Posts
 |
| Photo: James Vincent |
"Artificial intelligence has become very good at imitating human composers" insist James Vincent, London reporter for The Verge.
 |
| Photo: |
Artificial intelligence can imitate the works of Bach so well that you (probably) can’t tell the difference. Don’t believe me? Try it for yourself.
The computer-generated music in these samples is the work of DeepBach — a deep learning-powered program
created by Gaetan Hadjeres and François Pachet of the Sony Computer
Science Laboratories in Paris. This is the institution that previously
gave us the “world’s first” AI pop song
(though that had quite a bit of help from humans) and has used its AI
music software FlowMachines to mimic musicians ranging from Mozart to
the Beatles.
DeepBach: harmonization in the style of Bach generated using deep learning
Now, though, FlowMachines has conquered Bach — and its success says a lot about the current capabilities (and limitations) of current AI.
Deep learning, the method used by FlowMachines and tech
companies like Google and Apple, relies on sifting through large amounts
of structured data to identify (and later synthesize) patterns. When it
comes to applying these methods to music, Bach is a perfect fit. Not
only was he incredibly prolific, but his music is consistently
structured, following various patterns and formal intricacies that
defined Baroque music.
For DeepBach, Hadjeres and Pachet concentrated on Bach’s chorales — pieces of music that set traditional hymns to stately, four-part vocal melodies. Bach wrote 389 chorales in his lifetime, giving DeepBach plenty of material to study, and each chorale has recurring patterns.
For DeepBach, Hadjeres and Pachet concentrated on Bach’s chorales — pieces of music that set traditional hymns to stately, four-part vocal melodies. Bach wrote 389 chorales in his lifetime, giving DeepBach plenty of material to study, and each chorale has recurring patterns.
