[ad_1]
Microsoft and Google are very active in the field of artificial intelligence and have made a lot of noise these days; But other companies such as Meta also develop products and services based on artificial intelligence.
Meta company by publishing an article in On his blog, he announced that he was working on a generative AI tool for speech; The Voicebox tool can handle a wide range of speech production tasks such as editing, sampling, and styling; Tasks that Voicebox is not necessarily thematically trained to perform.
Meta describes Voicebox’s capabilities as follows:
In-context conversion of text to audio: By receiving a 2-second audio sample, it learns the tone and style of the voice and converts the text into audio with the same style.
Conversation editing and noise reduction: It can reconstruct a part of an audio conversation that is disturbed by noise or correct mispronounced words without having to re-record.
Transferring conversational style from one language to another language: It can take a sample of a conversation or part of a text and turn it into an audio file in English, French, German, Spanish, Polish or Portuguese.
Various speech sampling: It uses a variety of data to generate audio to more accurately reproduce the conversation type of people in the 6 indicated languages.
Meta believes that in the future, multipurpose AI models like Voicebox will be able to give natural voices to digital assistants and NPC characters in the Metaverse. Using this tool, people who are visually impaired or blind can hear their friends’ text messages as audio, Meta says, and content creators can create or edit audio clips more easily than ever before.
[ad_2]
Source link
