With more than 16 years’ experience as a Musicologist, Professional DJ and Music Curator/Consultant, Johnson has created thousands of playlists. He was a resident DJ at world’s most successful nightclub, Pacha Ibiza and classically ...
Muru Rethinks the Rules of Music Classification with AI-Powered, DJ-Inspired Curation
Meet the world’s first AI DJ brain, Muru. Inspired by an Aboriginal Australian term glossed as “journey,” the Sydney-based music technology company transforms metadata, the basic information about every digital music file, and generates extended playlists to fit every taste, every dynamic, by disrupting the way music is classified in the digital realm.
“We’ve built tools that allow streaming and other digital music services to optimize listening experience, streamline data flow, and create new revenue opportunities. Muru aims to add value to the entire industry. Our classification and recommendations helps artists, streaming platforms, publishers and of course the music fan.” explains Muru founder Nicc Johnson. “We’ve brought big data in harmony with the human response to music,” a response Johnson witnessed extensively as a veteran dancefloor DJ.
Muru playlists can be generated in a matter of seconds, even when they span hours, by selecting genres or artists as beginning and ending points. It can work on top of any streaming service. Muru can start cocktail hour with some cool jazz and morphs to deep house by the time everyone’s done with dessert. Users can nudge the playlist to evolve in real time, as they adjust parameters from energy to era to popularity. The approach is streamlined and intuitive, yet reflects years of serious thinking into how we relate and move to music.
Muru’s Journeys (the company’s name for its playlists) unfold so seamlessly thanks to a proprietary approach to music classification, one that embraces both the wonders of algorithms and the centuries-old study of what music does to people. Muru aims to make machines think more like dedicated DJs--and find the perfect groove in the process.
Muru’s founder was raised by voracious music listeners, on everything from rock to his father’s African music connection, from jazz to classical. As a young teen, he discovered electronic music and was hooked. He could not have grown up in a better spot for mastering the art of DJing: Ibiza, the Spanish island that transformed the club music scene. Johnson got his start in his mid-teens. By age 17, he was spinning 12 hour shifts starting at Pacha’s clothing store, later to return as a resident DJ for the club for 7 years in his mid-twenties. He honed his craft, but bigger questions fascinated him.
After a while, the Ibiza club world felt a little too small. Johnson headed for Madrid to study sound engineering, only to set music aside for a while. “I stopped doing music altogether for a few years,” he recalls. “I came back to it when I realized it was my passion. I was interested in understanding what exactly happens on the dancefloor, what happens to our body and mind. I wanted to understand what was lacking in the industry and what I could do to improve it.”
He dug into musicological literature on reception, genre, and other well-explored but complex topics. He sifted through his own experience of keeping the dancefloor lit for hour after hour. Then about five years ago, as Johnson was working as a consultant crafting lengthy playlists for retail and dining clients, he stumbled on an idea. “I was putting together these really long playlists, and it took forever to generate them with unique tracks, if I didn’t want to repeat things,” recounts Johnson. “I found myself losing potential business. I knew there had to be a different way.”
AI and other algorithm-based approaches were just coming into their own at the time, but no one had a product that did what Johnson wanted. In his irritation, he found his idea coalescing. “I noticed a few fundamental errors. Streaming services are led by incredible engineers, but they don’t have the same intensive experience with song sequences or their effect on people. Or they employ manual curation and suggestion approaches, run by humans, yet their knowledge and speed can’t match the huge scale of digital music. They miss a great deal of what’s out there,” reflects Johnson. “That was my point of frustration. I saw that if we could fix the knowledge gap, we could fix the whole industry.”
He began to wonder if he could provide the fix. He had a couple beers with a friend in Sydney, jotting down some concepts on a scrap of paper. The following week he got a call: he had found his first investor. Those brainstorms quickly turned into Muru.
Johnson and his engineering partner, Vijay Santhanam, began trying to teach Johnson’s DJ skill set to a computer. “We have been working on codifying the DJ part of my brain. If I look at a set of songs, I’ll put them in a certain order I think fits best, dependent on my audience” explains Johnson. “Not necessarily slow to up tempo, but juxtaposition. It’s a complex rule set, but one that is translatable to machine learning.”
To perfect user experience, Muru not only creates listenable playlists with elegant transitions between genres and styles, but also allows users to interact with the app to tweak the journey. “We’ve created ways to let Muru do what I used to do as a DJ when I got feedback from the crowd,” says Johnson. “People have the same reactions intuitively but don’t know how to express them. The things they wanted always came back to the same four questions.”
Muru allows you to play with four parameters to shift the playlist’s vibe: energy, era, popularity (for when you want more obscure or more popular tracks), and tempo. It lets you favorite or skip tracks, taking things beyond the thumbs up or down, telling the app whether it’s the song or the artist you object to. The interface uses color to indicate genre and has a visual representation for duration, perfect for set-it-and-forget-it party jams, for the car, for workouts.
“We’ve really worked to understand how to improve experience, by digging way deeper into the digital catalog and social data, by generating rules that reflect people’s response to music, and by creating something people can use easily to hear music that they love,” Johnson states. “And even with the same genres to start and finish, you’ll never get the same playlist twice.”