According to The Verge, the method developed by Google researchers can also convert some sounds into musical pieces performed by several instruments.
Experts say that this method is similar to other systems that rely on artificial intelligence, such as “DALL-E”, which produces images by relying on written texts.
Thanks to this method, the user will be able, for example, to provide a text describing jazz music, and then the artificial intelligence system will produce an audio and musical piece, according to the specifications that were provided.
This new system developed by “Google” researchers is called “Music LM”, and the company “Alphabet” has published models obtained in the innovative way.
People who viewed the clips said they were “amazing”, some of them are 30 seconds long, and some are five minutes long, but this feature is still not available to all users.
This method can mimic human voices, with great accuracy, so that they sound exactly the same.
The system, based on artificial intelligence, was trained with a huge music database containing 280,000 hours of music.
Leave a Reply