How do you turn a computer into a music geek?
This is Sandra Tsing Loh with the Loh Down on Science.
Computer scientists at Lawrence Technological University, have developed a novel artificial-intelligence algorithm. The goal of artificial intelligence is to have computers do tasks that normally require human intelligence. This algorithm can characterize and compare patterns in different musical styles. Interestingly, the work is an extension of earlier research into vocal communications of whales.
The program first runs through an audio track and produces a spectrogram. That’s a visual image of the track’s sound frequencies. It then converts the spectrogram into a set of mathematical measures. These describe the music’s style, and compares them to measures of other audio tracks.
When they tested the program with a set of Beatles tracks, they found that it placed the group's music releases in chronological order! That is, it detected the evolution in style over the Beatles’ career!
The team hopes an artificial-intelligence approach can determine what makes musical styles distinctive.
So, roll over, Beethoven! Or we’ll get a robot to do it. Won’t have your great hair, though.
***** For more 90-SECOND SCIENCE FACTS, click here.*****
The Loh Down on Science is produced by LDOS Media Lab, in partnership with the University of California, Irvine, and 89.3 KPCC. And made possible by the generous support of the Gordon and Betty Moore Foundation.