We taught computers how to sing 61 years ago, the very first recording of which now lives online. How you feel about IBM 7094’s fuzzy rendering of “Daisy Bell” will depend on your tolerance for voice synthesis. It either tickles the brain in all the right places or deeply confounds it. By definition, its voice is unnatural, the artificial product of programming. For some, it may sound unnerving and impersonal. But listen to the warm crackle of its tone, how it blithely calls out to the object of its affection—synthesised or not, that hulk of machinery and its silly little love song evoke real emotion.
Technology has now advanced to the point where software can sing anything and an algorithm can produce every possible melody. Vocal synthesisers sound more lifelike than ever, and Vocaloid pop stars are headlining shows and booking Coachella stages. Yet, the ghost of that lovestruck machine still lingers. You can hear it in the music of Yoasobi, a Japanese duo who adapt works of fiction into sprawling, MacBook Pro-made pop hits.
It always starts with a story, explains Ayase from Tokyo, where it’s already well into the evening. This is their magic hour, when dusk gives way to ingenuity and Yoasobi—which roughly translates to “nightlife”—come alive. Over Zoom, the 28-year-old composer and 22-year-old vocalist Ikuta are ruminant and composed, not quite as vivacious as their prolific oeuvre but no less introspective. A tinkerer by nature, Ayase approaches songwriting the way a scientist might conduct a lab study: with precision, control, and acute curiosity.
He pulls inspiration from a variety of written works like novellas, short stories, and even manga and fan submissions, immersing himself in imagined worlds until the characters begin to feel like his own creations. The process is amorphous and intuitive, he says. The melody takes shape as he reads, filling the margins of his mind with musical notes and flecks of colour. Ayase dissects its themes, picking apart the pieces and reconstructing them into music. He takes someone else’s story and interprets its meaning into an original work of his own.
“When I write these songs I reflect on my own feelings and experiences,” Ayase tells Teen Vogue via an interpreter. “Our music, although it is based on fiction, is not completely fiction, because if you just reflect the fictional story as it is, then it wouldn’t be interesting.”
Take their debut single, ‘夜に駆ける (Yoru ni Kakeru).’ Adapted from Mayo Hoshino’s melancholic short tale タナトスの誘惑 (Thanatos no Yūwaku), which translates to The Temptation of Thanatos, the song depicts a love between the narrator and a young woman who repeatedly attempts to take her own life. Ultimately, they both succumb to the allure of death, falling into the unknown together. It’s bleak. The music, however, is bright and energetic with drum kicks and piano interludes that propel the disjunct vocal melody forward at a relentless pace. Ikuta’s voice is both sweet and crystalline. This is the contrast that often lies beneath the surface of Yoasobi’s work—light and darkness, beauty and grotesque, it all goes hand in hand.
‘Yoru ni Kakeru’ topped the charts in 2020, resonating with a generation of people who were in the mood for a pensieve pop hit amid the pandemic. The song majorly broke through, bowing at No. 1 on Billboard Japan’s Hot 100 and accumulating more than 700 million streams to become the country’s most-streamed song of the last five years. It’s been a whirlwind ever since. Two EPs followed, The Book and The Book 2, which were re-recorded in English and released as E-Side and E-Side 2, the latter of which recently dropped. They held their first in-person concert at Tokyo’s legendary Nippon Budokan arena, and they’re set to embark on their first arena tour in Japan this spring. In 2021, they were crowned Artist of the Year at MTV’s VMAJs. They’ve also written and performed OST themes for popular anime, including Netflix’s Beastars and Mobile Suit Gundam: The Witch from Mercury. Their streaming numbers are now in the billions.
In just three years, Ayase and Ikuta have become permanent fixtures on Japan’s music charts while also making waves internationally. Despite their success, however, very little has changed in their process. The only difference, Ayase jokes, is that his to-do list keeps getting longer.
“Obviously, there’s a lot of attention on us,” Ayase says. “People think that now we’re spending a lot of money on our production. But I still feel that we have this very DIY spirit to us. It’s very hands-on and handmade.” He makes all of their songs on his MacBook Pro using Logic Pro, an audio production software that allows artists to create and arrange original compositions.
He cuts all of their demos using Vocaloid software, in which he inputs melody lines and lyrics for a synthesised voicebank (like Hatsune Miku) to sing. Ayase started making Vocaloid tracks a few years ago; he’d just parted ways with his band and began experimenting with making music on his laptop. He uploaded his Vocalo work to Nico Nico Douga, an online video-sharing platform. To Ayase, the computer is a tool, no different than a guitar or a metronome. “I do understand that some people don’t like it,” he says of Vocaloid music. “It’s not their taste. I hear sometimes that they don’t like it because it lacks emotion. But people say that about music with human voices too. It’s all a matter of preference.” Many of Japan’s top artists—like Yoasobi, Yonezu Kenshi, and Ado—got their start in the Vocalo scene.
With Vocaloid, Ayase can innovate; the possibilities are endless. “You can do whatever you want with it without any limitations,” he says. “And it’s also very easy to access. You don’t have to book a studio, you don’t have to worry about the mic placement or anything. There’s not any fuss. You just have to go to the computer and program it, and then you can make a song.”
In some ways, it’s also the purest form of art. Voice synthesis doesn’t sound human, and for Ayase, that’s exactly what he needs when he’s making music. “When you try to [record a demo] with another human vocalist, then the human vocalist’s habits or their characteristics or personality naturally seeps in,” he says. “That might alter what the song was intended to be before Ikuta hears it. So when it’s done by the Vocaloid, by Hatsune Miku, then I can show Ikuta the melody lines and exactly how I want the song to be so that there’s no other person’s interpretation in-between.”
When Ikuta receives the demo, the first thing she does is write her first impressions on her lyric card. “That first impression is very important to me,” she says. “I think, how should I sing this song? What kind of tone should I sing this in? What are the nuances? I have all of these different ideas that I take with me into the recording process.” There’s a gentleness to Ikuta as she speaks. It’s the same warmth that compelled Ayase to reach out after coming across one of her videos on Instagram. Ikuta is a singer-songwriter who releases her own self-penned music under the name Lilas Ikuta. She looks up to artists like Taylor Swift, who leave etchings on your soul in their handwriting. Personal and permanent. As a solo artist, she takes her time, allowing her feelings to metastasize into art. As part of Yoasobi, her contributions come after the song is already made and the lyrics are written. But her role is just as vital.
“When I’m creating and recording as Ikuta, I’m singing as the protagonist of that novel, or the storyteller or narrator of a story,” she says. She’s dictating the emotion of the song, matching Ayase’s propulsive rhythm beat for beat. Her breathless vocal agility gave Yoasobi its first viral moment after her appearance on the Japanese YouTube channel The First Take. The video of her singing ‘Yoru ni Kakeru’ in a single take from her home has over 132 million views. As Ikuta, she says, “it’s really about expressing what the song needs me to express.”
Ayase still makes his own Vocalo music and uploads it to YouTube, while Ikuta dedicates her free time to her solo work. It’s a balance they’ve perfected over the years—adapting other people’s stories while also making sense of their own. “I do feel a little bit of frustration for not being able to do as much solo work as I want,” Ayase says. “At the same time, I appreciate the fact that it’s because of Yoasobi that I’m able to have this career and do my solo work.” Yoasobi gives them the opportunity to “be more playful and adventurous,” he adds. “We’re able to take risks and challenge ourselves because we have each other to lean on.”
He describes that process as “everybody contributing what they can in a very human way.” It’s where mechanical and manual meet, and it’s not all that different from the process that taught an artificial device to sing in 1961. No matter how sophisticated they are now, machines aren’t capable of feeling. That task falls to the person behind the computer, using a program to process their own analog emotions.
This article was first published on Teen Vogue.
