advertisement
advertisement

This Primer On Generative Music Is A Masterclass In UX

It elegantly covers everything you ever thought to ask about “generative music”, and ton that you didn’t.

This Primer On Generative Music Is A Masterclass In UX
[Photo: Federico Rostagno/iStock]

We’re used to hearing about AI beating humans at everything these days. But machines have been making music on their own for decades already. It’s even got a whole genre: “generative music.” That doesn’t mean that humans are completely out of the creative loop, though. It just means that what we think of as “composing” gets abstracted one level away from the actual music, and becomes more like system design: the human being devises a mechanism that, once set in motion with certain inputs, creates the music itself in ways that the person may not be able to predict. And those systems may be composed of anything from linked reel-to-reel tape recorders to APIs that access mass-transit data from Helsinki. They may even involve other people.

advertisement

I learned all this from a spellbinding interactive explainer called “How Generative Music Works: A Perspective,” by a designer named Tero Parviainen. Think of it as an encyclopedia entry crossed with a mixtape crossed with one of those bad-ass software demos that UI futurist Bret Victor sometimes makes. Even if you’re familiar with genre—say, from classic works like “Music for Airports” by Brian Eno—you’ll probably learn something new from Parviainen’s delightfully multidisciplinary take on the history and practice of generative music. (Was I expecting passages about Noam Chomsky’s linguistics work or an early machine-learning technology called Markov chains? No, I was not.)

Explore the full site here. [Screenshot: teropa.info]
It even has some generative “musical instruments” built right into the website itself. This one can generate 12 trillion different variations. And here’s one that really does turn real-time mass-transit activity from Helsinki into music.

Parviainen’s writing is tight and approachable (no Wikipedia-style technical digressions). And his typography game is on-point, too (the text is set mostly in Averia Serif, which happens to be a work of generative art itself).

But the best thing about “How Generative Music Works” isn’t just the intriguing content, it’s the humane user experience. Web browsers can do amazing things, but interactive multimedia experiences are still just as hard to pull off as they were in the CD-ROM era. The trouble, as  Bret Victor himself convincingly argued more than a decade ago, is that “interactivity” is mostly a pain the rear end. So I’m always eager to call out interactives that successfully walk that fine line between “reading something interesting online” and “being forced to fuss around with a bunch of irritating widgets.” Parviainen nails it.

Explore the full site here. [Screenshot: teropa.info]
Take the blissfully simple navigation: There is none. You just hit the space bar to advance through it. No screwing around with sliders and buttons and hamburger menus. The transitions are cinematic without being overly flashy—Parviainen uses a Prezi-like pan-and-zoom method combined with simple dissolves. But the “camera movements” aren’t just useless eye candy that’ll send your laptop fan into overdrive. As you move forward in the piece, it becomes clear that the spatial arrangement of “slides” is part of an intentional design allowing Parviainen to make surprisingly elegant juxtapositions in the browser window like a film director.

The way video, animation, data, and music are seamlessly threaded together would make Bret Victor proud, too. It’s not immediately obvious how to deploy time-based media like music or film into a medium like the browser where the user can exercise control over the timing with which they read text and advance through the other content. Steven Soderbergh’s app “Mosaic,” for example, just splits the time-based stuff and the interactive UI “chrome” almost completely apart. But here, Parviainen is able to keep all his aural and visual aids comfortably in sync with static text by locking everything to those space-bar commands. Everything just autoplays smoothly when it should: you don’t have to “drive” it while you read and listen. (Except for the parts where you’re invited to noodle around with a generative musical instrument, of course. I skipped those parts and didn’t feel like I missed anything.)

advertisement

By keeping his interactivity to a subtle minimum, Parviainen has created an object lesson in generative system design of his own. If you tend to shy away from overdone interactives out of sheer weariness, this piece will be music to your ears.

About the author

John Pavlus is a writer and filmmaker focusing on science, tech, and design topics. His writing has appeared in Wired, New York, Scientific American, Technology Review, BBC Future, and other outlets.

More