Skip to content

Video

wan-3-reference

Direct a cast of images, motion and sound into one clip.

On the rail as wan-3-reference.

Building a clip up to thirty seconds long from references you assemble, a face, a movement, a voice, directed by naming each one in the shot you write.

It needs a written shot that names each reference, so images, motion or audio dropped in without a prompt will not produce a result.

  1. Use the card’s References area. Your cast: pictures for who is in it, clips for how they move, sounds for how they sound. Name each one in the prompt.
  2. Write what you want in Prompt.
  3. Set Duration on the card.
  4. Press Run.
Control What it does Choices
References Your cast: pictures for who is in it, clips for how they move, sounds for how they sound. Name each one in the prompt.
Prompt What you want, in your own words. The card sends it with your inputs. Your words
Duration Clip length in seconds. This is what drives the credit cost. 2 seconds, 5 seconds, 10 seconds, 15 seconds, 20 seconds, 30 seconds
Run Starts the run. The number on it is what this run costs in credits, worked out before you press.
×1 Runs the card several times in a row with the same settings. 1 to 99

These sit behind the settings button on the card.

Setting What it does Choices
Resolution Each step up doubles the per-second cost. 480p, 720p, 1080p
Aspect Ratio Adaptive lets the model pick the shape from your references. Adaptive, Horizontal 16:9, Horizontal 4:3, Square 1:1, Vertical 3:4, Vertical 9:16
Generate Audio Render synchronized audio with the video. Included at no extra cost. On or off
Seed The starting dice roll. The same seed and settings give the same result every time; change it for a fresh variation, or leave it at -1 for a new random one. A number