Meta's VoiceBox AI Generative Audio Model is INSANE! #shorts #voicebox

Announcing a breakthrough in generative AI for speech, Meta's new Voicebox AI generative audio model. Voicebox by Meta is a state-of-the-art AI model that can perform tasks such as:

Synthesize speech in the same style as an audio sample: Voicebox can learn the audio style from a sample as short as two seconds and use it to generate speech from text.

Edit speech and remove noise: Voicebox can fix speech that is noisy or has errors without having to record it again. For example, you can cut out a part of a speech that has a dog barking in the background and ask Voicebox to fill in the gap – like an audio undo button.

Transfer style across languages: Voicebox can read text in any of six languages (English, French, German, Spanish, Polish or Portuguese) in the same style as someone’s speech, even if the speech and the text are in different languages. This could help people talk to each other naturally and authentically even if they don’t share a language.

Sample speech from diverse data: Voicebox can create speech that reflects how people speak in the real world and in the six languages mentioned above.

#meta #ai #voicebox