Vidu Q3 Reference-to-Video: Core Parameter Breakdown

Use Vidu Q3's Reference-to-Video pipeline to make a short "parameter breakdown" film: take an info-card-style reference image as the visual anchor, pair it with the Image1-indexed reference method in the prompt, and demonstrate how a reference image locks the subject's appearance and visual style, while the camera language carries the pacing across 16 seconds. In terms of composition, the piece makes the pipeline's three controllable dimensions explicit: reference_image_urls (1–7 reference images, here a single key-visual card), duration (1–16 seconds; this example uses a longer duration to accommodate multiple shots), and size (720P/1080P tiers with five aspect ratios; this example uses 16:9 landscape). audio stays off by default, yielding clean footage with no audio track, which makes it easy to add your own voiceover or narration in post. It is suited to developers and content teams who want to quickly judge what the "reference image + text shot description" input paradigm can produce: whether subject consistency is stable, whether camera movement is predictable, and how to trade off duration and aspect ratio. It also works well as sample material for parameter-comparison content about the site's video models.