Musical machines: a different take on music and data

It’s easy to get bogged down in the seriousness of the music industry and the exciting potentials of how big data can change music for the better. However, just like music, there’s pockets of pure joy and weirdness to be found at the intersection of music and data, and we’re going to explore one of the more entertaining ones today.

We’ve mentioned before that data and automation will help musicians do their job and make music, but what if the data was the root of the music. Now we’re not talking about the sort of experimental music like Ryan Patrick Maguire’s moDernisT, in which he reconstructs the data lost when making an mp3 from Tom’s Diner, but something a little more akin to a digital Weird Al.

Who put the bot in the bomp bah bomp bah bomp?

Enter Botnik Studios — an artistic collective exploring what happens when machine learning is used to write music. Between machine written Christmas carols and the surprisingly compelling Morrissey/P90X workout DVD Amazon review lyrical mashup “Bored with this desire to get ripped”, this brilliantly bizarre group of engineers and musicians fed a recursive neural network with a full catalogue of country music’s greatest hits, just to see what would happen. It created something new and unprecedented — a listenable country song, complete with music and lyrics.

This (human performed) masterpiece, entitled “You can’t take my door,” is catchy, amusing, and quite interesting to examine. The music is generic but engaging, and it sounds like most country-pop hits of the past 30 years. The lyrics, on the other hand, are less comprehensible — or perhaps more profound. While “No you can’t take my door/I don’t wanna love you anymore” might not be the most credible lyrics ever written, “Barbed whiskey good and whiskey straight” is certainly more profound than Rhett Akin’s “that’s my girl, my whole world/but it ain’t my truck” — lyrics from a song that had three genuine human writers and spent 21 weeks on the Billboard Hot Country Songs list in 1995. It’s certainly food for thought.

Not Orwell, not yet

Machine written music isn’t a new idea. A minor plot point in George Orwell’s landmark dystopian novel 1984 was the versificator, a kind of manual songwriting machine that churned out popular, politically friendly music for the proles. This opens the door to an interesting notion — is machine-produced music art? How would it be managed — and more importantly, how would it be compensated? This is an issue of some sensitivity, with streaming giant Spotify first facing accusations of using fake artists in 2017 and continuing to generate questions to this day.

Now, the Spotify questions likely deal with fraudulent use of the platform, and aren’t a precursor to Skynet’s The Rise of the Machines World Tour 2020 (or at least not quite yet), but the ambiguity remains. If music is created by a machine, and even potentially published by a machine, who owns the rights? Who owns the revenue? And do consumers have a right to know? Will they even care?

Getting vocal(oid)

On the flipside, human songwriters in Japan are writing for machine performers, called Vocaloids. The first, and most famous vocaloid star is Hatsune Miku, who debuted as a sales avatar for the vocaloid voice synthesizer software in 2007. She’s a bonafide pop star despite being, well, not actually real — a digital Milli Vanilli as it were.

So what would happen if big data, machine learning, and artist synthesis were to join forces? Unstoppable musical success, or homogenized elevator music coming from an auditory Uncanny Valley? We’ll inevitably find out sooner rather than later — whether as a massively publicized launch of a new musical phenomenon, or creeping unnoticed into our earphones, the musical machines are here to stay.

Music, money, and data

At Utopia Genesis, we’re building and developing musical machines of the future. Our goal is to use better data and blockchain-powered technology to make sure that human artists get paid for every play, and that the music industry is empowered to continue making the music we love.

Curious to know more? Join our Discord and Telegram group — Utopia Genesis.