遊戲與娛樂

表情動作關鍵影格

根據上傳的參考圖,以一張靜態圖呈現角色單個表情動作的六個階段

更新日期:

完整提示詞

英文提示詞包含建議的範例需求。生成前請調整主體和細節。原始範例輸入及生成參數尚未完全核實。

推薦模型: GPT Image 2

Upload one character image as the reference for this single generation. Produce one still sheet of six keyframes of the single action named in the user request, in three columns and two rows, with no borders, numbers, or captions. If the reference is a photograph, redraw it as a simple sticker; if it is already an illustration, stay in that style. Keep the face structure, hair or fur, clothing, accessories, and main colors recognizable. Frame one is the start, frame two the wind-up, frame three the opening, frame four the peak, frame five the settle, and frame six nearly the start again. Change the mouth, cheeks, or brows, and let the body follow. Do not draw six unrelated emotions. Lock camera, light, scale, and body anchor. Animals keep their real anatomy. Nothing may cross into the next cell. Request a transparent background, and keep interior whites such as teeth and a sticker edge. Stop at the still sheet. Do not build a GIF, run code, or describe playback. User request: Use my uploaded character reference and show one delighted laugh across the six frames, from a closed smile to a wide open laugh and back.

提示詞技巧

  • 確認六格呈現同一個動作的連續階段;作品是靜態圖片
  • 提示詞要求透明背景;下載後檢查檔案是否具有真正的透明通道

來源與署名

由 OmniArt 改編及重寫,以下列出來源署名、修改說明及使用條件

  • Six-frame expression performance GIF · AdrianPunk115

    wangrunlin/awesome-gpt-image-2-5-prompts

    Repository MIT covers tooling and original documentation only; community prompts and third-party media retain their original rights and terms (CONTENT_POLICY.md).

    Rewritten as an original OmniArt still keyframe sheet. The source GIF export and code-execution steps were removed. The output is six static frames of one action. The prompt asks for a transparent ground, which must be verified after download and is not guaranteed. Localized into 19 locales. No source wording or media copied.

來源與署名

相關的提示詞配方