Andrson is a company with a music-matching algorithm designed to find songs that sound similar to ones you’ve already heard. Its product is, for now, aimed at making it easier for music labels to find would-be stars, but that’s not the only thing it’s useful for. Today, the company is launching ReWrapped, which analyzes your Spotify favorites to find you soundalike tracks from its roster of unsigned artists.
Back in the days before social media networks made their own stars, would-be acts had to catch the eye of a roving talent scout. These people — known in the industry as Artists & Repertoire (A&R) — were tasked with finding and grooming would-be stars. Andrson thinks that it can do the same job, only faster, by letting its algorithm do the legwork to find hits.
If you’re a young singer-songwriter, then you just need to upload your files to Andrson’s cloud, and it analyzes which existing act you sound most like. Then, an A&R person who is plugged in to the other end can search based on the existing hit acts that you sound like. So, for instance, if you’re looking for the next Dua Lipa, you could search for artists who sound a bit like the current Dua Lipa.
Co-founders Neil Dunne and Zach Miller-Frankel explained that the platform already has around 8,000 artists, who have uploaded more than 15,000 tracks. They added that users who upload their content can ask for what they’re looking for — including representation, or even to collaborate with other users on songs. They added that the algorithm looks for around 600 features in each song, and these data points create an “unbiased” audio fingerprint.
ReWrapped is, then, a neat way of introducing the power of Andrson’s music-matching algorithm to the wider public. It works by scanning through your Spotify Wrapped playlist, a cultural event that more than 90 million people participated in at the end of 2020. From there, it’ll offer up songs that sound similar to the ones it’s selected, for you to enjoy, or not.
I offered up my own Spotify Wrapped for the algorithm to judge and it plucked out five songs from the 100 to analyze. Through the site, each song has three suggestions from unsigned artists for me to listen to. You also get a link to a Spotify playlist including all of the songs from Andrson’s stable that are also available on the streaming service. In my case, 13 of the 15 tracks were ready to be played through Spotify, as well as on Andrson’s own mini-site.
You can certainly tell that the algorithm’s modus operandi is to take a sample song and find things that sound similar to it. Ninian Hawick’s “Scottish Rite Temple Stomp,” for instance, which is a piece of late ‘90s electro-pop, was met with three similarly crunchy electro-pop titles. Similarly, White Denim’s “Magazin” had three equally guitar-heavy pop songs that sounded plenty similar to what had been submitted.
It’s not a perfect system, of course, because the algorithm can, for now, only work with the catalog of songs in its database. So a mellow-ish piano-only track like Thom Yorke’s “Suspirium” was met with mellow-ish piano music. But it’s certainly impressive that the algorithm can find you songs that have a similar sound to the ones submitted. Unfortunately, very few of the songs I listened to I wound up enjoying, because just because something sounds similar doesn’t mean I’ll like it. But it’s a nice way to explore a different method of music discovery, and the algorithm can always develop in the future.