Runway 画像から動画生成: 画像から動画生成
通过添加运动效果或场景连续性，将任意图像制作成视频。RunwayML 的 I2V 模型通过推断深度、运动和时间动态，将静态视觉内容转换为短视频片段。
通过添加运动效果或场景连续性，将任意图像制作成视频。RunwayML 的 I2V 模型通过推断深度、运动和时间动态，将静态视觉内容转换为短视频片段。
このモデルについて
Runway Image-to-Video 是一款前沿 AI 模型，能够通过添加复杂的运动效果和场景连续性，将静态图像转换为动态视频。该模型利用先进算法推断深度、运动和时间动态，根据单张视觉素材创建简短且引人入胜的视频片段。这种创新方法让用户能够探索视觉叙事的新维度，将艺术与技术无缝融合。
该模型基于 RunwayML 的专有技术构建，并针对效率和质量进行了高度优化。其易用的配置、详细的プロンプト选项以及可自定义的出力设置（アスペクト比、分辨率和时长），为初学者和专业人士都提供了强大的视频生成工具。无论用于营销视觉内容、数字艺术项目还是创意叙事，该模型都是不断发展的画像から動画生成领域中的高性能解决方案。
コスト分析
| プロバイダー | 費用 | 備考 |
|---|---|---|
| muapiapp | $0.15 per generation | muapiapp offers exceptional affordability, being 20-50% cheaper than key competitors while maintaining comparable or superior quality. |
| Fal.ai | $0.25 per generation | Fal.ai's pricing is nearly identical to Replicate, but muapiapp provides a more cost-effective solution at 20-50% lower costs. |
| Replicate | $0.26 per generation | Replicate offers similar capabilities at a slightly higher price point. In comparison, muapiapp is 20-50% more affordable without compromising on quality. |
muapiapp offers exceptional affordability, being 20-50% cheaper than key competitors while maintaining comparable or superior quality.
Fal.ai's pricing is nearly identical to Replicate, but muapiapp provides a more cost-effective solution at 20-50% lower costs.
Replicate offers similar capabilities at a slightly higher price point. In comparison, muapiapp is 20-50% more affordable without compromising on quality.
** 競合サービスの料金は類似のモデル構成および利用ティアに基づいて算出された推定値です。
設定スキーマ
| パラメータ | 型 | 説明 | デフォルト |
|---|---|---|---|
| プロンプト | string | 用于生成動画的プロンプト | The camera smoothly zooms in on the sleek, futuristic race car as it speeds through a neon-lit urban tunnel at twilight, its glossy white surface reflecting the vibrant pink and blue lights streaking past. The precise detailing of the car’s aerodynamic curves and glowing accents is highlighted as droplets of water spray from the spinning tires, adding a palpable sense of motion and intensity. The driver’s black helmet, contrasted against the car’s gleaming body, remains sharply in focus, emphasizing the thrilling high-speed chase through the city. The blurred cityscape and illuminated digital billboards in the background create a high-tech, cyberpunk atmosphere, intensifying the scene’s adrenaline and futuristic vibe. |
| 画像 URL | string | URL of the input image. Provides visual reference for AI | https://d3adwkbyhxyrtq.cloudfront.net/ai-images/186/809475541516/a895607c-b724-427c-9020-6c36ebc4af25.jpg |
| アスペクト比 | Enum（5個の選択肢） | 出力動画的アスペクト比。 | 16:9 |
| 解像度 | Enum（2個の選択肢） | The resolution of the generated video. If 1080p is selected, 8-second video cannot be generated. | 720p |
| 長さ（秒） | Enum（2個の選択肢） | The duration in seconds. If 8-second video is selected, 1080p resolution cannot be used. | 5 |
用于生成動画的プロンプト
The camera smoothly zooms in on the sleek, futuristic race car as it speeds through a neon-lit urban tunnel at twilight, its glossy white surface reflecting the vibrant pink and blue lights streaking past. The precise detailing of the car’s aerodynamic curves and glowing accents is highlighted as droplets of water spray from the spinning tires, adding a palpable sense of motion and intensity. The driver’s black helmet, contrasted against the car’s gleaming body, remains sharply in focus, emphasizing the thrilling high-speed chase through the city. The blurred cityscape and illuminated digital billboards in the background create a high-tech, cyberpunk atmosphere, intensifying the scene’s adrenaline and futuristic vibe.URL of the input image. Provides visual reference for AI
https://d3adwkbyhxyrtq.cloudfront.net/ai-images/186/809475541516/a895607c-b724-427c-9020-6c36ebc4af25.jpg出力動画的アスペクト比。
16:9The resolution of the generated video. If 1080p is selected, 8-second video cannot be generated.
720pThe duration in seconds. If 8-second video is selected, 1080p resolution cannot be used.
5開発者ドキュメント
準備入力：
16:9、1:1）和分辨率（720p 或 1080p）。5 秒或 8 秒）；请注意分辨率与时长之间的兼容性限制。生成视频：
結果の確認：
FAQ
主体清晰、视觉噪声较少的高质量图像通常能获得最佳结果。图像中的细节越丰富，运动推断的数据越充分，过渡效果也会越流畅。
可以。你可以自定义生成视频的アスペクト比、分辨率和时长。但请注意，由于处理限制，`1080p` 分辨率不能与 `8` 秒时长同时使用。
该模型每次生成的费用为 $0.15。费用按生成的视频片段计收，因此对个人创作者和企业来说都是经济实惠的方案。
目前支持生成 `5` 秒或 `8` 秒的视频。这两种时长在质量和处理效率之间取得了平衡。