In this age of technological marvels, it seems hard to believe that none of this existed a half-century ago. Only in the minds of a brilliant few did the kind of world we have now even begin to exist, and even then, it was as a fanciful shadow more suited to science fiction than reality. Thanks in large part to the genius of Alan Turing, the father of the modern computer, we now have a world that relies almost solely on the extraordinary abilities of computers and technology. Computers do more than simply enable our world; they are becoming citizens in it. Artificial intelligence is a hot field. It’s one that inspires equal parts fear and fascination, and now classical music has a very strong stake in it. This latest stake is a computer program called Iamus, and it composes music. We’re not talking midi-files or little jingles here. We’re talking big-kid, real-sounding music. In fact, the music that it composes is almost indistinguishable from music composed by a flesh-and-blood musician–that is, if The Guardian‘s own “Musical Turing Test” is any indication (I guessed correctly, but it wasn’t based on anything more than a hunch).
The Iamus program, which is named after the son of Apollo whose chief claim to fame was his ability to understand the language of birds, was developed by a team at the University of Malaga. Of the program and its uncanny oeuvre, team leader Francisco Vivo says, “When we tell people that, they think it’s a trick. Some say they simply don’t believe us. Others say it’s just creepy.” Either way, the public at large will have a chance to weigh in when an album of Iamus’s music is released in September. The music on the album may be computer-generated, but the performances will not be: A-list musicians including the London Symphony Orchestra will be performing the works.
Even the team’s explanation of the way Iamus works is a bit jarring, especially for a community which is often more fluent in 18th-century Italian musical conventions than computer jargon. Put most simply, Iamus composes by mutating simple starter material in a process that mimics biological evolution. The compositions each have a musical core, a “genome” (in the words of Guardian reporter Philip Ball), that gradually becomes more complex. “Iamus generates an initial population of compositions automatically,” Vico says, “but their genomes are so simple that they barely develop into a handful of notes, lasting just a few seconds. As evolution proceeds, mutations alter the content and size of this primordial genetic material, and we get longer and more elaborated pieces.” All Iamus needs from the human researchers is the desired length and instrumentation of the piece to be composed.
So how does Iamus sound like anything more plausible than a computer writing music? Why do we hear its compositions as so similar to the music being written today? “A single genome can encode many melodies,” explains composer Gustavo Díaz-Jerez of Musikene, the Higher School of Music of the Basque Country in San Sebastian. “You find this same idea of a genome in the western musical canon – that’s why the music makes sense.” Díaz-Jerez has been collaborating with the University of Malaga team since the outset of the Iamus project, and he is the pianist performing in Iamus’s compositions. Iamus is not limited to any given musical aesthetic. Says Ball, “Although most of its serious pieces are in a modern classical style, it can compose in other genres too, and for any set of instruments. The Darwinian composition process also lends itself to producing variations of well-known pieces or merging two or more existing compositions to produce offspring – musical sex, you might say.”
The process itself is part of what makes Iamus so much more successful than its forerunners–and there were indeed forerunners. Of course, the role of the computer in composition is not a new one. Almost as soon as there were computers, composers were intrigued by their potential, often choosing to use computer algorithms to select music. Greek composer Iannis Xenakis began doing this in the 60′s, and in the following decade two Swedish researchers came up with an algorithm to create music in style of the Swedish composer Alice Tegnér. In the 1980′s, the computer scientist Kemal Ebcioglu created a program that harmonized chorales after Bach’s examples. Other musical efforts include programs that can improvise music in performance, and those have met with some success as well. However, until Iamus, efforts to create a program that convincingly produced music from scratch have been largely unsuccessful.
So how successful has Iamus been, really? After some initial and very understandable skepticism, Díaz-Jerez says that musicians have been pleasantly surprised by the output–albeit some more than others. Says LSO chairman Lennox Mackenzie, “I felt it was like a wall of sound. If you put a color to it, this music was grey. It went nowhere. It was too dense and massive, no instrument stuck out at any point. But at the end of it, I thought it was quite epic. The other thing that struck me was that it was festooned with expression marks, which just seemed arbitrary and meaningless. My normal inclination is to delve into music and find out what it’s all about. But here I don’t think I’d find anything.” But Makenzie’s verdict on a whole is one of cautious interest: “I didn’t feel antipathy towards it. It does have something. They should keep trying, I’d say.”
What does this mean for the future of classical composition? Only time–and the reactions to Iamus’s music–will tell.