MMAudio V2

MMAudio V2

MMAudio V2 generates audio for existing footage using the video’s visual content and a written sound description. The footage provides action and timing cues, while the text defines the desired environment, material sounds, and acoustic texture.

It produces both background ambience and distinct sound effects. Timing-aware generation connects audio events to visible movement, supporting sounds such as footsteps or impacts alongside environmental layers. This makes it useful for giving silent footage an audible sense of place and action.

What it's good for

Product demonstrations and food close-ups can gain sounds associated with pouring, slicing, or sizzling. Outdoor scenes can receive weather and nature ambience, while mechanical footage can be paired with material sounds that reflect what is happening on screen.

Editors can also explore sound-design treatments before the final mix. Trying different textures against the same footage helps establish the scene’s mood and the emphasis placed on individual actions.