One way to track the evolution of popular music is to examine its subgenres. Think of how “rock” begat “punk rock,” which begat “post-punk,” as a simple example. Electronic and ambient music include an even bigger universe of subgenres, with hyperspecific names like “UK bass,” “chillwave,” and “electroacoustic.”
But what happens when a genre emerges not because of its artistry, but because of its discoverability?
This is the place “YouTubecore” finds itself in. YouTube famously hinges on an algorithm that guesses viewers’ interests to keep them clicking and viewing, and we’ve seen how weirdly that algorithm can go, both in innocent and diabolical ways.
In the case of music, however, YouTubecore has emerged in ways we never saw from MTV, radio, or other traditional platforms: as an explosive response to average computer and smartphone users wanting chill, ambient music. Through this, the new-age trend of the ’80s has made a surprising return, fueled by Generation Z’s musical interests and some Silicon Valley code, and those combined forces are unearthing ethereal surprises from the past and present.
Traits and early examples
The concept of YouTubecore is admittedly open-ended in terms of genre and style, but for our purposes, we can limit it to soft, instrumental fare—specifically, an algorithm-driven hierarchy of ambient albums that leans, for one reason or another, to the island nation of Japan. The YT uploads in question tend to include complete albums as opposed to individual songs, and some of the most popular examples were uploaded by anonymous users, not the original artists, often decades after their original releases. And none of the albums previously enjoyed particular commercial success.
Some consider Midori Takada’s forgotten 1983 album Through the Looking Glass to be one of the first YouTubecore albums. Uploaded in 2013, the original video has since been delisted, but it did go on to accrue millions of views—which was followed by Takada playing a set of worldwide tour dates, including her first in the United States. Other albums by different artists followed suit, many from the same 1980s Japanese ambient scene.
The most famous upload of them all (not ambient, but too known to not mention) came in 2017, when a video of the 1984 city pop song “Plastic Love” by Mariya Takeuchi became mind-bogglingly popular. Once a Japanese bargain-bin staple, people started buying it for $60 a pop in the United States. It has 45 million views today, along with an Olympic swimming pool’s worth of fan art, vaporwave remixes, and memes.
From YouTube to the hotel lobby
Benjamin Wynn, who performs under the name Deru, is an LA-based composer and television sound designer known in part for his work on Nickelodeon’s Avatar: The Last Airbender. His ambient work 1979, named after the year of his birth, has gathered almost 4 million views since the account Tape Counter uploaded it in 2015, one year after the album’s original release. The video strips away much of the album’s context, as 1979 is a mixed-media project with peripheral content including a collaborative photo album, an invented philosophy, and a limited run of pico projectors (created with the assistance of Robert Crespo, who made circuit boards for Mars rovers) containing visuals for each song.
Wynn’s label owner first noticed the uncanny YouTube popularity of 1979, which was soon followed by YouTube revenue payouts for each video play. Typically, YouTube’s Content ID system identifies and tags copyrighted material, then redirects view-based revenue to performers instead of faceless uploaders. But YouTube is a different revenue beast than services like Spotify, primarily because it pays per complete play; in Wynn’s case, a play of 1979 is 44 minutes long.
Wynn watched the video comments skyrocket into the thousands. Then he and his wife were vacationing in Tokyo when he heard 1979 play on hotel-lobby speakers—without any Japanese promotional efforts that he knew of. And while YouTube revenue for the video hasn’t been huge, its exposure has had one noticeable effect: physical sales. The 1979 vinyl edition is now on its fourth pressing.
Wynn has never had contact with the uploader. “At one point I was thinking, ‘I should just give my next record to this person!'” Wynn says. “But they have a lot of uploads that didn’t take off, so clearly this isn’t a 1:1 correspondence.”
“My only complaint is that it feels utterly random,” Wynn continues. “I can’t bank on the algorithm associating my name with this video; I’ve put out videos since then that haven’t received the same attention.”
Research on trends like “Hair Dryer Sound”
Without official answers from YouTube parent company Alphabet, musicians and fans alike are left guessing how its algorithm has driven this subgenre’s millions of views.
“Maybe [YouTube] scrapes through the actual sound waves, and it finds [and suggests] something similar?” record reissuer Yoskue Kitazawa says, calling to mind sound-analysis services like Shazam. “YouTube does have an auto-caption function, and it might do the same thing with audio.”
Massimo Airoldi, a professor at Emlyon Business School, co-authored a 2016 paper titled Follow the algorithm: An exploratory investigation of music on YouTube. It proposes that the algorithm partially leans on sequential viewing: if a significant number of users watch video B after video A, the two are considered related and therefore recommended. Within this framework, genres stop being simple technical distinctions and become granular concepts based on crowdsourced human-behavior patterns. Utilizing neural networks, the study estimates that viewing habits cause the algorithm to connect videos via recommendations, thereby knitting tight genre cliques in the process.
Seven out of 50 video clusters the researchers identified are deemed “situational” music. This designation doesn’t operate under the standard concept of genres but rather the context in which the music takes place. This includes relaxation music like “Ambient/Chillout,” “Sounds of Nature,” and the ASMR-affiliated “Hair Dryer Sound.” The paper concludes that situational music, sometimes deemed trivial by musicologists, is growing in popularity. They also found a cluster of “Ethiopia/South Sudan Music,” suggesting the context of a local scene comparable to ’80s Japanese ambient music.
This prediction was, of course, correct, with the rise of ambient YouTubecore being fueled by twin elements: “[The music] can be seen in both ways, either as relaxing instrumental backgrounds or as high-art examples of some avant-garde scene,” Airoldi says.
Watch time is also mentioned in Airoldi’s research, which makes sense as YouTubecore’s album-length videos typically exceed 40 minutes.
Setting the stage with GeoCities searches, vinyl translations
In the years before YouTubecore, Western DJs and bloggers set the stage for it to come into the mainstream. Musician Spencer Doran released an influential Japanese ambient mix in 2010 called Fairlights, Mallets and Bamboo. Online mixes in general remain popular to this day: since I began researching for this article, a video titled “Japanese jazz while driving on a warm night” has been popping up in my recommendations relentlessly; it’s up to 1.2 million views as of press time.
Since 2014, Jen Monroe’s blog Listen To This has brought Japanese music to English-speaking audiences, often with an emphasis on out-of-print music. Before the YouTubecore movement took off, her work required jumping through serious hoops: “Cold emails to strangers begging for records I suspect they have, sending PayPal payments to Japan for CDs hoping that they ever show up, [and] clawing through pop-up ads on Google-translated content scraper sites and ancient Blogspot posts.”
Diego Olivas followed in Monroe’s footsteps with his blog Fond/Sound and connected YouTube channel. He discovered music through old GeoCities websites and ordered vinyl from Japan. Then, as a way to expose this data to the English-speaking Internet, he took pictures of those albums’ liner notes, ran them through OCR (optical character recognition) software, and copied the text into Google Translate. As YouTubecore arose, labels sent him takedown notices. Some Discogs record slingers posed as label owners and sent fake takedown notices to manufacture scarcity.
Both Monroe and Olivas tell me that quite a few blogs like theirs are written in Japanese.
How much authenticity drives the algorithm?
Leyland James Kirby has made music as the Caretaker since the late ’90s, employing a trademark sound created from distorted waltz records. Driven by the concept of memory, his initial work focused on the ballroom scene in The Shining before moving on to memory conditions—specifically anterograde amnesia and dementia.
A 2011 upload of his album An Empty Bliss Beyond This World by user alteredzones currently has 3.6 million views. Kirby’s own 2019 upload for Everywhere At the End of Time, his six-hour album portraying dementia, currently has 5.2 million views and 95,000 comments. Videos about that album also recently blew up on TikTok.
Kirby has never promoted his work save for giving the occasional interview. “When I saw videos of my work getting millions of listeners, I thought to myself that something must be happening, as I knew I hadn’t paid for views or gamed the system,” he says. He attributes it to the quality: it’s “based on the sound contents and ideas contained within the work,” Kirby says. “For the algorithm to pick this kind of work up, it already needs existing engagement from an audience.” Based on the data he’s seen, 12 percent of the video’s recent views have come from the algorithm, while over 50 percent have come from direct searches.
Wherever the views come from (Kirby’s work certainly appears relentlessly in my YouTube sidebar), Kirby is careful to make room for at least some authenticity driving listeners to his music: “I think it’s genuine in the sense nothing has been bought,” he says. “It’s a straight success.”