AI Programs Give Composers a Little Help
In wired.com, Matt Jancer recently shared a secret about the writing of the music that may be on your sound system even now. Just as Autotune turns iffy singers into bearable for listening, now composers have been turning to AI to help them write the hits. So far, AI seems to have found its groove.
Jancer demonstrated how programs employed various styles and algorithms to create successful songs accepted by humans. Here are two of several AI music tools:
Created by Sony’s Computer Science Laboratories, Flow Machines analyzes tracks from around the world, then suggests scores that artists—including electropop musician ALB and jazz vocalist Camille Bertault—interpret into songs. For its debut album, Hello World, the AI also surveyed syllables and words from existing music to create original (albeit gibberish) vocals. Example: The Beatles-inspired “Daddy’s Car”
Originally developed to compose background tracks for user-generated videos; now it’s being adopted by K-pop stars like Kim Bo-hyung and Highteen. Using deep neural networks, the AI predicts note sequences to compose new songs. After users select parameters such as mood, genre and beats per minute, the AI creates a track that artists can use as a foundation. Example: Highteen’s ultra-processed hit “Digital Love”