Will AI write music better than humans?

Turn on your favourite song, crank up the volume and watch as your toes start to tap and your head begins to bob. The power of a good tune can transform us from being slumped at an office desk to an unstoppable karaoke star belting out a chorus line about how we can take on the world.

The ability of music to change our mood is powerful and many of us will have recovered from a break-up using Toni Braxton's Un-Break My Heart or found ourselves headbanging to the chorus in Queen's Bohemian Rhapsody.

So what is it that makes a piece of music successful and can we predict the next hit knowing the components of the songs of the past?

New research out this week has attempted to answer these questions by using mathematics to predict the key elements in chart-topping hits.

Using computer algorithms, the researchers examined different components of successful and unsuccessful songs to come up with the recipe for success in what makes a hit tune.

Published in the journal Royal Society Open Science, the study looked at musical hits and pop flops released over the past 30 years on the Official Singles Chart Top 100 in the UK.

Analysing half a million songs, the scientists looked into many factors, including a song's danceability, genre, rhythm, key and the gender of the lead vocalist. They then compared this data to the songs peak chart position and the number of weeks it spent on the charts.

The results found that the perfect ingredients for a chart-topping song include being upbeat, danceable and sung by a female vocalist in a major key with a tempo around 120bpm.

Unsurprisingly, the research also found that having a celebrity artist and a song that ties to a blockbuster movie significantly helps too.

The data also uncovered that our musical tastes have changed over the years as we transitioned from Wham! goodies like Wake Me Up Before You Go-Go to Adele's more powerful Hello ballad.

As well as shifting from brighter, more upbeat-style songs to slower and sadder ballads, our favourite tunes have also lengthened, moving from an average of two minutes and 30 seconds to a new norm of four minutes and 30 seconds.

As the music industry progresses, so do the skills of what a successful musician needs. Most people would assume that the ability to create a top charting music hit was determined by songwriting abilities, singing talent and marketing budget.

However, in the age of technology - mathematics and computer science may be the future skills needed to create the next hit.

New software can now empower computers to write new songs based on the style of the initial songs that are fed into it.

The very listenable song called Daddy's Car was written using an artificial intelligence-powered piece of software called FlowComposer. After being fed a series of Beatles' songs, the software analysed the music's key components and went on to write a whole new melody and accompanying chords based on what it had heard.

A human then added some lyrics to the tune and created a new song that could easily be mistaken for a top Beatles hit.

Some say that the world will always need humans to be artistic and creative, however, this move by artificial intelligence into music writing is starting to question that thought.

Who knows, perhaps the chart-topping hits of the future will be made by computers that understand the art of music through reading academic research papers about what songs humans prefer to listen to.