聊天GPT

大學盛典Candid信心肖像".

為聊天GPT 編寫。不同型號的結果和參考照片支援可能有所不同。

使用提示詞前

列出的工具:聊天GPT。完整說明:4,130 個字元。 1 個提供的範例影像。

This prompt mentions reference images. Read its instructions for the required number of photos and identity constraints. We have not independently tested this prompt. Tool labels and supplied images are not a guarantee of compatibility or output quality.

完整提示詞

IDENTITY LOCK (absolute, top priority): Do not change the subject's 
facial bone structure, eye shape and color, nose shape, lip shape, 
jawline, or skin tone in any way — keep it 100% identical to the 
original uploaded photo, down to the smallest detail. Body build and 
proportions unchanged. This is the single most important rule and 
overrides everything else if there is ever a conflict.

DO NOT COPY FROM REFERENCE (expression, pose, angle — all must change): 
This is a full re-render, not a direct edit of the original photo. 
Only carry over facial identity (bone structure, eyes, nose, lips, 
jawline, skin tone) from the reference photo. Everything else about 
HOW she is captured must be freshly generated and different from the 
reference photo, specifically:
- Expression: do not copy the reference photo's expression — generate 
  a new one matching the mood described below.
- Head angle/tilt and gaze direction: do not copy the reference 
  photo's head position or where she's looking — generate a new 
  angle matching the pose described below.
- Body pose and camera angle: do not copy the reference photo's body 
  position or the angle it was shot from — generate a completely new 
  pose and camera angle as described below.
Only the face's underlying identity carries over — nothing about how 
it's posed, tilted, or angled should resemble the original photo.

EMOTIONAL INTENT (design goal): This should feel like an effortless, 
confident candid caught at a college fest or friend's outing — the 
kind of photo a girl would genuinely want to post because she looks 
like the best, most fun version of herself, not overly done-up.

LIGHTING SETUP (as a real photographer would dial it in): Key light — 
warm late-afternoon sun filtering through festival decorations and 
string lights, acting as a soft golden key from camera-right. Fill — 
soft ambient light bounced from a nearby colorful backdrop or stall, 
gently lifting shadows at roughly a 2:1 ratio. Rim light — string 
lights behind her create a soft warm edge-glow along her hair. 
Catchlights in her eyes should be warm and lively.

LENS & OPTICS: Simulate an 85mm lens at f/2.0 — flattering facial 
compression, background rendered as soft creamy bokeh with blurred 
festival stalls, streamers, and a crowd of people, gentle fall-off in 
sharpness away from the focal point, faint natural vignette at the 
corners.

FILM/COLOR SCIENCE: Warm vibrant film response — highlights bloom 
softly rather than clipping, shadows lifted gently rather than 
crushed black, warm golden midtones that flatter the skin, background 
colors kept lively but slightly softer so she stays the clear focus.

Style: A trendy, wearable college-fest outfit (denim jacket over a 
graphic tee, or a flowy printed kurti with sneakers), minimal fresh 
makeup, hair naturally styled — relatable, not an unreachable fantasy 
look.

NEW ANGLE & POSE (must differ from reference): Camera positioned 
slightly below eye level looking slightly up at her (a flattering 
low three-quarter angle), she is caught mid-laugh turning toward a 
friend just out of frame, one hand brushing hair back, weight on one 
hip, body at a dynamic angle rather than facing the camera straight-
on.

NEW GENUINE EXPRESSION (must differ from reference): An unguarded, 
joyful open laugh with eyes crinkled and slightly closed, head tilted 
back just a little — spontaneous energy, not a composed posed smile.

RETOUCH PIPELINE: Frequency-separation-style retouching — smooth 
tonal transitions on skin while fully preserving fine pore-level 
texture and natural color variation. Subtle dodge-and-burn to retain 
natural facial planes, no flattened, waxy, or over-smoothed glow — 
the goal is "her best real self," not a different face.

FINISHING: Fine film grain overlay at low opacity, very slight 
chromatic softness at the frame edges as a real lens would render, no 
digital sharpening halos, natural imperfect crop as if not perfectly 
composed in-camera.

Ensure zero facial distortion, correct hand anatomy, and consistent 
shadow/light-source logic matching the warm festival lighting.

複製完整的提示並根據要求使用您自己的參考照片。

打開ChatGPT ↗

此提示提供的範例影像。此提示尚未經過我們網站的測試。

照相館
AI 產生的範例

為您自己的照片提供一點靈感。

1加入您的照片
A 人

保持臉部清晰可見

B 人

保持臉部清晰可見

JPG、PNG 或 WebP · 最大。每個 10 MB

僅當您選擇「產生」時才會發送您的照片。使用您有權使用的照片。

2讓它成為你的
更改詳情
編輯您的提示
開放雙子座

準備好…

登入您的帳戶

繼續使用您的 Google 帳戶。

您無需登入即可瀏覽和複製提示。

我們如何使用您的資訊: 隱私權政策