Endel uses artificial intelligence to create personalized music. The soundscape app promises to improve concentration or facilitate relaxation, depending on what the user requires. The app even claims to alleviate sleeping problems. A self-experiment.
Written by Ralph Hofbauer
Illustration by Reza Bassiri
March 31, 2021
I sit in my home office, launch Endel, and activate the “Focus” mode. The app plays spherical sounds, a synthesizer weaves meditative harmonies. It’s an ambient sound reminiscent of Brian Eno, unobtrusive and relaxed – more atmosphere than music.
The Berlin startup does not want to describe what Endel outputs as music, either. Oleg Stavitsky, founder and CEO, prefers the term “functional sound”. The sounds are intended to reduce stress and improve concentration. This is in line with the spirit of the times – after all, mindfulness is in vogue.
It actually works: I get into my flow.
The sound blanket is so soft that I hardly notice it. I have to force myself to listen actively. The sounds are constantly in motion. A subtle beat kicks in, the synthesizer gives way to a marimba.
Endel is producing this soundscape just for me. Artificial intelligence composes the sounds in real time and adjusts them to parameters such as weather, biorhythm, and pulse. In addition to the Focus, On-the-go, Relax, and Sleep modes, the app offers sound for scenarios such as workout, reading, or yoga.
The app informs me that it is currently producing sound for my “afternoon energy slump”. The beat has picked up and I imagine myself sitting in a lounge on a Balearic island, a drink in my hand. Focused, I work my way task by task toward the end of my workday.
What works for me apparently also works for others: Endel was tested using the method of Mihaly Csikszentmihalyi, the discoverer of the flow phenomenon. The test subjects’ concentration increased by a factor of 6.3 and also, their feelings of anxiety were significantly reduced.
I hop onto my bike and switch to the “On-the-go” mode. The transition is seamless, the soundscape changes smoothly: Strings join in, then a piano. Ambient transforms into modern classical, but the relaxed mood remains the same as before in my home office.
Even though AI generates these sounds, they still bear the human signature of the musician and sound designer Dimitri Evgarov. His original compositions appear on the 130701 label, where the renowned neo-classical artist Max Richter also released his first albums.
My bike carries me out of the city and into the countryside. The dreamy, melancholy sounds blend in well with the rural idyll. My movement data control the sounds, but the composition has sparse dynamics. The melodies ripple along, without any tension or friction whatsoever.
The untroubled harmonies are no accident. The AI uses pentatonic scales, a 5-note pattern found in many folk and children’s songs. Suspenseful notes are impossible, so the AI cannot really do anything wrong. Any sequence of notes sounds right – i.e. harmonious.
Before tackling a climb, I activate the workout scenario. The percussion becomes more dominant and picks up a notch. Mentally, however, I remain in the spa: The sound pushes me on, but never becomes so energetic that I develop ambitions to win the “Maillot Jaune” jersey. It sounds a little like Nils Frahm on steroids.
Endel is not meant to rival musicians, its creators emphasize. To highlight this, they arranged for the AI to collaborate with artists: The Canadian singer Grimes fed the algorithm with her own sound. The result is called “AI Lullaby” and sounds different for each user.
I turn off the light and launch “AI Lullaby”. Cosmic sounds surround me. I feel like an astronaut lying in a berth on a space station light years away from Earth. Now and then scraps of language filter through to me, like radio messages from the mothership.
The song – music written down for eternity – is no longer a thing with Endel. Nevertheless, the AI has immortalized itself on soundtrack: The Warner music label markets 20 albums composed by the algorithm. The major deal caused quite a stir in 2019. The headline: “First AI with a record deal”.
Last night, I fell asleep after what felt like five minutes. This evening, I try out the “Sleep” mode. Once again, the spherical sounds can be heard, but today I am lulled to sleep by the sound of the waves. It brings back memories of my last vacation by the sea – and I’m off.
In the meantime, Endel has more than two million users and was chosen by Apple as the “Watch App” of 2020. The startup supplies its sound to airlines and car manufacturers, but also has its sights set on the healthcare sector: Music as medicine, the app as a non-harmful psychotropic.
The next morning, after a
week with Endel, I put on a record from my collection. The song puts a smile on
my face, my foot starts tapping to the beat. Endel was never able to do this.
Its makers are right: This is not music, rather it is a sound tool for
This is not music, rather it is a sound tool for self-optimization.
A manifesto explains, not entirely immodestly, how Endel as a “tech-aided bodily function” wants to advance the evolution of humanity. The app resolutely continues what popular Spotify playlists such as “Workday Zen” initiated. The sound bubble turns into a constant companion in everyday life, and music is reduced to its most basic function. No artist seeking to be revered and no melody that sticks in your mind.