There is such a scene in the biographical movie creative story Here, Ewan Brenner channels Creation Records founder Alan McGee in a scene with a therapist, ranting about demands to find the next big thing in music. “I spend millions noise I don’t know if anyone will like it! ” Welcome to the record industry.
What makes a song a hit? nobody knows It is a mysterious organic process that no one has been able to explain.
As one record company president said: “Running a record label is risk-based. Final songs we offer hope the public will love it. A song can be objectively great, but if there’s no public interest, you can’t spend a lot of money on marketing and promotion. make they like it “
Even in this situation, people are still trying to figure out how to accurately predict hit songs.
When rock and roll was young, some promoters thought in their head that the process of writing hits could be reduced to a formulaic process. In 1959, Jo Malhall and Paul Neff sent her 3,000 girls a questionnaire about their musical likes and dislikes. Their idea was that if they could incorporate as many positive data points as possible into the song, it would guarantee a hit for their favorite pop star, a 15-year-old American weightlifter named Johnny Restivo. bottom. As a result of synthesizing all the answers, this song was completed.
This approach didn’t work. the shape i am in It only reached number 80 on the pop charts.
Many attempts have been made to find ways to predict hits, most of them looking for people with “golden ears”. It’s an amazing natural instinct that certain people have to elicit the success of things they didn’t know the public wanted. For example, in the early 1960s, an American indie president of his label began performing songs for his teenage daughter. She showed a real talent for predicting which would work, with success rates around her 80%. But it turned out to be beginner’s luck, and after about 20 attempts her prediction failed.
Meanwhile, the record and radio industries built their businesses around people with golden ears like Clive Davis (discoverer of Janis Joplin, Barry Manilow, Patti Smith, Whitney Houston, and more). . Mo Oistin (Fleetwood Mac, Prince, Red Hot Chili Peppers); Seymour Stein (Ramones, Talking Heads, Madonna); Rosalie Tromblay has grown from a receptionist at CKLW/Windsor (The Big 8) to someone with an uncanny ability to pick hits.She didn’t just convince Elton John to release her benny and the jets As a single, she took hits from Guess Who, Bob Seger, Kiss, and many more, despite all his reservations.
Some have taken a different approach. Weezer’s Rivers Cuomo uses a spreadsheet approach to songwriting, believing his next hit is hidden in the data. In 2003, Polyphonic HMI brought in Barcelona-based AI company Hit Song Science. The company used machine learning to analyze millions of data points collected from Billboard hits dating back to 1955. The company believed this would unlock the underlying audio foundations of music. In addition to explaining the popularity of popular songs, you can also use that information to create new hits.at the same time in May Predicting the Grammy-winning success of Norah Jones’ debut album, Come Away With Me (This is debatable), the certified U2 hits performed through the project were rejected as failing.
Hit Song Science isn’t the only company looking to harness the predictive power of AI. MusicXray, Bandmetrics, Mixcloud and a few other companies are also in this space. The potential payoff is huge. At least 100,000 new songs are uploaded to streaming music services every day, much of it junk. If someone could come up with an idea to improve the filtering algorithm to select only the best, everyone from record labels to radio stations to streaming his platforms would want to participate.
Perhaps this old data-driven approach is too limited. Welcome to the new field of “neuroforecasting” music.this is the real precog minority report Thing: Using the neural activity of a small number of people to predict future population influences and behavior.
According to a Neuroscience.com report, US researchers are using neural responses from living humans, or brain wave responses, to enhance AI machine learning. Research subjects were set up with off-the-shelf physiological sensors that collect brain activity related to mood and energy levels. Various statistical approaches were applied to the data, machine learning was introduced into the mix, and his AI was applied to the neural responses recorded when real humans listened to the songs.
The results were surprisingly good. Researchers claim 97% accuracy in predicting which songs will be hits. This is much higher than the 50 percent from other more traditional methods (actually, a coin toss). To be fair, the test involved only 33 people and their neural activity, and included 24 songs. But when the technology actually works as advertised, the 97% success rate is outstanding.
Will people with golden ears and intuition for music become unnecessary? I hope not. AI is great, but it can only mimic what you extract from your data and your prompts. Only humans (so far) can get excited about something new and different.
But as neuro-prediction becomes more sophisticated, it will find endless applications in product testing and focus groups. For companies and institutions that can afford to do so, of course. Music would be a great starting point. But where will it take us? I have no choice but to wait and find out.
—
Alan Cross is a broadcaster for Q107 and 102.1 the Edge and a commentator for Global News.
Subscribe now for Allan’s ongoing new musical history podcast on Apple Podcasts or Google Play
© 2023 Corus Radio, a division of Corus Entertainment Inc.
