Quantcast
Channel: Technology | The Atlantic
Viewing all articles
Browse latest Browse all 7173

Could a Robot Write the Perfect Pop Song?

$
0
0

Grab some blank sheet music. Choose a key (a major key, preferably). Set your tempo at something moderate, 4/4 time. Write a melody, a chorus, and a bridge, with each line getting eight bars. Add lyrics about a relatable topic—love? loss? loneliness?—and, voilà, you've written a pop song.

Okay, okay, a 32-bar form and karaoke-friendly lyrics do not a hit song make. But most do follow the same basic guidelines—major key, moderate tempo, broad topic—before they're produced and packaged to become the latest earworms.

Which begs the question: If pop songs can so easily be written and then distributed into an unbreakable cycle of hits, can't they also be reverse engineered and reproduced? Can't a songwriter feed a topic into a machine and have that machine regurgitate a melody and lyrics, forming a pop song that's packaged and ready to go?

Not if you want the song to find an audience, says John Covach, the director of the Institute for Popular Music at the University of Rochester. Pop music has always been more about cultural significance than musical inventiveness. The tunes that become hits today may not have five years ago. (The Beatles, he says, probably wouldn't have had the success they did if they released their music for the first time today. Our EDM- and autotune-trained ears probably wouldn't allow it.) Besides, there are so many songs under the pop umbrella. "Popular music might as well just be called 'music,'" Covach told me. "It's socially and culturally constructed."

In other words, because pop music mutates to match what's en vogue, no one really knows what will succeed. Chart-topping pop songs—ones that are dance-along-able, that get stuck in your head—will always need a little something extra, something catchy. (Think the current '80s-influenced sound of Mark Ronson and Bruno Mars' "Uptown Funk," Maroon 5's "Sugar," and Taylor Swift's "Style," or the "Soar" described by Daniel Barrow at The Quietus in 2011 that dominated songs like Katy Perry's "Firework.") Anyone can grab some blank sheet music and scrawl out the beginnings of a pop song, but no one can predict whether that song will resonate with enough listeners to make it a sensation.

That said, record companies and artists have tried to reverse-engineer songs, thinking that doing so would solve the problem. Susan Schmidt-Horning, the author of Chasing Sound and a professor of history at St. John's University, points out that people have reverse engineered specific sounds—rock and roll in particular. Musicians would study the distortions made by electric instruments, recycling elements into their songs and creating new ones.

And pop music does get, well, formulaic because of it. Musical copies appear again and again. Artists try to recreate the sound that made them pop phenoms. It's easy to find mashups of any combination of hits, given the right beat.

But tearing a song apart and building a similar one from its components has never produced consistent success. Instead, the world of pop music depends on the way we create and listen to music. The pop star of the future is not going to be a hybrid Katy Perry-Miley Cyrus-Carly Rae Jepsen robot that blends hooks and sounds together to write the ultimate catchy pop song. The pop star of the future is going to look something more like Hatsune Miku.

Miku is a Japanese pop artist, an avatar of sorts who can "sing" and "perform" on stage when she's projected onto a screen in 3-D. Created by (the aptly named) Crypton Future Media, she is a Vocaloid, a program that uses recorded human voices in a database to sing. And she's got die-hard fans—enough of them to make her a household name in Japan, an opening act for Lady Gaga, and a mind-boggling guest on Late Show With David Letterman:

She was also the subject of a New York magazine profile in November that breathlessly declared, "She is, depending on whom you ask, a harbinger of a radically collaborative future in pop music or a holographic horsewoman of the apocalypse."

It's easy to understand the latter sentiment. Miku doesn't physically exist, and if people are becoming rabid fans of an avatar, how soon do we all begin worshipping computer-generated, soulless pop culture entirely? (Not anytime soon.) How "live" is a live Hatsune Miku performance if she can perform at the same time in two places halfway around the world from one another? (As "live" as it can be.) Can an avatar have fans, even if she can't pose for photographs or dole out autographs? (Yes.) Can an avatar be a celebrity? (Of course.)

But as popular as Miku is, her success relies on human ability. As Lindsay Zoladz writes for New York:

Far more revolutionary is the fact that all her music—including the songs performed in concert—is written by fans, some of whom cannot read music and never felt empowered to write a song before Miku came along. "Miku is seen less as this really special person, like Lady Gaga or somebody," [Tara] Knight [a professor at UC San Diego working on a documentary about Miku] says, "but rather a conduit through which you can express yourself."

Miku, therefore, is the embodiment of human creativity. She sings pop songs written by fans, with innocuous titles like "Sharing the World" and "Packaged," and those same fans cheer each other on.

Which makes her basically like YouTube or SoundCloud, but with a face and a "personality." She captures how music production has been changing: Musicians can write, produce, and distribute their work on a variety of platforms to a wide audience. "Artists don't need a label as much for distribution anymore," says Greg Milner, the author of Perfecting Sound Forever. "They're evolving independent of that."

Of course, artists are not about to all become blue-haired avatars. Easier distribution just means a more saturated pop music landscape, which poses a new challenge: Musicians need to reverse-engineer popularity, not pop songs.

No one has figured that out, but streaming proves one part of the answer. There are myriad ways to listen to music today—on the radio, on our devices, on anything but (R.I.P.) CDs and tapes—but the method that gives the listener the most control is the Internet, through sites like Spotify, Pandora, and Rdio. Users input their music tastes into the sites, essentially teaching them what songs and artists are popular. They then are able to play music similar to or match those tastes. It's a type of reverse engineering that begins with understanding what a listener likes.

But it's not perfect. Pandora, for example, occasionally rotates the covers of a song a user already disliked to cycle into a playlist or suggests a song that barely matches a user's tastes. And on the artist's end, if an artist hopes to build an audience, they'll have to reach the ones that already like the type of music they're making, and they have to capture those listeners' attentions before they skip ahead to the next song. "Music is all around us because it's so flexible and so easy," Milner says. "There are just too many other distractions now, so our attention spans have lessened... I think we're going to be listening to music both a lot more and a lot less closely."

On top of that, the use of data to track what we like listening to can backfire. The more we listen to a song, the more ubiquitous it becomes, and the more studios want to copy that song's success. We like the familiar, so the same sounds resurface, lending pop music its reputation for being repetitive and generic.

At this rate, we'll be discovering new music either by seeking out artists and songs ourselves or by, well, having samples of new music gifted to us. All this has prompted three new models for how to release new music: the album leak (Madonna, Björk, countless others), the surprise album (Beyoncé, Drake, Kendrick Lamar), and the U2-Apple Songs of Innocence infiltration—an album that appeared on people's iPhones without them asking for it—that's unlikely to be replicated by another artist in the near future.

Altogether, these factors predict a bleak road ahead for pop music, but in reality, pop music is doing just fine. At least Covach thinks so. "No matter what happens," he says, "people like to hear a song with some kind of melody or something that's got a catchy quality. And, in order for the song to be popular, people have to have a sense that there's a real person behind the music."

It's easy to poke fun at the state of the pop music industry, but it wasn't too long ago when people were in awe of being able to hold thousands of songs in the palms of our hands. These days, it just takes some more effort to capture listeners' attention—and if that means an avatar with pigtails projected onto a screen, well, there are still more puzzling spectacles in pop music-dom than that.

This article was originally published at http://www.theatlantic.com/technology/archive/2015/05/the-pop-star-of-the-future/387757/










Viewing all articles
Browse latest Browse all 7173

Trending Articles