Can a computer algorithm predict hit music?

Scientists at the University of Bristol believe they’ve cracked the science of what makes a hit with a 60% success rate.

They’ve come up with an equation that measures metrics such as tempo, timing, song duration and loudness along with more complex stuff such as chords and harmonies. The AI computer has listened to over 50 years of top 40 hits so far and what is even more interesting is that music tastes have changed over the years.

I’ve spoken before on how text to speech computers may replace voice overs and now it appears a computer could replace the judges on reality TV shows like The X Factor.

Dance was not a big factor in the 1980s with many ballads becoming hit tunes. In the 1990s hit music became simpler with many 4/4 time hits and today it seems that, in general, dance songs are more likely to become hits.

Music streaming services Spotify, Last.fm, Deezer and Rdio to name a few already make use of clever algorithms to predict the kind of music you’d like to hear based on what you’ve listened to in the past.