LTU Computer Science student Joe George has been studying computational methods for analyzing music. With assistance from LTU faculty, he digitally processed songs and albums of some well-known popular music bands, and applied novel computational tools to study the machine perception of their music, quantifying the differences between their albums. By doing that he was able to make a dramatic scientific breakthrough, demonstrating that computers can profile the progression of the artistic styles of popular music bands. For instance, his algorithm was able to sort all of the Beatles albums in their chronological order by just processing the music data, and without using any information other than the sound itself.
The work published in the scientific journal “Pattern Recognition Letters” – one of the world’s most prestigious machine learning journal. It was also featured by numerous articles published in the world’s premier popular press such as NBC News, Discovery, Scientific American, Science, Yahoo News, Bloomberg Business Week, and many more, including daily newspaper and TV stations all over the world, from Trinidad to Kazakhstan.