handheld movement, premium fashion UGC feel, white bed and neutral room softly blurred in the background.

A 15-second vertical (9:16) UGC try-on video filmed on a smartphone. The young man from the attached reference image films himself in front of a full-length mirror in a minimalist bedroom with neutral beige walls and natural daylight from a window. Handheld selfie-style camera, authentic casual fashion-influencer energy, slightly warm natural tones. 0–3s: He faces the mirror camera wearing a simple basic white tee and relaxed shorts. He holds up the outfit pieces on hangers — a cream textured knit henley sweater and a pair of relaxed dark charcoal drawstring trousers — showing them to the camera with a subtle “watch this” expression. He raises an eyebrow slightly and gives a small confident look. Accessories, including simple sneakers, sunglasses, and a shoulder bag, are laid out casually on the bed behind him. 3–5s: Quick jump cut — he is now wearing the cream textured knit henley sweater. Preserve the soft loose knitted texture, ribbed collar and cuffs, relaxed oversized fit, and partial button placket matching the attached character's fashion aesthetic. He adjusts the front, lightly smooths the fabric, and turns side to side while checking the silhouette in the mirror. 5–8s: Another quick jump cut — he steps into the relaxed dark charcoal drawstring trousers, pulling them up and adjusting the waistband. The trousers have a loose, casual straight fit with a visible drawstring detail. He tugs the waistband lightly and makes a small turn to show the overall shape and proportions. 8–11s: Jump cut — the full outfit comes together. He slips on clean minimalist sneakers, adds a pair of slim oval black sunglasses, then picks up a simple understated shoulder bag and lets it rest naturally over one shoulder. He steps back to show the complete look: soft cream knitwear, dark relaxed trousers, and minimal accessories. He performs a relaxed slow turn, one hand casually holding the bag strap. 11–15s: He faces the mirror straight on and shifts his weight casually into a relaxed confident pose. Chin slightly raised, effortless calm expression. He briefly holds the pose, gives a subtle satisfied smile, then reaches toward the phone camera and the video cuts naturally. Style: Authentic UGC / TikTok men's fashion try-on haul aesthetic. Quick clean jump cuts between each stage of getting dressed. Handheld smartphone framing with slight natural camera shake. Natural bedroom lighting, no ring light, no studio setup. The attached character remains visually identical throughout every shot. The outfit is the star: soft cream knit texture contrasted with relaxed dark charcoal trousers creates a minimal, cozy, slightly vintage, effortless fashion-editorial look. Calm confidence, understated masculine styling, natural movement. No music specified. No text, no logos, no filters, no beauty smoothing, no CGI appearance.

Style: Founder pitch, skincare UGC, iPhone front camera, confident but casual energy, clean bathroom or cozy bedroombackground. A 15-second vertical (9:16) founder pitch video filmed handheld on an iPhone. A young woman speaks directly to the camera naturally while holding a minimal aesthetic lotion bottle — slight camera shake, authentic lighting, no heavy editing. 0–3s — HOOK: She walks toward the mirror camera holding the lotion bottle close to her chest. “Okay, so we made this because every lotion we tried either felt greasy... or did absolutely nothing.” 3–6s: Quick cuts showing the lotion texture pumped onto her hand, smooth creamy consistency, applying it onto arms slowly. “We wanted something lightweight, clean, and actually hydrating.” 6–10s: She rubs the lotion into her skin while talking casually to the camera, showing the glow and texture on her arm. “No sticky feeling, no heavy perfume smell... it just makes your skin feel really good.” 10–13s — OFFER PUSH: She smiles softly and pushes the lotion bottle toward the camera lens with one hand, presenting it naturally to the viewer. “And honestly... once you use it, you get it.” 13–15s: She laughs softly, relaxed expression, looking down at the bottle then back at the camera. “So yeah... that’s why we made it.” Style Notes: authentic founder-story TikTok skincare ad, conversational tone, natural pauses, soft clean lighting, minimal aesthetic packaging, no text overlays, handheld movement, real skin texture visible.

hero character sheet @@hero Style: Photorealistic vertical social video, 9:16, shot as if on a handheld iPhone — no 3D render, no game engine, no digital gloss, no cyberpunk color grade. Lighting: [LIGHTING — swap per world, see below]. Color: naturalistic true-to-life color science, no teal-orange grade, no neon cast, no stylized LUT. Camera: Simulated handheld iPhone — natural hand-shake micro-jitter, slight rolling-shutter wobble on fast moves, occasional realistic autofocus micro-hunt, natural lens flare only when a light source is directly in frame. Skin: Pore-level realism — vellus hair, asymmetric features, capillary flush, natural oil sheen, pore-shadow matching on-set light, no airbrushing, no waxy uncanny-valley smoothing. Acting: Naturalistic and unscripted-feeling — micro-pauses before reactions, living eyes with catch-lights, chest rise from breathing, small imperfect human movement. Character never standing still, always reacting. Physics: Gravity and inertia respected — real weight, correct contact shadows, fabric and hair move with true momentum. No floating props. Composition: Handheld social framing — slightly imperfect rule-of-thirds, natural camera-operator sway, subject occasionally drifting off-center like a real person is filming. Continuity: Character, outfit, and identity identical across every cut. No identity drift. Technical: 24fps with natural handheld motion blur and real phone-sensor grain/compression character, no jitter artifacts, no AI smoothing. Audio: [AUDIO —sound effect only, no background music, no subtitles]. no logo no ip Prompt — creature (bird-dragon) "A large flying creature, a bird-dragon hybrid: a feathered crest running along its neck and spine, wide membranous wings with visible veining, a long scaled tail ending in a feathered tuft, powerful realistic flight anatomy and physics — heavy, weighted wingbeats, a woven leather harness saddle on its back, not a cartoonish design, cinematic realism." Prompt — sky-battle world "Sky over a naval/aerial battlefield: layered storm clouds, thick battle smoke drifting between cloud layers, distant silhouettes of warships and cannon-fire flashes, dramatic backlighting breaking through the clouds, epic scale, cinematic camera work, realism with light stylization." Prompt — hero in the same form as reference "The same hero, facial identity preserved, leather armor-corset with straps and arm guards, bow and quiver on his back, hair pulled into a battle braid, dynamic pose while diving on the back of a flying creature — body leaning forward, fabric and straps whipping in the wind, focused and determined expression, realistic motion and lighting physics." Prompt — shot and explosion "A the hero warrior on the back of a flying creature rips the bow from her back, draws the bowstring mid-dive at full speed, releases the arrow; a beat later — a massive explosion on a ship below: bright flash, flying debris, expanding shockwave; the creature and rider are thrown and knocked off balance and fall back into the sea, the blast, realistic explosion and turbulence physics, cinematic VFX realism, no cartoonish distortion." POSITIVE LOCKS: @@hero hero's face, hair, stubble, and outfit (brown leather jacket, grey henley, dark denim, tan boots, canvas satchel worn crossing the chest) are identical in every world — no outfit swaps, no accessory changes. Every world's CUT 1 opens on the exact stride/foot-strike that closed the previous world's CUT 2 — this is what makes the cut feel like one continuous step between realities, not five unrelated clips. No world's lighting or color grade should read as neon, teal-orange, or synthetic — every light source in every world must be a plausible practical (torch, sun, task light, lamp, sunrise) with naturalistic color temperature. Handheld tremor/sway stays consistent in amplitude across all five worlds — this is what reads as "one iPhone" rather than five different camera operators.

Three-shot hard-cut sequence, 8 seconds total, vertical 9:16, hyper-real documentary macro cooking footage intercut with non-photoreal 2D line-art cartoon inserts, 24fps, real-time speed throughout, no slow motion. SHOT 1 (0:00–0:01): Overhead top-down framing on a kitchen counter, a hand's fingers gripping a light-blue capacitive stylus enters frame from the lower right and taps once onto a steel ladle of pale yellow pancake batter mid-pour onto a hot griddle. On contact, the entire frame instantly converts into a flat black-and-yellow 2D line-art cartoon sketch — bold black outlines, solid yellow fill, no gradients, no shading, screen-printed poster look. The transformation happens exactly once, is instantaneous (no dissolve, no morph animation, a hard visual swap on the tap), and the batter's pour position and shape are identical in both the live-action and cartoon frame at the moment of the cut. SHOT 2 (0:01–0:04): Hard cut, same overhead top-down angle, now rendered as the black-and-yellow cartoon: the animated batter sits on a cast-iron griddle, hand-drawn bubbles rising through it one by one, a cartoon knob of butter melting and spreading in thin outlined rivulets at the edges. A stylus tap enters frame and touches the center of the pancake; on contact the cartoon instantly converts back into live-action — a real golden-brown pancake mid-flip above the same griddle, sizzling, small droplets of butter spitting at the edges, steam rising. The pancake's size, position, and browning state at the instant of the cut match exactly what the cartoon version showed the frame before — nothing is re-shown or reset. SHOT 3 (0:04–0:08): Hard cut to a dark ceramic plate shot from a slight high angle, first as the black-and-yellow cartoon: a stack of three outlined pancakes with looping hand-drawn steam lines rising off the top, animated powdered sugar sifting down from just above frame in a fine dotted trail. A final stylus tap touches the plate and it instantly converts into a live-action gourmet dish: the same three-pancake stack now photoreal, dusted with real powdered sugar, fresh raspberries and blueberries scattered beside it, two mint leaves placed on top, and maple syrup pouring in a slow amber ribbon from a small wooden syrup boat entering from the upper right, pooling at the base of the stack. The stack's height, position on the plate, and camera framing stay identical across the cut. CAMERA: fixed overhead rig, 47° FOV equivalent to a 40mm lens, locked focus at griddle/plate height, no camera movement within any shot — only the hard cuts change what's shown. No fisheye, no zoom, no gimbal drift, no camera shake. PHYSICS: batter has real fluid viscosity and settles under gravity; butter melts and pools believably; steam in the live-action shots rises and dissipates with natural buoyancy and air current drift; syrup pours as a continuous ribbon with correct trailing thickness, no floating or clipping of any prop. LIGHTING: single soft overhead kitchen light, warm 4000K, consistent direction and intensity across both the live-action and cartoon renders of each shot so the transformation reads as a style swap, not a lighting change. AUDIO: diegetic only — batter sizzling and popping on contact with the griddle, a light plastic-on-glass "tap" sound each time the stylus touches down, butter crackling, powdered sugar's soft sifting hiss, syrup's slow pour and light splash into the pool at the base of the stack. No music, no narration, no voiceover, no on-screen text. CLOSING LOCK: no logos, no watermarks, no additional hands or people beyond the single tapping hand, no on-screen text of any kind. The video ends holding on the finished gourmet plate — no fade out, no freeze frame.

A 15-second hard-cut multi-shot sequence, pure first-person POV, no face or body ever visible except hands and forearms, photorealistic cinematic action-comedy, 2.39:1, 8K, ARRI Alexa 35 with a wide rectilinear lens equivalent to 24mm (84° diagonal FOV), no fisheye, no gimbal smoothness, no digital push-ins, real body-driven camera motion with weighted inertia and overshoot on sudden direction changes. The rider sits in a sleek futuristic roller-coaster vehicle at a premium theme park, restraint bar down, gripping it with both bare forearms in frame; the world starts as an ordinary, glossy, big-budget theme-park attraction and escalates into complete absurdity while never breaking the first-person illusion. 0:00–0:02 — BOARDING: the POV rushes forward and drops into the seat with a weighted downward jolt; hands snap onto the safety bar; a sharp mechanical CLACK as it locks; the vehicle launches instantly with a hard forward g-force pull, motion blur streaking the platform away. 0:02–0:04 — THE TRACK ENDS: the coaster rockets through a tunnel lit in shifting saturated colors, wind noise rising; the track ahead visibly terminates mid-air; the vehicle flies off the rails, weightlessness hitting for a beat; hands clench the bar white-knuckled, a visible tremor of panic in the grip. 0:04–0:07 — FLYING THROUGH THE CITY: instead of falling, the vehicle rockets forward between skyscrapers, glass facades strobing past; a giant illuminated billboard fills the frame and rushes past inches from the lens; the vehicle clips a rooftop edge, camera kicking hard sideways with a violent impact shake and a metallic scrape sound, then recovers into forward motion. 0:07–0:10 — ABSURD ESCALATION: the vehicle launches off the rooftop into open air; a helicopter crosses directly through the frame at close range, rotor wash audible, missed by inches; the camera whips into a full upside-down spin, the entire city rotating beneath/around the rider, inertia visibly dragging the hands and forearms with the g-force. 0:10–0:12 — BACK TO THE PARK: the vehicle dives in a steep controlled plunge back toward the park; the coaster track rushes up to fill the frame; a second sharp CLACK as the vehicle re-seats perfectly onto the rails, a hard suspension-absorbing jolt on landing. 0:12–0:15 — THE PUNCHLINE: the vehicle decelerates smoothly into the station; the restraint bar releases with a pneumatic hiss; the hands slowly relax and drop, everything reading as calm and normal; a brief still BEAT; then a sudden WHOOSH as the vehicle launches forward again without warning; both hands instantly snap back onto the bar; hard cut to black. Camera stays locked to the rider's own eyeline at seated height throughout every shot; hard cuts change time and location but never viewpoint; only hands/forearms ever enter frame, no face, no reflection of the rider's body anywhere. Physics stay grounded even in the absurd beats: correct weight transfer on impacts, correct inertia lag on the hands during the spin, no floating or clipping geometry, no slow motion, no speed ramping — everything plays at real-time urgency except the brief still BEAT before the final launch. Lighting shifts per beat: warm practical park lighting at boarding, saturated shifting tunnel color washes through the tunnel, cool daylight/glass-reflected city light through the skyscraper run, overcast open-sky light during the spin, warm station lighting again at the return — all sources are practical/environmental, nothing stylized or artificial-looking. Audio is entirely diegetic: the mechanical CLACK of the restraint locking, wind roar and rail rumble, the metallic scrape of the rooftop clip, helicopter rotor wash, rushing city ambience, the pneumatic hiss of the bar releasing, the sudden WHOOSH of the second launch — no music, no narration, no dialogue, no voice-over, no on-screen text or subtitles. No face or body of the rider ever appears on screen. No logos, watermarks, or brand names. No freeze frame — the sequence ends abruptly on the hard cut to black immediately after the second launch grab.

Continuous 15-second home-video reel built from three hard-cut segments, one with an internal match cut, shot entirely on an early-2000s Sony DCR-TRV19 MiniDV handheld camcorder, 4:3 throughout, imperfect DV resolution, digital noise in shadows, exposure hunting and autofocus hunting recovering a beat late, white balance shifting warm-to-green indoors, motion blur, no stabilization/gimbal smoothness/push-ins/rapid pans — every camera move is body-driven handheld sway from an unseen operator among the action, never a crew move. Compositions asymmetrical: subjects off-center, never centered/mirror-balanced. Lighting flat/desaturated, muted grade, lifted blacks, gentle highlights, never glare. Skin matte/dry: no shiny/greasy/sweat-sheen/wet-look skin; rain sits on fabric/hair, not faces. Each segment/match-cut half is a different day; @jun/@mei wear unrelated, non-repeating outfits each time; no timeline continuation. Timestamp flickers bottom-right. @jun: mid-20s Malaysian Chinese man, slim, short black side-parted hair, rounded face, downcast eyes, faint awkward smile. @mei: mid-20s Malaysian Chinese woman, wavy shoulder-length hair, oval face, bright eyes, warm smile that closes thin when hurt. Identical throughout. [0.0s–3.0s, HARD CUT open] @cny_gathering, warm overhead light, muted lights, no red glow, neutral color, relatives laughing. @jun wears sage-green short-sleeve shirt. @mei wears dusty-blue blouse, simple collar, hair half-pulled back. They stand apart, eyes meet; neither smiles. Camera slowly zooms toward off-center @mei; autofocus hunts/settles; she lowers gaze as @jun looks away. HARD CUT. [3.0s–8.0s] @walkway, one continuous segment spanning two earlier days years apart, joined by an internal match cut. Opens on visibly older tape: heavier grain, duller/flatter color, one generation of quality lost, like copied from earlier cassette. Older: @mei white cotton t-shirt/denim shorts, loose ponytail; @jun striped short-sleeve polo, slightly younger. Camera follows behind/slightly side, never centered. @mei laughs, holds @jun's hand, swings it once: "你以后不可以放开哦。" (You can never let go, okay?) @jun smiles without hesitation: "嗯。" (Mm.) As fingers tighten, footage match-cuts — same grip, walkway and off-angle height — into cleaner recent tape, a later day: @jun wears charcoal t-shirt; @mei light denim jacket over floral sundress. @jun's hand drifts toward hers, hesitates, pulls back half a beat; without turning, @mei closes her hand around his. His shoulders loosen, faint smile. Operator zooms too far; image blurs then settles. @mei laughs: "不要拍啦。" (Stop filming.) HARD CUT. [8.0s–15.0s] @entrance, present day, separate from every prior segment, steady rain, weak dim/desaturated entrance light, no glare. @jun wears dark grey rain-damp jacket over white shirt, unlike earlier outfits. Single continuous over-shoulder: operator close behind @mei, framing past her; only rain-soaked shoulder/back and sliver of wet hair visible foreground, out of focus; never face. She wears a dark oversized sweater, color/texture indistinct in blur. @jun faces her/lens beyond her shoulder, clear focus, rain darkening jacket. He looks past her shoulder, unsteady: "Mei……" Throat works, hard swallow; eyes redden. "那天……如果我没有走……" He stops, unable to continue, holding her gaze. Silence. From just in front of lens, @mei asks quietly: "那你现在还会走吗？" (So would you still leave, now?) His expression breaks — lips part, shaky exhale — no answer. Neither moves. Frame holds on him; rain thickens and focus softens into blur. CUT TO BLACK. Rain realism: fabric clings with weight, water sheets off hair/shoulders, no floating/clipping, camcorder housing occasionally jostled by operator grip; skin remains matte/dry-toned, water only on fabric/hair/surface texture, never glossy. Audio diegetic-first, matching degradation: gathering hum/relatives' laughter; older walkway memory has worn optical-track texture, muffled with faint hiss under dialogue, clearing to cleaner street sound instantly at match cut. Final: steady rain; one low piano note under @jun; total silence when @mei finishes; one final low note at black. No narration, non-diegetic SFX, logo, watermark. Closing lock: no crew, camera operator or bystanders with visible faces. Camcorder never visible, including reflections in windows/glass/puddles/mirrors/polished surfaces. No centered/mirror-symmetrical composition. Final @mei face never shown — only shoulder, back of head and rain-soaked hair silhouette foreground. Skin never shiny, greasy or sweat-slicked. Every segment and match-cut half has distinct non-repeating wardrobe, reinforcing separate days across years, not one timeline. No freeze frame/fade; black cut is instant/hard. After black frame, flickering camcorder-style timestamp burns in briefly exactly "2007.06.18 21:43", then title exactly "《如果那天没有下雨》" and beneath exactly "有些人，走散了，才知道从来没有放下。", plain worn white tape-burn font, no logos/graphics.

Continuous 15-second home-video reel built from three hard-cut segments, shot entirely on an early-2000s Sony DCR-TRV19 MiniDV handheld camcorder, built-in 2.5x zoom lens, image locked 4:3 throughout, imperfect DV resolution, visible digital noise in shadows, occasional exposure hunting and autofocus hunting recovering a beat late, imperfect uneven white balance shifting slightly warm-to-green indoors, natural motion blur on fast movement, no stabilization, no gimbal smoothness, no cinematic push-ins, no rapid pans — every camera move is body-driven handheld sway from a real unseen operator standing among the action, never a crew move. Every composition is asymmetrical: subjects off-center from natural handheld imprecision, never centered or mirror-balanced. Lighting is flat/desaturated, muted grade, lifted soft blacks, practical-source highlights gentle, never hard glare. Skin is matte/dry in every segment: no shiny/greasy specular highlight, sweat-sheen, or wet-look skin even in rain; rain sits on fabric/hair, not as glossy coating on faces. Each segment is a different day; @jun and @mei wear completely different unrelated outfits every segment — no repeats, no timeline continuation. Characters: @jun — mid-20s Malaysian Chinese man, slim, short black side-parted hair, soft rounded face, gentle downcast eyes, faint awkward smile rarely full width. @mei — mid-20s Malaysian Chinese woman, shoulder-length wavy black hair usually loose, oval face, bright expressive eyes, easy warm smile that closes into a thin flat line when hurt. [0.0s–4.0s, HARD CUT open] @kopitiam, late afternoon, warm fluorescent tube overhead, softened/diffuse, greenish shadow fall-off, motorcycle and plate-clatter ambience. @mei wears mustard-yellow short-sleeve blouse with small round collar, loose hair slightly damp from heat. She sits across table from lens, off-center, looks straight into camera, head tilting a fraction. Laughing: "你一直拍我做么？" (Why do you keep filming me?). @jun, plain grey crewneck t-shirt, heard breathing an awkward laugh behind camera, swings lens down fast toward table; frame briefly blurs then re-settles on fries. @mei reaches into frame, steals a fry, grinning at lens; first spark of chemistry easy and unforced. HARD CUT. [4.0s–9.0s] @rainy_street, different night, overcast/flat, diffuse ambient light, no neon glow/colored shopfront spill — plain streetlight and grey wet asphalt, muted/ordinary, not moody. @jun wears plain white t-shirt, soaked through and clinging, cradling camcorder under hem for cover as they run. @mei wears pink t-shirt, sleeves damp/heavy, screams a laugh and jumps onto his back from behind; he staggers catching her weight, footing sliding on wet ground before steadying. Footage shakes hard on impact, rain streaks lens, both laugh uncontrollably over downpour — skin stays matte/dry-toned; water visible on fabric/hair, not glossy on faces. HARD CUT. [9.0s–15.0s] @apartment, separate daytime, flat overcast window light, desaturated/even, no artificial key, no dramatic contrast. @jun wears faded olive-green button-up, sleeves rolled forearm, sitting stiffly on chair edge. @mei wears loose grey oversized t-shirt and pulled-back hair, at-home clothing distinct from earlier outfits. Camcorder static on table, angled off-axis, not square to room, asymmetrical wide shot — no operator holding it, frame does not move. @jun and @mei argue quietly mid-distance, voices slightly muffled by built-in mic. @mei: "你每次都这样。" (You're always like this.) @jun says nothing; jaw tightens almost imperceptibly, eyes drop floor. She waits, watches him; expression shifts tight anger to slower, heavier disappointment. Quietly: "算了。" (Forget it.) She exits frame left. @jun stays seated off to one side, unmoving, camera holds without cutting. Rain-soaked physical realism where relevant: fabric clings/drags with real weight, footfalls skid on wet pavement with visible weight transfer, water sheets off hair/shoulders rather than static, no floating/clipping props, camcorder housing occasionally jostled by operator grip, brief handheld micro-shake distinct from character motion. Audio diegetic-first: kopitiam plate-clatter/motorcycle traffic opening; heavy rain/distant traffic hiss running; thin nostalgic acoustic guitar fragment softly under first two segments, thinning near nothing by argument; near-total room tone with faint distant chatter under apartment argument. No narration, non-diegetic SFX, logo, watermark. Closing lock: no crew, camera operator. Camcorder itself never visible, including reflections in windows, glass, puddles, mirrors, polished surfaces. No centered/mirror-symmetrical composition. Skin never shiny/greasy/sweat-slick. Each segment wardrobe distinct/no repeats, reinforcing separate days, not one continuous day. No freeze frame, fade, or on-screen text this segment.