视频 · Seedance 2.5
秋日城市的一天
制作一支30秒真实自然的秋日城市生活短片,跟随同一位女性从出门、乘地铁、买咖啡到夜间回家,呈现多场景纪实式 vlog 片段。
Prompt
场景背景
一位20岁的东亚女性度过秋日的一天,穿行于大城市之中——离开公寓、乘坐地铁、在推车上买咖啡、穿过落叶纷飞的公园、浏览书店、在屋顶捕捉黄金时刻,并在夜间乘车回家——以纪实般的混合格式都市一日vlog拍摄,总时长30秒。
有效参考
@image1:20岁的东亚女性,圆润柔和的脸庞,温暖的象牙色肌肤,深棕色头发松散地挽成高髻,脸颊旁留有几缕碎发,自然水润的妆容,温柔而富有表现力的眼睛,身穿酒红色罗纹针织一字肩毛衣,佩戴一条小巧的银色花朵吊坠项链。100%在每个镜头中都与参考保持一致。
格式模式
限时多镜头。仅在规定秒数处进行硬切;镜头之间不会自行切换。混合的场景风格——并非每个镜头都有对白,部分为无声的纪实空镜。
0.0s–4.0s — 硬切
4.0s–8.0s — 硬切
8.0s–12.0s — 硬切
12.0s–17.0s — 硬切
17.0s–21.0s — 硬切
21.0s–26.0s — 硬切
26.0s–30.0s — 结束
光学参数
镜头1:63° 广角,近似视角的手持。镜头2:47° 中性,几乎静止的微倾。镜头3:29° 人像压缩感,快速交易式。镜头4:18° 自然人像,慢动作细节。镜头5:29° 人像压缩感,抓拍式、未察觉镜头的角度。镜头6:47° 中性,近似双人镜头的单人举杯。镜头7:18° 自然人像,贴近的沉思感。段落中途不产生漂移。
摄影机
全程手持,带有自然的1–2cm手部抖动。镜头1:手机高度的近似视角,以4 km/h 的速度行走。镜头2:相机贴靠在杆子旁低位放置,几乎不动,观察式。镜头3:从手部快速甩向脸部再甩回。镜头4:以半速慢放,相机跟随飘落的树叶,以1 km/h 向侧面漂移。镜头5:相机似乎让她猝不及防,她稍有延迟才注意到。镜头6:近乎静止,支撑放置,她走入画面。镜头7:透过车窗玻璃静止拍摄,反光掠过她的脸。
动作
0.0s–4.0s:她从公寓楼走出,踏上落叶遍地的人行道,调整包带,瞥一眼手机,清晨低矮温暖的阳光穿过光秃的树枝。她短暂地抬头看向镜头,半笑着说:"Okay — let's do this."
4.0s–8.0s:在地铁车厢内,她倚着立杆,戴着耳机,看着站台灯光从窗外模糊掠过,城市声音沉闷而有节奏。无对白——只有她安静、从容的表情,随列车晃动偶尔轻摇。
8.0s–12.0s:在街边一个小咖啡推车旁,机器冒着蒸汽,她递上现金,接过纸杯,立即双手捧住杯子取暖。简短的交流,她只说:"Thanks, you're the best."
12.0s–17.0s:慢动作穿过铺满金色落叶的公园小径,她在迈步中途仰起头,转身一圈,树叶落在她的头发和毛衣上,发出真实不设防的笑声,没有台词——只有呼吸与笑声可闻。
17.0s–21.0s:在一家小书店内,她用手指沿着书架划过,抽出一本书一半,翻了几页,没有察觉镜头,柔和的窗光洒在她脸上,店内环境安静。无对白。
21.0s–26.0s:黄金时刻的屋顶上,身后是朦胧的城市天际线,她将外带咖啡杯微微举向镜头,像一个小小的祝酒,温暖的低角度夕阳给她的头发打上边缘光,并说道:"To days like this."
26.0s–30.0s:夜间出租车内,温暖的琥珀色路灯光带在车窗玻璃上滑过,并以柔和的流动光带掠过她的脸,她将头靠在座椅上,带着疲惫而满足的微笑,望向镜头并说:"Best day, hands down."
表演
克制、自然的反应——没有刻意表现的兴奋。镜头2和5特别呈现出抓拍、未察觉镜头的感觉。每个特写镜头中眼睛都要有神并带有眼神光,镜头1中可见呼出的白气,各处皮肤与织物质感自然。
物理效果
树叶飘落并散开,具有真实的重量与漂移,短暂地挂在头发和毛衣纤维上,然后飘落离开。咖啡蒸汽自然升起并消散。路灯光带随车辆移动,在她的脸和车窗玻璃上呈现正确的视差。列车晃动以微妙而可信的方式改变她的重心。
灯光
镜头1:低角度的温暖晨光穿过光秃的树枝,柔和的阴影线条。镜头2:冷色荧光地铁灯光,窗外闪过温暖的实用光源。镜头3:阴天日光,柔和均匀的主光。镜头4:斑驳的温暖阳光穿过公园树冠,树叶边缘逆光。镜头5:柔和的窗户漫射日光,低反差。镜头6:温暖的低角度黄金时刻主光,强烈的轮廓光,柔和的薄雾。镜头7:温暖的琥珀色钠灯路灯光带,其余内部保持黑暗。
色彩分级
温暖的秋季色调——琥珀、锈红与柔和的金色,搭配灰度偏低的城市灰色。夜晚为温暖的琥珀色路灯,画面中任何地方都不出现霓虹灯或高饱和彩色招牌。不带冷青色或品红色偏色。
服装
酒红色罗纹针织一字肩毛衣与银色花朵吊坠项链,全程保持不变,并在每个镜头中清晰可见。在地铁和出租车镜头中,可以有一条随意搭着的围巾或敞开的外套盖在其上,但不遮挡它。
音频
镜头1:安静的街道环境音,树叶沙沙作响。镜头2:闷闷的火车轰鸣声,远处的报站声,无对白。镜头3:意式浓缩咖啡机的嘶嘶声,简短的交流。镜头4:落叶的沙沙声,自然的笑声,无台词。镜头5:安静的书店环境音,翻页声,无对白。镜头6:远处城市的嗡鸣声,风声。镜头7:引擎嗡鸣,她台词下方有柔和的收音机低语声。
风格
真实的日常生活vlog摄影,梦幻的35mm胶片美学,细腻自然的颗粒,真实的自动对焦搜索,穿插说话与无声抓拍镜头,营造生活化、非剧本式的质感。
输出设置
4K,24fps,35mm颗粒,除镜头4(半速慢动作)外为实时速度,真实的手持运动,色彩方案中任何地方都没有霓虹灯或高饱和招牌。
正向锁定
所有镜头中同一位女性的脸部与体型保持一致,与@image1完全匹配。酒红色毛衣与银色项链在每个镜头中都保持可见。头发全程保持同一款松散高髻。切换仅发生在规定秒数,镜头不会自行切换。镜头2和5保持无对白且具抓拍感。温暖的琥珀色调全程保持——任何地方都没有霓虹灯,没有冷青色偏色。SCENE CONTEXT
A 20-year-old East Asian woman spends an autumn day moving through a big city — leaving her apartment, riding the subway, grabbing coffee from a cart, cutting through a park in falling leaves, browsing a bookstore, catching golden hour on a rooftop, and riding home at night — filmed as a candid, mixed-format day-in-the-city vlog, 30 seconds total.
ACTIVE REFERENCES
@image1: 20-year-old East Asian woman, soft round face, warm ivory skin, dark brown hair in a loose high bun with face-framing strands, natural dewy makeup, gentle expressive eyes, wearing a wine-burgundy ribbed knit off-shoulder sweater and a small silver flower pendant necklace. 100% matches the reference throughout every shot.
FORMAT MODE
Timed multishot. HARD CUTs only at the stated seconds; camera does not cut on its own between them. Mixed scene styles — not every shot has dialogue; some are silent candid b-roll.
0.0s–4.0s — HARD CUT
4.0s–8.0s — HARD CUT
8.0s–12.0s — HARD CUT
12.0s–17.0s — HARD CUT
17.0s–21.0s — HARD CUT
21.0s–26.0s — HARD CUT
26.0s–30.0s — end
OPTICS
Shot 1: 63° wide, POV-adjacent handheld. Shot 2: 47° neutral, static-ish lean. Shot 3: 29° portrait-compression, quick transactional. Shot 4: 18° natural-portrait, slow-mo detail. Shot 5: 29° portrait-compression, candid unaware angle. Shot 6: 47° neutral, two-shot-adjacent solo toast. Shot 7: 18° natural-portrait, close reflective. No drift mid-segment.
CAMERA
Handheld throughout with natural 1–2cm tremor. Shot 1: phone-height POV-style, walking at 4 km/h. Shot 2: camera rests low against a pole, minimal movement, observational. Shot 3: quick whip from hand to face and back. Shot 4: slowed to half real-time, camera drifts sideways at 1 km/h following falling leaves. Shot 5: camera appears to catch her off-guard, slight delay before she notices it. Shot 6: static-ish, propped, she moves into frame. Shot 7: static through car window glass, reflections passing over her face.
ACTION
0.0s–4.0s: She steps out of an apartment building onto a leaf-scattered sidewalk, adjusting her bag strap, glancing at her phone, morning light low and warm through bare branches. She looks up at the camera briefly, half-smiles: "Okay — let's do this."
4.0s–8.0s: Inside a subway car, she leans against a pole, earbuds in, watching the platform lights blur past the window, city sounds muffled and rhythmic. No dialogue — just her quiet, unbothered expression, occasional sway with the train's motion.
8.0s–12.0s: At a small street coffee cart, steam rising from the machine, she hands over cash, receives a paper cup, wraps both hands around it immediately for warmth. Brief exchange, she says only: "Thanks, you're the best."
12.0s–17.0s: Slow-motion through a park path thick with falling amber leaves, she tips her head back mid-step and turns once, leaves catching in her hair and sweater, genuine unguarded laugh, no spoken line — just breath and laughter audible.
17.0s–21.0s: Inside a small bookstore, she runs a finger along a shelf, pulls a book halfway out, flips a few pages, unaware of the camera, soft window light across her face, ambient store quiet. No dialogue.
21.0s–26.0s: On a rooftop at golden hour, city skyline hazy behind her, she raises a takeaway coffee cup slightly toward the lens like a small toast, warm low sun rim-lighting her hair, says only: "To days like this."
26.0s–30.0s: Inside a taxi at night, warm amber streetlights sliding across the window glass and over her face in soft passing bands, she rests her head against the seat, tired contented smile, looks toward the lens and says: "Best day, hands down."
PERFORMANCE
Restrained, natural reactions — no performed excitement. Shots 2 and 5 specifically read as candid/unaware of the camera. Living eyes with catch-light in every close shot, visible breath in shot 1, natural skin and fabric texture throughout.
PHYSICS
Leaves fall and scatter with real weight and drift, catching briefly in hair and sweater fibers before falling away. Coffee steam rises and dissipates naturally. Streetlight bands move across her face and the window glass with correct parallax as the car moves. Train sway shifts her weight subtly and believably.
LIGHTING
Shot 1: low warm morning sun through bare branches, soft shadow lines. Shot 2: cool fluorescent subway light with warm practicals flickering past outside. Shot 3: overcast daylight, soft even key. Shot 4: dappled warm sun through park canopy, backlit leaf edges. Shot 5: soft window-diffused daylight, low contrast. Shot 6: warm low-angle golden hour key, strong rim light, soft haze. Shot 7: warm amber sodium streetlight passing bands only, otherwise dark interior.
COLOR GRADE
Warm autumn palette — amber, rust, and soft gold tones with muted city grays. Warm amber streetlight at night, never neon or saturated colored signage anywhere in frame. No cool-cyan or magenta tint.
WARDROBE
Wine-burgundy ribbed knit off-shoulder sweater and silver flower pendant necklace, unchanged and clearly visible in every shot. A loosely draped scarf or open coat may sit over it in the subway and taxi shots without covering it.
AUDIO
Shot 1: quiet street ambience, rustling leaves. Shot 2: muffled train rumble, distant announcement, no dialogue. Shot 3: espresso machine hiss, brief exchange. Shot 4: crunch of leaves, natural laughter, no line. Shot 5: quiet store ambience, page-turning, no dialogue. Shot 6: distant city hum, wind. Shot 7: engine hum, soft radio murmur under her line.
STYLE
Authentic day-in-the-life vlog cinematography, dreamy 35mm film aesthetic, fine natural grain, realistic autofocus hunting, mix of talking and silent candid shots for a lived-in, unscripted feel.
OUTPUT SETTINGS
4K, 24fps, 35mm grain, real-time except shot 4 (half real-time slow motion), realistic handheld motion, no neon or saturated signage in the color palette anywhere.
POSITIVE LOCKS
Same woman's face and build in every shot, matching @image1 exactly. Wine-burgundy sweater and silver necklace stay visible in every shot. Hair stays in the same loose high bun throughout. Cuts only at stated seconds, camera does not cut on its own. Shots 2 and 5 stay dialogue-free and candid. Warm amber tones hold throughout — no neon, no cool-cyan tint anywhere.
使用建议
使用建议:先上传人物参考图,再在支持多镜头时间轴、角色一致性、对白与环境音的图生视频工具中使用;核对每段硬切时间、服装连续性和夜景色调,必要时分镜生成后再剪辑合成。
信息
- 类型
- 视频
- 模型
- Seedance 2.5
- 许可
- Source-author-rights
- 发布时间







