Fisheye Bedroom Punk Aesthetic Video
Seedance 2.5· Video
Create a single continuous 15 second live action shot in full 16:9 with no cuts. A young adult woman with platinum blonde hair, wispy straight bangs and two loose low side buns performs directly into an extreme fisheye camera inside a dim intimate bedroom. She wears dramatic black eyeliner, subtle pale makeup, a fitted black sleeveless crop top and several thin layered black necklaces with small pendants. Give her a confident alternative punk aesthetic without making the styling look like a costume. Place the camera completely still at approximately face height using an extremely wide fisheye lens. Preserve strong optical barrel distortion so the room bends naturally around the edges and her facial proportions change realistically whenever she approaches the glass. Fill the entire widescreen frame with real environmental detail rather than black side bars. Behind her are rough aged walls, small grid like architectural openings, muted blue bedding and miscellaneous lived in bedroom details disappearing into darkness. Lighting is very low and imperfect, dominated by cold cyan and green illumination from near the camera with faint ambient light in the room. Her pale face occasionally approaches mild highlight clipping while the background has deep crushed shadows. Maintain authentic high ISO sensor noise, subtle chromatic contamination, soft low resolution texture, slight motion smear, imperfect white balance and gentle exposure breathing. Avoid cinematic beauty lighting and avoid an artificially polished image. At the beginning her face is extremely close to the lens while she mouths a song with relaxed half closed eyes. During the first three seconds she gradually leans backward until her upper torso becomes visible, shoulders subtly moving to the rhythm. Around four seconds she lowers her chin and then leans toward the camera again. Between five and seven seconds she looks upward while continuing to mouth the vocal, slowly lifting her chin and exposing her neck. From seven to nine seconds she rolls her head sideways in one fluid natural movement, allowing her loose hair strands and necklaces to react independently. From nine to twelve seconds she returns toward center, dips her chin, gives a soft pout and briefly lowers her eyes. Keep tiny breathing motion in her chest and shoulders. From twelve to fourteen seconds she straightens and slowly approaches the camera again. During the final second she becomes noticeably more playful, moves extremely close to the fisheye glass and finishes with a wide spontaneous smile and teasing expression directly into the lens. All body motion must obey realistic inertia. Hair, necklaces, clothing and facial muscles move independently and consistently. Maintain stable identity, correct hands, natural teeth, realistic eyes, skin pores and believable blinking. No morphing, floating jewelry, duplicated features, frozen hair, beauty filter skin, artificial depth of field, camera cuts, text, logos
Prompt terkait
Mirror Selfie with Your Favorite Character
GPT Image 2· Gambar
A realistic cinematic mirror selfie in a dark bedroom, photographed vertically in 9:16 format on an iPhone 17 Pro Max. Strong natural motion blur and soft focus caused by slight hand movement. Warm dim ambient lighting only, no flash. Grainy texture, low exposure, indie-film atmosphere.
Exact composition:
Three people standing extremely close together in front of a tall mirror.
[picture 1] = reference for the LEFT person
[picture 2] = reference for the LEFT person
[picture 3] = additional reference for RIGHT person
[picture 4] = additional reference for RIGHT person
[picture 5] = reference for the CENTER person
CENTER person:
Using the phone to take the mirror selfie. Positioned slightly lower than the other two people. Neutral, detached expression. Body facing the mirror directly.
LEFT person:
Standing very close on the left side of the center person, leaning inward toward them. Head tilted slightly downward toward the phone. One shoulder slightly forward.
RIGHT person:
Standing even closer on the right side, leaning their head toward the center person. One hand gently holding the center person’s chin/jaw.
Camera/framing:
Mirror selfie composition.
Upper bodies and thighs visible.
The phone partially blocks the center figure’s torso.
The right figure occupies slightly more space in the frame than the left figure.
Shot at a slight diagonal angle instead of perfectly straight.
The bed appears in the lower-right background with messy white blankets and crumpled sheets.
Dark empty wall in the background.
Subjects nearly filling the entire frame vertically.
Style and atmosphere:
extreme realism, cinematic photography, soft shadows, muted blacks and browns, realistic skin texture, shallow depth of field, heavy natural blur, melancholic atmosphere, intimate tension, luxury editorial aesthetic, Korean drama mood, indie movie still, imperfect framing, accidental-looking composition, emotionally charged body language
Important:
The identities, hairstyles, facial structures, and overall appearance of each figure MUST follow the uploaded reference photos exactly. Keep the pose and composition consistent with the described scene while adapting all faces and features to the provided references.
Negative prompt:
anime, cartoon, illustration, CGI, 3D render, overexposed lighting, studio lighting, smiling at camera, sharp focus, extra limbs, deformed hands, bad anatomy, duplicate people, low realism, clean bedroom, colorful background, symmetrical pose, stiff posture, watermark, text, logo
Flashlight in a Night Cornfield (A24 Indie Horror)
Seedance 2.0· Video
Wide angle camera view of a large corn field with a woman at its edge pointing her flashlight into the field at night. The sky is pitch black, the cornfield is taller than her, the grass is wet with dew and dark green. She is tall with very long full blonde layered highlighted hair draping down her back, she has long wispy bangs, blue eyes, big soft pink lips, light skin with subtle texture and pores, she is wearing a hot pink tube top, low rise flared jeans, a small gold cross necklace, and flip flops. The only light comes from her flashlight casting into the field. Realistic rendering, shallow depth of field, A24 indie film style.
Cat Sneaking Out of the House at Night
Flux 3· Video
A dark blue night sequence following a house cat: it slips off a bedside cabinet while its owner sleeps under a duvet, pads across a beige kitchen counter past fruit bowls and a kettle in the dim under-cabinet light, drops down into a black garden and moves through wet leaves and tall plants, is seen from outside as a silhouette in the lit kitchen window of a brick house, walks along the ridge of the roof tiles against the night sky, then disappears into dense dark foliage. Continuous cinematic night photography, deep shadows, cool moonlight, very low key. Audio: house quiet, distant traffic, rustling leaves.
Ancient-Style Costume OOTD Short Video
Seedance 2.5· Video
7秒，9:16，一镜到底，甜美古风OOTD短视频。7秒音频完整播放，配合人物舞蹈。 严格以@人物参考图人物设定图中的成年女性为唯一人物：以右侧脸部特写作为人物长相参考，以左侧正面图 and 中间背面图作为服装、发型及背部结构参考。完整保持她的小巧鹅蛋脸、清秀五官、红色眉心花钿、深棕黑色高盘古典发髻、浅金色花叶发簪、珍珠流苏和两侧垂落珠链；完整保持浅香槟白与淡桃粉古风舞衣、刺绣抹胸内搭、浅杏色透明长袖纱衣、浅金色腰封、金属链饰、多层垂坠长裙。全程禁止换脸、换装、改变发型或增减饰品。 仅参考@视频人物的动作轨迹、姿态、节奏、景别变化和构图，不继承原视频人物的脸、发型、服装、颜色或材质。纯白无影棚，高机位轻微俯拍，超广角轻微透视，镜头基本固定，人物始终保持全身或大半身清晰入镜。 0–1秒，人物从画面右侧提着少量裙摆轻快走入，长裙和薄纱袖自然摆动； 1–2秒，人物靠近镜头，上身轻微前倾，抬眼注视镜头，神情清甜克制； 2–3秒，双手轻扶腰间，形成自然收腰姿态，俏皮微微歪头； 3–4秒，优雅侧身转过半圈，展示服装背面、盘发和腰封结构，再回眸看向镜头； 4–5秒，转回正面浅浅微笑，双臂向身体两侧自然舒展，轻纱袖缓慢垂落； 5–6秒，双手轻放身后，先低头整理姿态，再微微歪头看向镜头； 6–7秒，再次轻扶腰间，人物先扭头看向与镜头呈90°角的侧前方，视线先聚焦在镜头外某个远方，然后再转向镜头，以甜美、自信、端庄的古风姿态定格结束。 动作自然连贯、轻盈克制，不做大幅度舞蹈动作。裙摆、薄纱衣袖、珠链流苏和耳侧发丝随动作产生轻微真实惯性，避免夸张飞舞。明亮柔光，清透高亮，柔和浅香槟色调，古装影视广告与韩系时尚短片结合的高级质感，节奏轻快，无对白，搭配轻盈空灵的国风流行音乐。人物身份、五官、发型、服装纹样和饰品位置全程稳定，面部清晰稳定，真实皮肤纹理，4K质感，细节清晰。 不要使用左侧正面图中的空白遮脸作为人物面部，人物长相必须来自右侧脸部特写；不要继承视频中原人物的脸、发型或服装；不要新增人物，不要换装，不要跳切，不要镜头大幅移动或抖动，不要快速旋转，不要剧烈甩袖，不要裙摆遮脸，不要肢体变形、重复手臂、多余手指、手部粘连、穿模、服装突变、饰品消失、发型变化或背景变化。
Smartphone-Style Fashion Model Sequence
Seedance 2.5· Video
[生成ゴール]
テンポは軽快で速く、繰り返しのリズムを持つ。広告映像でも映画的映像でもなく、友人がその場でスマホで撮った実写クリップに見えること。
[ステージ1｜0-5秒]
初期状態：固定フレームには日の当たる木壁だけが映っている。 主要イベント：<モデルA>が右端から速い足取りで入り、中央で立ち止まり、腰を一度切り返してプリーツスカートを翻して落ち着かせ、レンズへ正対しながら閉じていた口元が緩んで目を細めた笑顔になる。その後、左端から歩いて抜ける。 終了状態：フレームは再び無人になり、右端は空いている。
[ステージ2｜5-10秒]
前段からの継続：カメラは動いておらず、フレーミングは同一. 主要イベント：<モデルB>が右端から歩いて入り、中央でレンズに背を向けて立ち止まり、肩に羽織ったカーディガンとワンピースのティアードの裾を見せ、そこから肩越しに顔をカメラへ向ける。表情は作らない静かなもので、唇はわずかに開いている。身体を完全にレンズ側へ向け直し、左端から歩いて抜ける。 終了状態：フレームは再び無人になり、右端は空いている。
[ステージ3｜10-15秒]
前段からの継続：カメラは動いておらず、フレーミングは同一. 主要イベント：<モデルC>が右端から歩いて入り、中央で立ち止まり、上半身を一度右へ回してニットベストとワイドリネンパンツのシルエットを見せ、そこからレンズへ1歩近づいて顔を寄せ、肩を上げて声を出して笑い、太ももから上が収まるフレームになる。レンズから約1メートルより近づくことはない。中央の位置まで戻って身を起こし、笑顔のままそこに留まる。 終了状態：<モデルC>は中央でカメラに正対して立ち、手持ちのフレームが彼女の上で落ち着いて止まる。
Vintage Fisheye Fashion Film in a Parking Garage
Seedance 2.5· Video
Create a 15 second cinematic underground fashion film in 16:9 inside a real concrete multilevel parking garage at night. Two stylish adult women are the central subjects. One has long straight black hair with blunt bangs, narrow dark futuristic sunglasses, a distressed silver and black sleeveless graphic top, layered metallic necklaces, stacked bracelets, fingerless black gloves, loose dark trousers, chunky black footwear and a large worn black shoulder bag. The second has long pale blonde hair under a fitted black beanie, oversized rectangular black sunglasses, a fitted white top, oversized black jacket slipping casually from her shoulders, a short dark skirt, black thigh high socks, chunky shoes, silver jewelry and a large textured handbag. Use an extremely wide vintage fisheye lens with visible barrel distortion and strong perspective stretching. The camera should feel like a real person holding a small early digital camcorder at arm length. Preserve tiny framing mistakes, slight horizon drift, natural hand vibration, physical footsteps, unpredictable body proximity, realistic autofocus breathing and occasional motion blur. Never make movement perfectly stabilized or mechanically smooth. Open extremely close to both women crouching beside a black motorcycle, staring calmly into the lens while a fluorescent fixture blooms above them. Rapidly orbit and tilt around them as they adjust sunglasses, move their hair and casually interact with the camera. Cut to a dramatic floor level detail of the motorcycle engine, rear wheel, frame, dark metal and chrome reflecting harsh ceiling light. Show both women crouching over an open bag with believable personal objects scattered on the concrete. Insert a brief near black rhythmic cut. Return with an intimate overhead fisheye angle while they search through the bag. The blonde woman raises a tiny silver compact camera toward the lens. One woman stands and crosses the painted white pedestrian stripes while the other collects her bag. Follow them through the garage with loose handheld backward movement. Include concrete pillars, curved driving lanes, yellow curb paint, white directional markings, a black parked car and a white van deep in the background. Around the middle create one very brief three image editorial collage of the dark haired woman from different close perspectives, then immediately return to natural live action. Show her walking away carrying the heavy black bag before both women turn and confidently walk back toward the camera. As they approach, allow the fisheye perspective to enlarge their faces, hands and clothing naturally. Finish by returning to the motorcycle where they crouch close to the camera again, creating a seamless visual loop to the opening shot. Lighting must come almost entirely from practical cold fluorescent garage fixtures with slightly green white color, blown highlights around lamps, deep soft shadows under eyes and clothing, subdued concrete colors
Avant-Garde Editorial with Tech Overlay
GPT Image 2· Gambar
Ultra-realistic fashion editorial portrait of a confident woman in avant-garde white couture with dramatic structured shoulders, hands reaching toward camera (perspective distortion), intense gaze, long wind-blown messy hair, studio soft lighting, grey textured background, high contrast, Vogue-style.
Add overlays: face detection box centered on face, thin white frame with corner handles, subtle grid, UI markers, cursor, tracking points, experimental tech overlay, magazine layout with bold text “AVANT MODE”, barcode. Ultra-sharp, cinematic lighting, 8K, professional fashion photography, --ar 3:4 --v 6 --style raw
Chinese Surrealist Dark Mythology Scene Fused
GPT Image 2· Gambar
Chinese surrealist dark mythology scene fused with Chinese aesthetics — perfect details, multi-pipeline rendering, perfect modeling. Journey to the West setting, Lion-Camel Ridge, countless demons and monsters. On the left: massive throne with the great elephant king in heavy armored demon form; in the center: the lion king's throne; on the right: a giant roc. Three demon kings each with a spectacular visual. Ultra-detailed, dark fantasy atmosphere.
Flat Illustration
GPT Image 2· Gambar
Flat illustration, Eastern fantasy style premium city poster design, vertical 9:16 composition. Overall diagonal + S-curve flowing layout extending from lower-left to upper-right. Deep black background fading to dark crimson creates strong warm-cool contrast and spatial depth with subtle stardust and particle texture. A golden flowing energy line like a flame winds from the bottom upward, with fluid texture, particle light effects, and gradient highlights. Layered Guangzhou city landmarks emerge from the golden flow: Canton Tower as visual core, surrounded by Zhujiang New Town skyscrapers, Liede Bridge, modern and Lingnan architectural elements — detailed line art + glowing golden forms, outlines clear, hovering in void. An Eastern white-haired female figure at the bottom — flowing hair blending with the golden stream, translucent strands with gradient light, graceful posture, eyes gently closed, holding a bouquet of colorful flowers with light particles. Top text: "Guangzhou · China" in center Song typeface, below "2026/04/20", then "LIYUE". Ratio 9:16.
Premium Dark Mode SaaS Landing Page Design Mockup
GPT Image 2· Gambar
Premium dark mode SaaS landing page design mockup. Deep navy/black background, glassmorphism cards, neon blue accent highlights. Sections: hero headline, feature grid, pricing cards, CTA button. Korean typography labels '주요 기능', '요금제', '시작하기'. Modern Korean tech startup aesthetic. 16:9.
Automatically Generate a Top-tier Dark Epic Concept Poster / Cinematic Infographic Poster
GPT Image 2· Gambar
Automatically generate a top-tier dark epic concept poster / cinematic infographic poster based on [theme].
The only required input is:
[Theme]: ___ Trump's Thinking ___
Everything else is automatically determined by AI, including: main subject (auto-determine whether human figure, guardian, warrior, product, artifact, statue, or abstract symbol), central structure (auto-determine: throne, stone base, altar, mechanical platform, ruins, elevated stage), environment (auto-determine: cave, temple, ruins, abyss, underground palace, chamber), light source (auto-determine: moonlight, divine light, energy beam, judgment light, holy light), symbolic elements (skulls, heraldry, inscribed stones, emblems, runes, energy rings, divine symbols), color system, material combinations, title/subtitle/copy, layout and narrative tone.
Overall style: High-budget 1990s Hollywood epic movie poster aesthetic, combining cinematic matte painting, hyperrealistic photographic quality, extreme chiaroscuro, heavy spatial narrative, dark heroism, and ritualistic composition. Must look like a genuine movie main poster.
Core structural requirements:
1. A massive, oppressive dark enclosed space
2. A strong beam of directional volumetric light cutting from above — the primary visual order
3. Core subject on a central platform at the light endpoint
4. Lower left: dense title and information layout.
Dramatic
GPT Image 2· Gambar
A dramatic, high-contrast studio portrait of a {argument name="subject gender" default="male"} athlete or model in the visual style of a premium sports drink advertisement, centered and facing straight toward the camera in a tight head-and-shoulders crop. The subject has {argument name="hair style" default="short dark hair brushed back"}, visible ears on both sides, and a rugged lower face with a short beard or stubble. Dress him in a dark zip-up athletic jacket with the zipper centered and visible near the collar. Use an almost entirely black background and extremely low-key lighting, with subtle rim light and soft highlights catching the hair, ears, jawline, shoulders, and jacket texture while most facial features remain swallowed by shadow for a mysterious, intense mood. The image should feel monochrome or nearly monochrome, with deep blacks, muted gray highlights, cinematic contrast, gritty texture, and a sleek commercial sports-brand aesthetic reminiscent of a {argument name="brand style" default="Gatorade"} campaign. Vertical composition, minimalist framing, no text, no logo, no props, no visible environment.