Reference-to-Video

This page contains description of Reference-to-Video.

What is Reference-to-Video?

Reference-to-Video mode allows you to generate entirely new video content by referencing up to four images. Whether it’s a character, prop, or scene, the model will extract features from your references and generate a video that maintains consistency while introducing new creative scenarios.

This mode is perfect for creators who want to see their characters or objects in new environments, action sequences, or different camera angles without starting from scratch.

How to Use?

1

Select the "Reference-to-Video" Mode: From the video mode options click on Add image which will open the modes tray, choose Reference-to-Video to start the process.

2

Upload Reference Images: Provide one to four reference images of your subject(s), such as characters, props, or scenes. These images will serve as the basis for generating the video.

3

Describe the Scenario: Write a prompt describing what you want to happen in the video. For example:

• "The character walking through a futuristic city at night, neon lights reflecting on wet streets."

• "A woman in a red dress dancing in a ballroom."

4

Set Parameters: Choose video duration, resolution, and output format according to your preferences.

5

Generate the Video: Click Generate and allow the model to create your video, maintaining subject consistency while introducing new dynamic elements.

---

Input Media Supported

Reference-to-Video supports various input media to help guide video generation and editing, ensuring flexibility in your creative process.

Images:

• Up to 4 images can be uploaded for use in Reference-to-Video or Image-to-Video.

• Resolution: Minimum of 300px, max file size of 10MB.

• Formats Supported: JPG, JPEG, PNG.

Videos:

• 1 video can be uploaded for use in Reference Video.

• Duration: 3 to 10 seconds.

• Max File Size: 200MB.

• Max Resolution: 2K.

Elements:

• Multiple images (up to 4) can be uploaded or generated from different angles to form an element, providing more reference information for the model.

• Note:

  • When a video is present, you can upload up to 4 images/elements combined.

  • Without a video, you can upload up to 7 images/elements.

---

Key Capabilities:

• Multi-Reference Subject Creation: Build new subjects from one or more reference images, ensuring consistency in appearance, clothing, and accessories.

• Subject Consistency: Maintain the same identity for characters, props, and scenes across the entire video.

• Creative Flexibility: Generate dynamic new scenarios, actions, and environments while preserving the reference subject’s integrity.