聊天GPT

電影風格 — 傾盆大雨

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:3,866 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

हाँ—यहाँ पहली uploaded image को FACE/IDENTITY REFERENCE और दूसरी image को TARGET POSE + CLOTHING + BACKGROUND REFERENCE मानकर एक साफ़, मजबूत prompt है:

FINAL FACE-REPLACEMENT PROMPT

Create a hyper-realistic, cinematic wide-shot photograph of two adult men locked in an intense martial-arts fight inside an open, muddy arena during a heavy downpour. Raindrops, water spray, and mud splashes should be dynamically visible in the air. The central fighter wears a sleek, traditional all-black martial-arts uniform, while the defending opponent directly in front of him wears a contrasting traditional all-white martial-arts uniform. A realistic crowd of spectators surrounds the muddy combat area in the background.

🔒 PRIMARY IDENTITY REFERENCE — FIRST UPLOADED IMAGE

Use the FIRST uploaded image exclusively as the facial identity reference for the person whose face is being transferred. Preserve his identity with maximum possible fidelity.

Transfer and reproduce the exact facial identity and natural appearance from the first uploaded image onto the head of the target fighter. Preserve the person's recognizable facial characteristics, including:

exact face shape and proportions

forehead and cheek structure

jawline and chin

eye shape, size, spacing, and natural expression

eyebrow shape

nose shape and proportions

lip shape and mouth proportions

natural facial asymmetry

beard/mustache characteristics, if present

hairline and naturally recognizable facial details

authentic skin tone and realistic skin texture


Do not create a similar-looking person. Do not reinterpret the face. The result must remain unmistakably the same person shown in the FIRST uploaded image.

🎯 SECOND IMAGE — TARGET REFERENCE

Use the SECOND uploaded image strictly as the TARGET composition reference.

Strictly preserve the target image's exact pose, body position, head placement/orientation, camera perspective, framing, clothing, wardrobe structure, environment, background arrangement, lighting direction, and overall composition.

Do not copy the second image's facial identity. Replace only the target person's face/identity with the identity from the first uploaded image while keeping the target head position and body pose intact.

The transferred face must naturally conform to the target head angle and perspective, with realistic anatomical integration. Maintain correct facial proportions, natural neck connection, realistic lighting, shadows, wetness, rain interaction, and skin reflections.

🎬 REALISM

Ultra-photorealistic high-end commercial martial-arts photography, cinematic realism, physically accurate anatomy, realistic wet skin, soaked fabric, individual hair strands, detailed mud and water droplets, natural skin pores, realistic shadows and reflections, dramatic rainy atmosphere, sharp fabric texture, authentic photographic depth, subtle cinematic film grain, 4K/8K-level detail.

🚫 NEGATIVE PROMPT

Do not change the identity from the FIRST uploaded image. No look-alike face, no generic face, no facial redesign, no beautification, no face morphing, no plastic skin, no excessive smoothing, no artificial symmetry, no CGI, no cartoon/anime style, no altered facial proportions, no different hairstyle unless required by the target clothing/scene, no distorted head, no warped jaw, no malformed eyes, no incorrect nose or lips, no extra limbs, no extra fingers, no deformed hands, no anatomy errors, no pose changes, no clothing changes, no background changes, no unnecessary cropping, no watermark, no text, no artificial face-swap appearance.

HIGHEST PRIORITY:
FIRST IMAGE = exact facial identity.
SECOND IMAGE = exact pose, body, clothing, framing, and background.
Do not mix the two references incorrectly. Preserve the first person's identity while reproducing the second image's target composition as accurately as possible.

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