Well, we’re all in the mood for a melody,” Billy Joel crooned in Piano Man, his iconic 1973 bar-room ballad. That may have been true enough when Joel wore a younger man’s clothes, but a new study conducted by computational musicologists at Queen Mary University of London, UK, has found that vocal melodies in popular music have become much less complex over time.
The study, published recently in Scientific Reports, used mathematical models and algorithms to pinpoint three “melodic revolutions” — in 1975, 1996 and 2000 — that brought increasing simplicity to the two main components of melody: rhythm, or the pattern of sounds and silences in a piece of music, and pitch, the measure of how high or low the notes are.
The study looked at the top five Billboard songs every year from 1950 to 2023. Both rhythm and pitch became steadily less complex, the study found. “Conservatively, they have both decreased by 30 per cent,” said Madeline Hamilton, a graduate student at Queen Mary University who led the research.
The simplification of melody is not a new worry for some musicologists. “The place of melody as one of music’s basic building blocks is being reduced substantially,” the composer Yuval Shrem wrote in a 2014 article for Keyboard magazine, which was cited in the study. But the new research adds rigorous quantitative evidence of that trend.
Hamilton and her adviser, Marcus Pearce, the leader of the music cognition lab at Queen Mary University, found that other aspects of popular music, such as the number of notes played per second, actually increased over time, suggesting that the loss of melodic complexity amounted to a kind of trade-off. The shift may have been due to technological advances.
“Today, with the accessibility of digital music production software and libraries of millions of samples and loops, anyone with a laptop and an Internet connection can create any sound they can imagine,” the researchers wrote.
They cautioned against making value judgments about melodic loss, as the trend can easily fall victim to culture war narratives about classical versus contemporary music, they said in an interview. “It’s not that music is getting less complex,” Hamilton said. “The melody is getting less complex, but maybe the chords are getting more complex, or maybe the production.”
As Hamilton began researching melodic aesthetics in 2019, she encountered a shortfall that would dominate her career for the next few years. “I noticed that all the available datasets were classical and folk music, which I thought was kind of strange,” she said. “That’s not really representative of what people listen to every day.”
Hamilton decided to compile a dataset of her own using the Billboard Melodic Music Dataset, which includes 1,131 files of melodies from the top five Billboard songs for each year from 1950 to 2022. She listened to all 366 songs in the database and transcribed the vocal melodies — about three per song — into MuseScore, an online music notation program.
By that time, London was under the coronavirus lockdown and Hamilton was spending much of her time alone in a dormitory, listening to melodies for up to 10 hours each day.
Hamilton measured eight melodic metrics, four related to rhythm and four related to pitch, for each melody of each Billboard song in the data set. These included, for example, the number of notes per bar and the average melodic interval between consecutive pitches.
Roger Dean, a molecular biologist, musicologist and composer at Western Sydney University in Australia, offered crucial help, as well as what Hamilton called “sanity checks”.
Hamilton also used a statistical model developed by Pearce to measure how predictable each melody was in terms of both rhythm and pitch. “The model ‘guesses’ which note will occur next in the melody based on the previous notes in the melody,” Hamilton said. “Afterwards, it gives a value that represents how ‘surprised’ the model was on average throughout the melody.”
Then, Hamilton used algorithms that linguists have used to study shifts in language use to reveal the timing of significant moments in the evolution of pop. She found that melodic complexity dropped off sharply in 1975, around the time disco and stadium rock took hold. There followed a somewhat less steep drop in 1996, tracking with the growing appeal of hip-hop and electronic music, along with the popularity of MTV. Another significant melodic cliff loomed in 2000, most likely a product of the same forces that were at work throughout the 1990s.
Digital culture, including social media, may have also more broadly accustomed people to smaller units of information. “As our brains are getting used to reading and writing broken sentences in order to fit the various digital character limits,
our brains end up not expecting, requiring or even craving the fully formed sentence, musically or otherwise,” Shrem wrote in his 2014 article for Keyboard.
But others say the shift away from melodic variety makes sense because the human mind can handle only so much complexity. As music became more innovative in some ways thanks to the proliferation of digital tools and cultural shifts, it had to sacrifice creative nuance elsewhere, said Patrick Savage, a musicologist at the University of Auckland in New Zealand.
NYTNS