What’s your favourite kind of music? Rock, Pop, Hip-hop, Classical or Jazz? Are your musical preferences based on the performer, or the genre? Now, blurring the lines further, researchers have developed an Artificial Intelligence system that asks itself — how would Mozart have played Metallica, or how would Chopin have performed a song by The Chainsmokers?
UK-based researchers have been trying to unravel the answer with the help of machine learning and analyse the style of different musicians and mix and match them.
Sign up for your weekly dose of what's up in emerging technology.
Reportedly he system began as a smaller project to determine stylistic differences between Irish traditional flute players.
Islah Ali-MacLachlan, senior lecturer in sound engineering at Birmingham City University, England said:
“Each genre of music has its own nuances, so it’s important from a musicology point of view to extract the correct information for further classification,” he said. “With flute this is particularly difficult because a player can change notes without introducing a new attack, unlike piano or guitar. We have concentrated on algorithms that accurately detect note onsets and note timbre.”
Once they figured out how to unravel the mystery of the Irish flute, they fed the AI system with over 15,000 notes and sounds, each of which was tagged to a style of a particular musician or composer.
Now, the AI system uses a neural network to replicate notes with an 86% level of accuracy, and imitate nearly 75% of all individual note deviations.
But this is not the first time someone has tried to unravel the mystery of music and the effect it has on humans.
Dutch neuroscientist Jacob Jolij had earlier tried to figure out and analyse the “feel-good” factor of a song. He had also come up with a mathematical formula that describes the anatomy of such songs, to understand what made them warm and fuzzy.