Facebook’s AI Research (FAIR) team’s music translation network has the ability to translate music from one style to another.
FAIR has utilised AI to translate an audible input into different styles, genres and instruments. First reported by TNW, the neural network has the ability to hear, for example, a jazz track, and is then able to translate that exact song into a different style of music, such as rock.
According to TNW, this AI breakthrough is the ‘first time an unsupervised learning method has been created for the purpose of creating high-fidelity music with a neural network’. Neural networks are comprised of computer software or hardware, modelled on the human brain and nervous system.
According to the FAIR research paper, “A Universal Music Translation Network”:
“Our results present abilities that are, as far as we know, unheard of. Asked to convert one musical instrument to another, our network is on par or slightly worse than professional musicians. Many times, people find it hard to tell which is the original audio file and which is the output of the conversion that mimics a completely different instrument.”
You can make your mind up for yourself by checking out the video produced as part of the findings below. We think it’s pretty interesting!
FAIR’s AI breakthrough is another step forward in AI music composition
With rapid advancements in AI technologies including machine learning, many areas in the music industry are being introduced to tech which is delivering results impossible to achieve through manual processes.
A few months back, SKYGGE released the first full-length album composed using artificial intelligence. In a more mainstream application, British synth-pop band Years & Years have experimented with AI creative technology to craft remixes of tracks on their upcoming second album.