Culture

YouTube’s blissful dance music

“Lo-fi house,” a fuzzy, highly danceable subgenre of house, owes its rise to machine learning.
Culture

YouTube’s blissful dance music

“Lo-fi house,” a fuzzy, highly danceable subgenre of house, owes its rise to machine learning.

If you’ve ever gone down a YouTube rabbit hole of dance music, it’s possible you’ve encountered the track “Talk to Me You’ll Understand” by the British producer Ross From Friends. It’s likely that, after the song finished, the next track you heard was “Winona” by DJ Boring. With these as entry points, you may have then found yourself in a trance of sorts, as fuzzy, distantly familiar melodies ushered you through hours of bizarrely named producers’ discographies.

This is no accident. YouTube’s recommendation engine pulls from the platform’s scale to serve up tracks from even the most niche subcultures. The coterie of producers making what’s been dubbed “lo-fi house” by its fans and creators is benefiting from Google’s uncanny ability to predict what you’ll want to hear next. It’s an example of the type of digital discovery that is coming to redefine the machinations of the music industry. Rory Fresco, the rapper whose track autoplayed after a Kanye West release, gained thousands of fans thanks to SoundCloud’s algorithm. There’s an entire economy of up-and-coming hip-hop producers appealing to search terms on YouTube, pumping out beats that sound similar to popular rappers (usually affixed with the descriptor “type beats”) to gain customers. Lo-fi house doesn’t owe its popularity to being algorithmically associated with music’s biggest stars, though. In fact, it’s quite the opposite. YouTube’s recommendations are so smart — and account for subtleties in ways that other platforms still don’t — that it effectively became a radio station for underground labels to scoop up fans.

The lo-fi house sound wasn’t invented on YouTube, but the platform allowed it to flourish. Beginning around 2012, aficionados of fuzzy ’80s house music congregated on online forums and in a Facebook group called Strictly Lo-Fi to share their creations. One of the most common ways of sharing tracks was by uploading them to YouTube, where, over time, the community would algorithmically grow. YouTube, of course, remains the most popular music platform for young people.

YouTube’s recommendations are so smart that it effectively became a radio station for underground labels to scoop up fans.

Nostalgia is at the center of the lo-fi house sound. The genre is a smattering of musical sensibilities from the past: Think of the elastic energy of early acid house, merged with the rhythmic foundations of R&B. What separates these new offerings from their vintage counterparts is a distinctly modern, maybe even space-y, tinge. The popular Australian producer Mall Grab, whose name is derived from a pejorative for a poseur in skateboarding communities, flipped a sample of Aaliyah’s “Are You That Somebody?” for his track “Let U Kno.” He dubbed the late R&B singer’s vocals over a conventional four-on-the-floor drum pattern and some well-placed hi-hats, and then put the whole thing through a time machine of filters. The result is a nostalgic hiss that explains YouTubers’ propensity for uploading these cuts alongside grainy VHS footage. It’s a sound that, in the past year, has become a force on dance floors around the world. Labels like the Vancouver-based 1080p and the U.K.’s Lobster Theremin experienced a boom in popularity in underground circles after releasing a number of well-received lo-fi house offerings. Mall Grab’s airy track “Feel U,” released on 1080p, has nearly 1 million plays on YouTube — no small feat for an underground artist. Mainstream acts have flocked to the sound as well. Early on in Drake’s More Life playlist, listeners are transported to a smoky club where the legendary Detroit house producer Moodymann is urging another round of drinks. “Passionfruit” and “Get It Together,” the standout tracks from from the club-ready section of project, lean on existing dance sounds; “Get It Together” is built around a sample from South African house mainstay Black Coffee. But they feel right at home among the dreamy dance tunes of Mall Grab and DJ Boring.

According to a paper released last year by Google, YouTube’s Google Brain-powered recommendations are made up of two neural networks. The first one, called “candidate generation,” takes into account a user’s watch history and uses it to select a range of videos that they might be interested in. Rather than organizing videos based on their rating or other feedback directly generated by users, the system uses implicit details about viewing habits. Data such as the amount of time spent watching a video and a video’s click-through rate are analyzed to create a sampling from which to serve recommendations. For more niche genres like lo-fi house, this works perfectly. Once I interact with just enough of a small community’s uploads, YouTube knows what other videos I’d likely want to spend time with. Thanks to the sheer breadth of uploads to the platform — hundreds of hours of video are added every minute — producers like DJ Seinfeld and Mall Grab can bubble to the forefront of a dance music listener’s recommendations. And, thanks to the infectiousness and accessibility of the genre, they can soon break into more mainstream recommendations.

The visual nature of YouTube certainly helped the genre expand, too. The vintage rave clips that dominate the most widely played lo-fi house videos could serve as mood boards for trendy teens everywhere. This season’s most coveted clothing item, the tracksuit, owes a great deal of its appeal to British ravers from the ’80s, footage of whom has been co-opted by young, nostalgic house producers of today. Across mainstream culture, an obsession with club culture’s heyday — which happened to coincide with a period of similar global turmoil — is evident in everything from pop music (Kanye’s Larry Heard-sampling “Fade”) to film (the newly released sequel to the cult classic Trainspotting, which leans heavily on British rave culture). During New York Fashion Week this year, Alexander Wang threw a full-fledged rave to celebrate a new collection.

The phenomenon of lo-fi house — bolstered by YouTube’s machine learning — can be interpreted as a nostalgia for a time that one didn’t live. While certainly more tied to the underground than genres like chillwave and dubstep, which rose to prominence as distinctly “millennial” phenomena to be co-opted by corporate entities, lo-fi house is on a similar trajectory to ubiquity. There is no pretense to the simple 4/4 drums that much of it relies on or in the kitschy names of its producers. Lo-fi house, like the culture it idolizes, is welcoming — disarming, even. In our culture’s obsession with the past, the young producers behind the increasingly omnipresent sound have created something entirely right now.

A previous version of this story described the label 1080p as being Australian. It is based in Vancouver.

Culture

The playlist is the new album

Thanks to new listening habits, artists are assuming the roles of creator and curator.
Read More