Listen to These Odd Pop Songs Composed by a Computer
SONY Computer Science Labs has developed AI FlowComposer to write music.
Maybe you saw the short movie a few months back made from the bizarre script written by an AI program. Well, now here comes AI music courtesy of SONY Computer Science Labs’ (CSL) FlowComposer, and it’s every bit as…convincing. While FlowComposer does string together notes, it has no idea what combinations sound good to the human ear. So on a catchiness scale from 1 to 10, its songs rate about a, well, 0.
FlowComposer had its official debut in Paris, France on October 7, 2016 with a talk, “The Future of Music Making with the Flow Machines,” by François Pachet, director of Sony CSL Paris. Here’s the future-y gif on the event’s web page:
In the description for the talk, FlowComposer is described as a device that boosts creativity through what they call “reflexive interactions,” which “are human-machine interactions with a system that attempts to imitate the user’s style.” Basically, after analyzing a body of work by a human songwriter, FlowComposer attempts to generate new music based on the characteristic traits of the person’s work.
CSL has released two fully realized examples of FlowComposer songs. However, since what FlowComposer generates is just a lead sheet — a melody and basic chords — it’s a little hard to say what came from FlowComposer and what came from human musician Benoît Carré, who “arranged and produced the songs, and wrote the lyrics,” according to CSL.
This one, “Daddy’s Car,” is based on The Beatles’ music.
Here’s The Ballad of Mr Shadow” (sic), somehow based on Irving Berlin, Duke Ellington, George Gershwin and Cole Porter all at once.
Speaking of The Beatles, FlowComposer can reharmonize songs, too, though the results are equally daffy. Here’s what it produced for the Beatles’ “Yesterday.” (Here are the original chords.) FlowComposer re-harmonized the song “in the style of the Beatles,” which you would reasonably expect to sound a lot like the Beatles actual version. Not so much. (You may need to wait a moment for the songs to load in your browser.) There are also examples of “Yesterday” via “jazzman Cole Porter” and Michel Legrand.
(This post’s main image is “Yesterday” in iTunes, by the way.)
FlowComposer exists, presumably still in private testing, as a web app that Carré demonstrates here.
With a lot of settings to adjust to your tastes, it seems obvious — speaking as a musician here — that the process takes a lot longer than simply inventing a melody and finding its chords all by yourself. And what FlowComposer generates doesn’t sound nearly as good.
Amusing side note: As I watched the video above, a cat wandered across a keyboard of mine producing weird melodies that fit seamlessly with FlowComposer’s onscreen ditty.
I get why programmers enjoy figuring out how to create art with AI just to see if they can pull it off. It does feel mildly insulting that they assume human creativity — one of our most magically wonderful human gifts — can be reduced to an algorithm, but maybe it can. Someday. Not today.
Swipe right to make the connections that could change your career.
Swipe right. Match. Meet over coffee or set up a call.
No, we aren't talking about Tinder. Introducing Shapr, a free app that helps people with synergistic professional goals and skill sets easily meet and collaborate.
A completely unexpected discovery beneath the ice.
- Scientists find remains of a tardigrade and crustaceans in a deep, frozen Antarctic lake.
- The creatures' origin is unknown, and further study is ongoing.
- Biology speaks up about Antarctica's history.
Jeff Bezos, the founder of Amazon.com, explains his plan for success.
- Jeff Bezos had a clear vision for Amazon.com from the start.
- He was inspired by a statistic he learned while working at a hedge fund: In the '90s, web usage was growing at 2,300% a year.
- Bezos explains why books, in particular, make for a perfect item to sell on the internet.
It's one factor that can help explain the religiosity gap.
- Sociologists have long observed a gap between the religiosity of men and women.
- A recent study used data from several national surveys to compare religiosity, risk-taking preferences and demographic information among more than 20,000 American adolescents.
- The results suggest that risk-taking preferences might partly explain the gender differences in religiosity.
SMARTER FASTER trademarks owned by The Big Think, Inc. All rights reserved.