[Reference Role]
@image1 defines only the woman's facial identity, eye shape, hairstyle, hair ornament, yukata and the white baby spotted-seal plush, plus the rough composition and shot order of each panel. @image1 is an 8-panel storyboard grid — do not reproduce its grid layout, panel divisions, split-screen composition, timestamps or caption text. Read the panels left to right, top row then bottom row, and use them as the shot order of a single continuous vlog.
[Single Instance] — critical
All eight panels in @image1 show the SAME one woman at different moments. The output must contain exactly ONE instance of her at all times. Never show two of her, never show her duplicated, mirrored, split-screen or in separate panels. The background crowd must not contain anyone resembling her, wearing the same yukata, or with the same hair ornament.
[Subject Profile: HANA]
A Korean woman in her early 20s with hooded inner double eyelids — long, narrow, almond-shaped eyes with a calm, composed gaze, not wide or round. Jet-black hair in a soft updo with damp flyaway strands, a white tsumami-zaiku flower ornament on the left side. Fair skin with visible pores and a faint sheen of sweat. Off-white yukata with pale blue floral pattern and a cream obi. Identity, eye shape, hair ornament and yukata stay identical in every shot.
[Style]
Ultra-photorealistic handheld smartphone vlog footage, front-camera selfie POV unless stated otherwise. Realistic skin texture, natural fabric physics, mixed warm lantern light and cool stall light, high-ISO noise in shadows, slight motion blur, amateur phone-camera quality. Fast vlog-style cuts. No AI-looking smoothness. No background music.
[Shots]
Shot 1: Selfie close-up on a lantern-lit festival street, HANA smiles into the lens and says in Korean: {여름 축제 왔어요!}
Shot 2: Selfie at a takoyaki stall, steam rising, she holds the paper tray toward the camera. A male vendor beside her says in Japanese: {はい、熱いよ!}
Shot 3: Extreme close-up, low selfie angle, she bites a glazed dango off a wooden skewer, cheeks puffed, no dialogue. Red lantern bokeh behind.
Shot 4: Shot from behind — her yukata back and obi fill the frame as she points at a bright prize stall, arm slightly motion-blurred.
Shot 5: Selfie hugging a white baby spotted-seal plush against her cheek at the prize stall, small proud smile, no dialogue.
Shot 6: Back of her head in the lower foreground, hair ornament in focus, a large firework blooming in the dark sky above the crowd, smoke haze drifting.
Shot 7: Low selfie angle, she looks up off-camera as fireworks flash across one side of her face, lips parted in quiet wonder.
Shot 8: Selfie on the lantern-lit street, she waves an open palm close to the lens and says warmly in Korean: {함께 와줘서 고마워요. 다음에 또 봐요. 안녕!}
[Maintain Consistency]
Exactly one HANA in every shot. Same face, same
[参考角色]
@image1 仅用于定义女性的面部特征、眼型、发型、发饰、浴衣以及白色斑点海豹幼崽玩偶,并提供每个分镜的大致构图和拍摄顺序。@image1 是一个 8 分镜的项目网格图 —— 请勿复刻其网格布局、分镜分割、分屏构图、时间戳或说明文字。请从左到右、先上后下阅读这些分镜,并将它们作为一段连续 Vlog 的拍摄顺序。
[单一实例] — 关键
@image1 中的所有八个分镜展示的都是“同一个”女性在不同时刻的画面。输出内容必须始终且仅包含“一个”她的实例。严禁出现两个她,严禁出现复制、镜像、分屏或在不同分镜中同时出现的情况。背景人群中不得出现任何长相与她相似、穿着相同浴衣或佩戴相同发饰的人。
[主体资料:HANA]
一名 20 岁出头的韩国女性,拥有内双眼皮 —— 长而窄的杏仁眼,眼神平静从容,非圆眼或大眼。乌黑的头发梳成柔软的盘发,带有湿润的碎发,左侧佩戴一个白色つまみ細工(Tsumami-zaiku)花朵饰品。白皙的皮肤上可见毛孔和淡淡的汗水光泽。米白色浴衣,带有淡蓝色花卉图案和奶油色腰带。身份、眼型、发饰和浴衣在每个镜头中必须保持完全一致。
[风格]
超写实手持手机 Vlog 镜头,除特殊说明外,均为前置摄像头自拍视角。真实的皮肤质感、自然的织物物理效果、暖色灯笼光与冷色摊位光的混合光影、阴影处的高 ISO 噪点、轻微的运动模糊,呈现业余手机拍摄的质感。快节奏的 Vlog 式剪辑。拒绝 AI 式的平滑感。无背景音乐。
[镜头]
镜头 1:灯笼照亮的庆典街道上的自拍特写,HANA 对着镜头微笑,用韩语说道:{夏日庆典来啦!}
镜头 2:章鱼烧摊位前的自拍,蒸汽升腾,她将纸托盘举向镜头。旁边的男性摊主用日语说道:{给你,趁热吃!}
镜头 3:极近距离特写,低角度自拍,她咬下一串木签上的糖霜团子,脸颊鼓起,无对白。背景是红色灯笼的虚化光斑。
镜头 4:背面镜头 —— 她的浴衣后背和腰带充满画面,她指向一个明亮的奖品摊位,手臂带有轻微的运动模糊。
镜头 5:在奖品摊位前,自拍时将白色斑点海豹幼崽玩偶贴在脸颊旁,露出小小的自豪微笑,无对白。
镜头 6:前景下方是她的后脑勺,发饰聚焦,巨大的烟花在人群上方的夜空中绽放,烟雾缭绕。
镜头 7:低角度自拍,她看向镜头外上方,烟花的光芒闪烁在她脸的一侧,嘴唇微张,流露出安静的惊叹。
镜头 8:灯笼街头的自拍,她对着镜头近处挥动手掌,温暖地用韩语说道:{谢谢你们陪我一起。下次再见。拜拜!}
[保持一致性]
每个镜头中只能有一个 HANA。相同的面孔,相同的
Comments & Questions (0)
No comments yet
Be the first to comment!