Wan 5
  • 創作
  • AI圖像
    • 開啟圖像工作台
    • AI 修圖
    • GPT Image 2新上線
    • Nano Banana新上線
      • Nano Banana 2新上線
      • Nano Banana Pro
      • Nano Banana Flash
    • Seedream 5.0新上線
    • Flux 2
    • Qwen Image新上線
      • Qwen Image 2
      • Qwen Image 3新上線
    • Z-Image
  • AI影片
    • 開啟影片工作區
    • MiniMax H3新上線
  • 工具
    • 瀏覽全部工具
    • 圖片競品頁
      • SeaArt AI
      • Canva AI
      • Leonardo AI
      • Lovart AI
    • 影片競品頁
      • InVideo AI
      • VEED AI
      • Vidu AI
      • DeeVid AI
      • Sora AI
      • Qwen AI Video
    • 更多工具
      • 圖像轉 Prompt
      • 批次圖像轉 Prompt
      • 影片增強 AI
      • AI 圖生影片
  • 定價
繁體中文
  • English
  • 中文
  • 日本語
  • 한국어
  • Español
  • العربية
  • 繁體中文
Wan 5

Vidu AI Video Generator Workflow: Reference-Led Video Briefs

Structure a reference-led brief: name the subject, composition, environment, and identity cues before describing motion. Uses Google Veo 3.1 Fast.

Generation model: Google Veo 3.1 Fast.

Enter text, describe the content, movement... (e.g. A 3D boy skateboarding in the park.)

Generate video · 40 credits total

Reference-Led Video Briefs

Create video clips guided by a reference idea. Name the subject, composition, environment, and essential identity cues before describing motion. Begin with one readable action and one camera move.

  1. 01

    Define the reference elements

    Name the subject, composition, environment, and any essential identity cues that must remain stable.

  2. 02

    Add one motion

    Describe one readable action and one camera move. Keep the motion simple and focused.

  3. 03

    Review the output

    Check that the reference elements are preserved and the requested motion is clearly visible.

Character Introduction

A young woman in a red coat stands in a busy train station, she turns and looks toward the camera, natural daylight from the glass roof, commuters moving around her, cinematic style.

The subject, clothing, setting, and lighting remain stable; the turn and look are the only actions.

Reference elements are stable
The subject, composition, environment, and identity cues should remain recognizable after the motion.
Motion is readable
The requested action and camera move should be easy to identify without competing instructions.
The woman's red coat changes color during the clip. How do I fix it?

Add 'same red coat throughout, no color change' to the prompt.

The commuters are not moving. How do I add motion?

Specify 'commuters walking around her, busy station' in the prompt.

The turn and look is too fast. Can I slow it down?

Add 'slow turn and look toward camera' to the prompt.

資源
  • 創作
  • 定價
  • 圖像轉 Prompt
  • 批次圖像轉 Prompt
公司 & 法律資訊
  • 關於我們
  • 聯絡我們
  • 隱私政策
  • 服務條款
  • 退款政策
AI圖像
  • Nano Banana Pro
  • Nano Banana 2
  • GPT-4o
  • Seedream 4.5
  • Seedream 4.0
  • Flux 2 Pro
  • Flux 2
  • Qwen Image 2
  • Z-Image
  • NanoBanana Flash
  • Qwen Image 3
Wan 5

Wan 5|AI 圖像生成與編輯工作空間

[email protected]

本網站為獨立第三方平台,與本站所提及的任何 AI 模型提供商無關聯、未獲其背書,亦非其官方代表。所有商標及品牌名稱均歸各自所有者所有。

© 2026 Wan 5。保留所有權利。

隱私政策