Here’s Meta’s new AI model – making realistic voices and sounds

earthpressnews
Heres Metas new AI model making realistic voices and

Meta, the company behind Facebook and Instagram, among others, has invested heavily in AI over the past year. Among other things, they have developed the language model Llama 2 and the image creator Imagine with Meta.

Now Meta has developed and launched another AI tool – Audiobox. Audiobox is a kind of unified platform where the user can, among other things, imitate human voices and create their own sound effects with the help of AI.

pic.twitter.com/sCkCjgHSvG

Read more: Microsoft’s big change – for the first time in 30 years

You can do this with Meta’s Audiobox

In total, there are six functions within Audiobox: “Your Voice”, “Described Voices”, “Modified Voices”, “Sound Effects”, “Magic Eraser” and “Sound Fill”. Here’s how they work:

  • With “Your Voice” feature you can make the AI ​​mimic your own voice and say anything. Just upload a short audio file where you speak, and Audiobox will mimic your voice as it reads out the text you enter.
  • “Described Voices” is also a text-to-speech function, but here you instead describe in your own words the voice you want the AI ​​to have.
  • The third function, “Altered Voices”, combines the features above. Here you can upload an audio file where you speak – and also describe how you want Audiobox to change it.
  • “Sound Effects” allows you to generate sound effects based on a text description.
  • With “Magic Eraser”function, you can delete unwanted sounds from a bad audio recording, such as noise and background noise.
  • The last function, “Sound Fill”allows you to add sound effects to an audio file by providing simple text instructions.

    • Here you can test the Metas Audiobox and all its functions.

    Read more: OpenAI to New York Times after lawsuit: ‘Manipulated’

    Read more: Microsoft’s new AI for mobile – it can do that

    Read more: This is how you can make money with AI – the new feature



    nh2-general