
Meta, one of the largest technology-based companies, has finally taken its first step in the Artificial Intelligence (AI) world. Meta has released their latest AI-based invention called Audiocraft. The release of Meta’s newest technology was announced on their official website on Wednesday, August 2nd.
Audiocraft allows its user to automatically convert a text into the audio version of it. Their latest technology consists of 3 different types of AI models, such as MusicGen, AudioGen, and EnCodec.
Meta also explained the work flow of these 3 different types of AI models. MusicGen will generate music from the text provided by the user. It is said that MusicGen have been trained for a total of 20.000 hours. AudioGen will focus more on making voice effects, to make the sound of the voice sounds as natural as possible. While EnCodec will decompress audio signals with high fidelity. (theshabazzcenter.org)
Through the release announcement of Meta’s latest AI invention, the company hopes that their invention can help people to convert text into audio easily just by typing. They hope people can do experiments practically without any instruments.
“That’s Audiocraft promise, our new AI tool that was made to make high quality audio and realistic voices from music and texts,” Meta stated in their website.
Read also : Meta Release Virtual Reality Service For USD 7.99 A Month
Now, Audiocraft already has several samples of music and voice effects posted on their website for people to listen or experience. All of the music and voice still sounds classic. The company also stated that they are open for voice or music volunteers to increase the diversity of their voice samples.
Voice and music generator technology is predicted to gain a massive amount of interest, spotlight, and users if a lot of people start to use this kind of technology.