fbpx

Song-writing computers on the cards?

Rate this item
(0 votes)

Read  Mia Andric's comments here...

People who are asked what makes a song a hit might say it’s the fact that it’s catchy, or that it’s got great lyrics, or even that it’s sung by an attractive person. They would be wrong, according to science.

What hits like Abba’s “Knowing Me Knowing You”, James Taylor’s “Country Roads”, UB40’s “Red, Red Wine” and The Beatles’ “Ob-La-Di, Ob-La-Da”, and hundreds of other songs that have topped the Billboard charts have in common is the right combination of uncertainty and surprise, a new study has found. Researchers analysed 745 classic U.S. Billboard pop songs to better understand the process of musical enjoyment and discovered that they could narrow it down to one of two possibilities: Low uncertainty and high surprise, or the opposite, highly uncertain but not surprising.

Song writing computers on the cards

To rule out factors like memories associated with particular songs, the researchers focused only on the chords of the songs, which were stripped of other aspects of the original material, such as the lyrics and melody. The chord progressions remained the same but they were no longer recognisable as famous songs. They then used a machine learning model to mathematically quantify the level of uncertainty and surprise of the chord progressions relative to one another.

These were then played to around 80 human test subjects connected to functional magnetic resonance imaging (fMRI) brain scanners. The results were consistent across the entire group. When the test subjects were relatively certain about what chord to expect next, they found it pleasant when they were instead surprised. Conversely, when individuals were uncertain about what to expect next, they found it pleasurable when the subsequent chords weren't surprising.

The MRI results indicated that the interactive effect between the uncertainty of the upcoming chord and its level of surprise was associated with brain activity changes in emotion- and auditory-related areas. However, they found that the nucleus accumbens – a region that processes reward expectations and had been thought to play a role in musical pleasure – only reflected uncertainty, which differs from previous research findings.

This effectively means that the role of the nucleus accumbens in music-evoked emotions may be more nuanced than previously thought, the scientists said. Musical pleasure was reflected in the brain's amygdala, hippocampus and auditory cortex. These regions are associated with processing emotions, learning and memory, and processing sound, respectively.

The researchers are planning on investigating whether the findings will remain similar for other genres like jazz and for non-Western musical traditions such as those from China and Africa. Once that has been established, they believe that their results could be applied to assist composers or even computers in writing music. Algorithms could also be developed to predict musical trends and how well a song would do based on its structure.

And the uses of the research wouldn’t have to stop at music. Dancing and joke-telling are also linked to expectancy, so the study results could theoretically be used to extrapolate whether people would enjoy one dance move over another, or find a specific joke funnier.

With the development of improved artificial intelligence and machine learning technologies progressing swiftly, this study could well form the foundation of a new generation of computers dedicated to producing art. This was the one area humans have been confident of keeping in their hands in the face of ever-smarter robotics and the resulting automation of human occupations.