Adam Wakefield assess how music has been changed by machine learning, algorithms and artificial intelligence (AI).
Machine learning recognises music so you don't have to
Music recognition apps such as Shazam and SoundHound have been built to provide listeners with song names, albums or artists they are otherwise unable to identify.
For example, if you hear a song on the radio and do not know the artist's name, song name or the lyrics, you can use Shazam or SoundHound to identify the song. Both platforms lean on machine learning to recognise music at the swipe of a thumb.
According to a blog post on data analytics and consulting company Principa’s website, 'by analysing millions of songs and extracting features that are characteristics of each song and then storing it in a database', each app is able to compare the characteristics of the song being listened to, and compare it to songs in its database until it finds a match.
“The more versions of the song it listens to (acoustic, instrumental, etc.) the better it becomes at identifying the song – to the point where it can even identify a song by someone simply humming it.”
Algorithms are shaping the music you listen to
Machine learning and algorithms help us identify songs, and these same technologies are shaping the music we listen to in our spare time. Spotify, which has recently launched in South Africa
, makes users go through what it calls a 'Taste Onboarding' process so it can identify users’ music tastes and present users with various playlists based on the data it receives.
Ben Ratliff, a music critic, writer, podcast host and published author, noted in a piece
for The Guardian
(February 2016) that the biggest impact made in music over the last 10 years was not by a single artist but, in fact, by listeners themselves.
“Access is power. A great deal of the history of western music, and much else besides, is now on your phones or on your desks, via YouTube or a streaming service – Spotify, Pandora, Apple Music, Tidal, whatever – for free, or for a modest price,” Ratliff says.
Spotify and its peers, says Ratliff, have altered listener’s relationship with music since no time in history have people had access to so much music as they do now.
“Listening to anything, especially when you haven’t heard it before, is a highly creative act; but a little less so, I think, when you let the computers do the choosing for you,” he adds.
AI and data is being used to increase profitability, popularity and music itself
Sergey Bludov, senior vice president of media and entertainment practise at technology consultancy DataArt, points to how record labels, producers and artists use machine learning to make more intelligent and profitable business decisions.
Whether it is finding ways to use data to cater offers, create hit music, drive collaborations and partnerships, schedule concert dates or simply sell more music, Bludov says data is changing how the music industry relates to fans and impacts merchandising, engagement, relationships and experiences.
Music is no longer just written by people
Stuart Dredge, a freelance journalist writing for The Guardian
, spoke to Siavash Mahdavi, CEO of AI Music.
AI Music is one of two firms that in August 2017 was taking part in the Abbey Road Red, a start-up incubator run by the studios that aim to forge links between new tech companies and the music industry.
Dredge says there are other companies in the field, and that they are trying to answer the question of "Can machines create music [by] using AI technologies like neural networks, to be trained upon a catalogue of human-made music, before producing their own?”
Ed Newton-Rex, CEO of JukeDeck, says
that a few years ago, AI was not at a stage where it could write a piece of music good enough for anyone. Now, it’s good enough to be used in some cases.
Want to stay up to date with the latest media news? Subscribe to our newsletter
Machine learning is having a major impact across different industries, with the news media no exception. Read more in our article, How machine learning is changing the newsroom.