The 'Vidu Q2: Reference to Video' workflow is designed to generate videos that maintain character and style consistency across sequences, making it ideal for projects requiring cohesive visual storytelling. This workflow leverages the Vidu Q2 and Vidu models, which are specifically tuned for handling multiple subjects while preserving their distinct features and styles throughout the video. By using the Vidu2ReferenceVideoNode, this workflow ensures that up to three subjects can be referenced and consistently represented in the output video, allowing for cinematic camera movements that add a professional touch to the final product.

Technically, the workflow begins with the LoadImage node, which allows users to input reference images for the subjects they wish to include in the video. These images are processed by the Vidu2ReferenceVideoNode, which applies the styles and characteristics from the reference images to the subjects in the video. The SaveVideo node then compiles these frames into a coherent video sequence. This workflow is particularly useful for creators looking to produce videos with a high degree of visual continuity, such as in animated storytelling or character-driven content.