[ad_1]

Generative AI is making great strides in various aspects of content creation. Now, tech giant Meta has published generative AI models for speech-related tasks. The company has unveiled Voicebox, a super tool that can help with audio editing, sampling, and staging. Voicebox is a type of technology that can assist content creators with a range of tasks, help the visually impaired hear written messages, and enable people to speak any foreign language.

The company claimed to have made a breakthrough in generative speech AI. “We developed Voicebox, the first model that can generalize to speech generation tasks and hasn’t been specifically trained to accomplish with cutting-edge performance,” the company writes on its blog.

Voicebox creates outputs in a variety of styles and can create them from scratch. While the usual generative AI models produce images from text prompts, Voicebox produces higher quality audio clips. Currently, the model can process speech in six languages ​​and perform tasks such as noise removal, content editing, diverse sample generation, and pattern transformation.

Meta has also said that its multipurpose AI models like Voicebox can provide natural voices to virtual assistants and NPCs in the metaverse. The sample comes with contextual text-to-speech syntax that allows Voicebox to match the voice pattern to generate text-to-speech from an audio sample as short as two seconds.

The model can recreate a section of speech that was interrupted by noise or replace mispronounced words without having to re-record the speech again. Voicebox can produce speech from text in French, Spanish, English, German, Polish, and Portuguese from a sample of a person’s voice. This capability is known as pattern transfer across languages. “This capability could be used in the future to help people communicate in a more natural and authentic way even if they don’t speak the same languages.”

Besides, the tool with its various speech samples can generate speech that reflects how people speak in the real world.



[ad_2]

Leave a Reply

Your email address will not be published. Required fields are marked *