> ## Content Index
> Fetch the complete content index at: https://montagea.com/blog/llms.txt
> Use this file to discover other available public pages before exploring further.

# Wan 2.2 Input Video: When to Use Animate and How to Prepare It
- URL: https://montagea.com/blog/prepare-inputs-motion-transfer-ai/
- Published: 2026-10-24T14:22:37.000Z
- Updated: 2026-10-24T14:26:12.000Z
- Description: Choose the appropriate Wan 2.2 workflow before preparing your assets. This guide distinguishes Animate video inputs from ordinary image-to-video and highlights preprocessing checks.
- Author: Tim
- Tags: video production, ai filmmaking, character consistency

For a Wan 2.2 workflow driven by an existing performance video, start by checking whether you are using **Wan2.2-Animate**. The official repository describes Animate as taking a driving video and a character image. Ordinary image-to-video examples instead start from an image. Choosing the right workflow matters before you spend time converting files. See the [official Wan2.2 instructions](https://github.com/Wan-Video/Wan2.2?ref=montagea.com).

## Choose the workflow for your input

| Workflow       | Input distinction                   | Planning question                                            |
| -------------- | ----------------------------------- | ------------------------------------------------------------ |
| Wan2.2 I2V     | Starts from an image                | Do you want to animate a still?                              |
| Wan2.2 TI2V    | Supports text or image input        | Which input route does your selected setup use?              |
| Wan2.2-Animate | Uses a video plus a character image | Do you want the character to follow an existing performance? |

Animate provides animation and replacement modes. Animation transfers the driving performance to the character; replacement inserts the character into the video.

## Check the preprocessing outputs before generation

The [official preprocessing guide](https://github.com/Wan-Video/Wan2.2/blob/main/wan/modules/animate/preprocess/UserGuider.md?ref=montagea.com) distinguishes the driving-video path, character-image path and output directory. Animation produces face and pose videos. Replacement additionally produces background and mask videos. Inspect these intermediate files before rendering.

- For basic pose retargeting, the guide requires front-facing, stretched starting poses for the reference and driving characters.
- The supplied mask-extraction approach is intended for a single person; multi-person footage needs another solution.
- Treat resolution and frame-rate settings as workflow-specific choices. Follow the documentation for your actual implementation.

These notes describe the documented workflow, not a hands-on benchmark. They do not establish a universal requirement that every image and video have identical dimensions, or that one frame rate will prevent all artifacts.

## Keep a small, reviewable first experiment

As an editorial planning suggestion, choose a short performance you are entitled to use and record the intended action in one sentence. Save the workflow version, input filenames and settings with the output. Watch for changes to the face, lost gestures, unstable edges and unwanted background changes. Record what you can see before guessing at a cause.

For example, an acceptance note might read: “The character turns toward the doorway; the face remains recognizable; the background suits the next shot.” This is a suggested review criterion, not a promise of model performance. If the result is unsuitable, identify one visible issue and change one input or setting for the next trial.

## Connect the test to the finished scene

A usable experiment should serve the story. Keep the chosen take, rejected alternatives and revision notes together. When planning work in [Montagea](https://montagea.com/?ref=montagea.com), use the creative brief with the tools actually available there. This article does not establish a Wan2.2 integration in Montagea. For a related planning comparison, see our [image-to-video and text-to-video guide](https://montagea.com/blog/image-to-video-vs-text-to-video-product-ads/).