Skip to content

Draw an Audio

Synthesize synchronized audio from video with multiple instructions

Visit official site

What Draw an Audio is

The technique is described as a video-to-audio synthesis method using multi-instruction control.

What you can do with it

  • Generate synchronized audio from video content
  • Mix audio across multiple stages
  • Follow text, video and signal instructions together

Application scenarios

  • Add underwater sounds to a diving scene
  • Build engine sounds for a driving scene
  • Create train and cheering crowd effects

Pros & cons

✓ What we like

  • Text, video and signal instructions can be combined
  • Several audio generation demos are provided

! What to watch out for

  • Audio output depends on the instruction modules
  • The code is listed as coming soon

FAQ

What is the base model?

A latent diffusion model is listed.

Which modules handle instructions?

A text-conditioned model, a masked-attention module and a time-loudness module are named.

Last reviewed: 2026-11-15

More AI video generation tools

View all →

How we review