Intriguing program, just a heads up for anyone interested in this sort of thing...
The Secret Science of Pop
"Evolutionary biologist Professor Armand Leroi believes data science can transform the pop world. He gathers a team of scientists and researchers to analyse over 50 years of UK chart music. Can algorithms find the secret to pop success?
When the results are in, Armand teams up with hit producer Trevor Horn. Using machine-learning techniques, Armand and Trevor try to take a song by unsigned artist Nike Jemiyo and turn it into a potential chart-topper.
Armand also takes a scientific look at pop evolution. He hunts for the major revolutions in his historic chart data, looking for those artists who transformed the musical landscape. The outcomes are fascinating and surprising, though fans of the Fab Four may not be pleased with the results. As Armand puts it, the hallmark of The Beatles is 'average'.
Finally, by teaming up with BBC research and development, Armand finds out if his algorithms can discover the stars of the future. Can he predict which of thousands of demo tracks uploaded to BBC Introducing is most likely to be a hit without listening to a note?
This is a clash of science and culture and a unique experiment with no guarantee of success. How will the artists react to the scientist intruding on their turf? And will Armand succeed in finding a secret science of pop?"
~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~
60 min program, available on iPlayer for 17 days, I'll let you do the maths !
http://www.bbc.co.uk/iplayer/episode/b08gk664/the-secret-science-of-pop
Comments
Simplez
Twisted Imaginings - A Horror And Gore Themed Blog http://bit.ly/2DF1NYi
All that aside, could machine learning churn out a decent song in the future? I reckon so.
A truly terrible and disappointing program! Just because you are applying scientific methods doesn't mean you are being scientific.
Reminded me of a series on years ago with Marcus du Sautoy, spouting bollocks about "The Code."
For anyone really interested in the detail one of the original papers is here by the way (couple of years old):
http://rsos.royalsocietypublishing.org/content/2/5/150081