• Ben's Bites
  • Posts
  • Google's new AI brings sound to silent videos.

Google's new AI brings sound to silent videos.

Google Deepmind just dropped some hot new research on how they're making those silent AI-generated videos actually watchable. Think it’s time for AI video to move on from awkward mime acts.

What's going on here?

Google Deepmind's video-to-audio (V2A) tech can create rich, synchronized audio for AI-generated videos using just the video pixels and a text prompt for guidance.

What does this mean?

This V2A is like giving AI videos their voice (literally). It can generate stuff like:

  • Dramatic scores and sound effects that match the on-screen action

  • Realistic ambient noise to set the scene

  • Dialogue (still a bit wonky on the lip-syncing, but they're working on it)

You can even mix it with Google Deepmind's souped-up Veo video generator for the full AI-powered cinematic experience. Or use it to resurrect vintage silent films with immersive audio.

The best part? You can generate endless soundtrack options for a single video and use text prompts to nudge it in the direction you want. Handy for perfecting that chase scene score.

Why should I care?

Because the future of AI entertainment is looking pretty darn immersive. V2A is a major leap toward ultra-realistic AI-generated movies, shows, games - you name it.

Google’s not releasing this tech into the wild just yet. It's still got some kinks to iron out and rigorous safety checks to pass (to avoid another fiasco). Let’s ope we can hear the popcorn pop soon!

Join the conversation

or to participate.