Slaves to the algorithm beat? Not quite
Algorithms get bad press. Once an obscure computing term for a sequence of steps to perform a task, the word has acquired a sinister connotation in the era of Big Data.
We imagine algorithms enmeshing us, invisibly shaping our behaviour as they process information about us, slotting each person into place in a network of marketable personality types.
“Algorithmic” is assuming the same pejorative meaning as “generic” and “formulaic”.
Pop music is particularly prone to such sentiments. Ed Sheeran is a “pop algorithm”, the New York Times notes sniffily of his new album, “able to produce reasonable re-creations of a whole range of styles”.
Writing about The Chainsmokers last year, the music website Pitchfork said that the critically reviled but highly popular “tech-bro” duo use “a finely tuned algorithm” to write their EDM (electronic dance music) hits.
It was not intended as a compliment: The record being reviewed was given Pitchfork’s sixth worst score of 2016.
Suspicion of algorithms in pop connects with two long-term themes. One is the factory-inspired approach to making music, the Motown model whereby acts are standardised in order to appeal to as big a market as possible.
The other is the encroachment of technology into music, which in turn continually provokes anxieties about where humans end and machines begin. Are the computerised cadences of modern music ushering us into the post-human age?
“Technology is human,” counters the electronic musician and digital artist Robert Henke. “It is a human impulse to want to create a beautiful piece of computing code that generates a beautiful piece of music.”
Born in Munich in 1969, Mr Henke straddles the divide between science and art.
He came to music from an engineering background, forming the techno duo Monolake in Berlin in 1995. He and his Monolake partner Gerhard Behles helped to create the software programme Ableton Live in 2001.
It has become a popular tool for composing and performing digital music, with users ranging from laptop neophytes to Daft Punk.
“For a long time, I struggled with this contradiction,” says Mr Henke, who is not formally trained in music.
“Was I an engineer or an artist? Then I realised I could use my scientific knowledge to create something that would allow me to write music. But when I work, I try to keep the technical side separate in my mind from the artistic side.”
His new show is called Lumière III. Currently touring, it uses custom-designed software to trigger a rapid patterning of lasers and sounds, a high-tech son-et-lumière whose visual action unfolds on a screen on stage.
At its premiere at the Barbican in London last month, Mr Henke stood at the back of the auditorium, a shadowy presence at the mixing desk, overseeing the operation of the algorithms that he had developed for the project.
The use of algorithms in composition falls into two general categories.
One is for people to compose music on a computer.
The other is for computers to compose music on their own.
Numerous artificial intelligence (AI) projects are currently engaged in trying to write pop songs. Unlike beating grandmasters at chess, however, the task turns out not to be straightforward.
Last year, researchers at Sony used their Flow Machines software to produce an AI-generated song, Daddy’s Car.
The result is an awful Beatles pastiche that required input from French musician Benoit Carre, who arranged the music and wrote the lyrics. It is all a bit nul points, alas.
But one can imagine AI mastering a formula-based art form such as pop music.
Indeed, the Flow Machines software recently created quite a groovy acid house song (at least until the vocals kick in).
Where AI stumbles is in developing new trends rather than analysing old ones. A computer can be programmed to find coincidences but not to recognise “happy accidents”.
“Mistakes are part of the work,” Mr Henke says.
“You have to go back and analyse why the accident happened and what difference it makes to the piece.”
The software program Auto-Tune was released in 1997 as a tool for correcting vocal pitch.
By chance, it was found that a certain setting imparted a strange electronic wobble to a singer’s voice.
Pioneered by Cher’s 1998 hit I Believe, the effect is now a staple of pop. (“We never thought anyone in their right mind would do that,” Auto-Tune’s creator Andy Hildebrand told The Financial Times in 2008.)
“There seems to be a strange misunderstanding of algorithms,” says Ms Shelly Knotts, a classically trained violinist who has switched to computer music.
“It’s like they’re some dark thing with no human connection. Somehow people don’t always connect the algorithm to the programmer. I would see an algorithm much the same as I would see any other tool, like using a piano.”
A PhD student at Durham University, Ms Knotts, 31, is a “live coder” who operates at an unusual intersection between computer science, electronic music and improvisational performance.
Live coding involves interacting with algorithms in real time, altering sounds and beats by tapping commands into a computer.
The practice began in the early 2000s as a reaction against the preprogrammed nature of electronic music shows.
The small but growing scene — of which Ms Knotts is no longer one of the few women involved — has coalesced around “algoraves”, a hacking-raving hybrid.
Algoraves take place globally, although the movement is centred on the United Kingdom.
The events involve a programmer such as Ms Knotts writing code on her laptop, which is simultaneously projected onto a screen. The code generates changes in the algorithmically driven music.
When not lost in Dionysian ecstasy at the tunes, the algoravers can pause to admire the mathematical intricacies of the code unfurling on the screen.
“It’s to do with thinking through a system and how to interact with a running algorithm,” says Ms Knotts.
“Quite often the algorithms do something quite different from what you’re expecting. It’s much more like free improvisation in jazz.
“You have to react to the unexpected and come up with some reasonable response from what resources of algorithms you have in your brain.” Ms Knotts constructs her own music software.
“Whenever I used other people’s software I felt really restricted because so many design decisions had been made in advance,” she says.
Algorithms raise questions of authenticity and authorship, the percolation of individual expression through the circuitry of a computer. But they are an expression of creativity, not its nemesis. THE FINANCIAL TIMES
ABOUT THE AUTHOR:
Ludovic Hunter-Tilney is Financial Times’ pop critic and freelance feature writer.
