Vidu Q4 combines 15 image and 3 voice references in one video

What is changing?

ShengShu Technology’s Vidu Q4 is now available with new video-generation options built around multiple references. Reference-to-Video can use 1–15 images for characters, scenes and style, plus up to three voice samples intended to preserve a character’s voice and lip synchronization. According to Vidu’s official product page, the model can generate multi-shot stories with smooth camera movement at 2K or 4K resolution for up to 16 seconds.

Image-to-Video turns a single image into a 3–16 second clip while preserving the source aspect ratio. The company claims fewer deformations and smoother motion in running, combat and action sequences, alongside more natural smoke, particle, explosion and energy effects. These are provider claims, not an independent quality comparison.

Vidu Q4 is positioned for short-form series, advertising, social content and cinematic production. The model is offered through Vidu’s web product and API connection. Its official page contains example outputs, but no redistribution permission for that media could be verified, so the assets were not copied into this report.