LTX-2 姿態轉影片

透過姿態引導產生高品質影片,LTX-2 支援音訊與影片同步產生,姿態控制精準、動作富有表現力、執行高效。

Input

The person to animate, full body and facing the camera; it becomes the first frame.

A clip of someone moving; their body pose (DWPose) drives the character. Match the framing of the character image.

Describe the scene, the motion and any sound.

Number of frames to read from the pose video (8n+1, e.g. 97, 121); 121 frames at 24 fps is about 5 seconds.

Default: 121

1280×720 is the default; the model snaps to the nearest size it supports.

1280×720 is the default; the model snaps to the nearest size it supports.

Advanced settings

Frames per second; set it to the pose video's rate.

Default: 24

Don't use the character image as the first frame (the pose still follows the video).

What to avoid.

Leave empty for a new result each run. Enter a number to repeat one.

Output
00:00
Example

試試範例

關於此工作流程

LTX-2 姿勢轉影片工作流程,專為將姿勢引導轉換為高品質影片內容設計,並支援音訊與影片同步生成。採用 Lightricks 的 LTX-2 模型,提供精確的姿勢控制與豐富動作表現,適合創作動態影片內容。流程整合多個節點,如 GetVideoComponents、PreviewImage 與 SaveVideo,分別負責影片分解、影格預覽與最終輸出。核心特色為 comfyui_controlnet_aux 自訂節點,進一步提升姿勢控制與動作表現的精度。

技術上,流程從 LoadVideo、ResizeImageMaskNode 等節點進行影片與影像預處理,接著用 DWPreprocessor 微調影片設定,確保影格數與解析度最佳化。支援最高 1920x1088 解析度,預設為 720p 以提升效率。影格數需為 8 的倍數加 1,以確保流暢過渡與同步。此流程特別適合需要精細動作與音訊同步的創作者,善用 LTX-2 模型實現專業級成果。

問答