Google’s Artificial Intelligence is able to create music based on text descriptions
Google’s model, trained using a database loaded with over 280,000 hours of music, is capable of combining genres and instruments and writing lyrics around abstract concepts. Google’s Artificial Intelligence
Google published a study on an artificial intelligence model, called MusicLM, which, as the name implies, was developed to create music based on a text description. This is not the first tool of its kind, but, of all, it is the one with the most capacity to create tracks with depth, of different genres.
“We introduce MusicLM, a model generating high-fidelity music from text descriptions such as ‘a calming violin melody backed by a distorted guitar riff’,” the research paper read. “We demonstrate that MusicLM can be conditioned on both text and a melody in that it can transform whistled and hummed melodies according to the style described in a text caption,”
it added.
Trained with the help of a database with more than 280,000 hours of music, the model is able to combine genres and instruments and write lyrics around abstract concepts that, in general, remain an obstacle to the artificial understanding of these models. The complexity is attested when we realize that this same tool is capable of creating melodies based on a short hum, a whistle or the description of a painting.
Despite all the advantages, MusicLM has its problems. Some of the compositions sound bad and the vocalists’ voices can sometimes be incomprehensible. Some outputs stand out positively, but there is still room for improvement.
The technology will not be made available to the public, at least for now, because Google reveals some copyright concerns. During the experiments that preceded the launch of the study, the North American giant detected that 1% of the songs created by the AI were copies of the songs included in the database that trained it. Google’s Artificial Intelligence
Artificial intelligence already has a foot in music. Artists like Holly Herndon and Arca have used AI tools to add something to their creations, but all of these examples are collaborative or conceptual projects, heavily controlled by the human hand. MusicLM isn’t ready to be a solo artist, but it does give us a window into what the future of the music industry might hold.
