From megabytes to masterpiece

From megabytes to masterpiece

University of Manitoba composer Örjan Sandred uses computers to create new music
November 1, 2007
Just as the Rolling Stones were about to release their 1997 single Anybody Seen My Baby, a band member discovered that the chorus sounded a lot like k.d. Lang’s Grammy Award-winning song Constant Craving, which had been released five years earlier. Hurried arrangements were made by the Stones to add Lang and her co-writer’s names to the credits.

An iconic rock band and a country singer—the two have notably different music styles, and yet, with infinite possibilities for composing a song, almost identical scores were written. Surprising? Not really, says Örjan Sandred, a composer and researcher at the University of Manitoba. “It’s almost impossible to compose music with truly new structures,” he says.

Most artists use their creative intuition to write music—their ideas come from their subconscious, which is based on past experiences, including music they’ve heard or studied before. In other words, new music is inspired by old music. “But, if we can develop a new method of creating music, we will get a different result,” explains Sandred. He hopes to develop just such a method at the university’s Studio FLAT, a state-of-the-art studio for computer music research, slated to open in September 2008 .

Studio FLAT, the first of its kind in Canada, will build on Sandred’s expertise in Computer-Assisted Composition (CAC)—a multi-disciplined approach to composing music using sophisticated software and one’s own talent—and explore how computers can be applied to composition.

When Sandred started composing at age 13, instruments weren’t even considered the best vehicle for composition. “I had a teacher who said we should not even use a piano when composing music, just a pen and paper,” he recalls. “But there is no right or wrong way to create art, and the idea of how a composer should work is changing.” A pivotal year spent at Institut de Recherche et Coordination Acoustique/Musique (IRCAM)—an internationally recognized music research institute in Paris where computer science, acoustic research, and music composition coexist—convinced Sandred that using computers would help him to create new music.

Gérard Assayag, head of the Music Representation Research Group at IRCAM , explains: “Örjan cannot be satisfied with existing ideas, rather he is keen to invent new ideas people didn't think of before”.

When it comes to using computer programs to write music, Sandred is one of only a few composers worldwide who deploy what are called rule-based expert systems. Using his computer, he writes programs that determine how music should be structured, how rhythms relate to metric pulse, and how chord progression evolves in relationship to rhythm.

“My main goal is to open my eyes to music structures I did not think about,” he says. Using advanced software, the computer takes Sandred’s rules and generates a musical score. Ten years ago, it would sometimes take two full days to do this. Today’s more powerful machines can not only generate compositions quickly, but can also solve complex musical riddles. So instead of working with a single set of rules, say about rhythm, as most computer composers do, Sandred will assign rules to rhythm, harmonic progression, metric pulse as well as other components simultaneously. “I can get really creative with what I tell the computer to do now,” he affirms.

After the computer finishes its work, Sandred takes the score and makes it “playable.” The computer-generated piece may not sound precisely the way the composer intended, and so manual adjustments are often necessary. Sandred tweaks the tunes, but “I try not to change too much. I try to stay true to the results as much as possible,” he says.


“Music research is an attempt to create new music that reflects our time,” says Sandred. For example, during the early 1980s, Stanford University composer and music researcher John Chowning helped develop the Yamaha DX7 Synthesizer, possibly the most popular digital keyboard ever. The sounds musicians could create with the DX7 revolutionized music and helped defined a decade of tunes.

“Such people (like Sandred and Chowning) act on two levels—for the benefit of civilization and for technical progress,” says Assayag. “An advanced civilization always defends and supports its brave artists seeking new paths in the exploration of creativity. But even if that was not the case, such artists are essential to technical progress because they bring unique challenges to technical people, which result in new tools for average people.”

Sandred isn’t working on the next synthesizer, but he is trying to achieve the same goal—bringing music to a new technical level and exposing CAC to a larger audience. Beyond the creation of new original music, encouraging more composers to use CAC may also result in more compositions that successfully connect with listeners.

“Music is very abstract,” Sandred explains. “You listen and it’s gone. To be able to communicate something, you have to give it lots of structure so the listener is able to grasp something. All music depends on structure, and all structure requires a mechanical aspect. That’s where the computer comes in.”

Critics say the computer takes away from the human component of music, but Sandred argues the opposite is true. “There are a lot of mechanical things in writing music and that’s what the computer is good at,” he says. “The computer allows me to focus and reflect more on the human element. And the computer develops music in a way I couldn’t develop by myself.”


Örjan Sandred’s network of research partners is strongest on the other side of the Atlantic, where he’s been composing since childhood. In Europe, he is part of PRISMA (Pedagogia e Ricerca Internazionale sui Sistemi Musicali Assistiti), a centre for musical research, production and education that brings together a network of prestigious musical academies and universities, at the Centro Tempo Reale, University of Florence, Italy.

With the creation of Studio FLAT, Sandred hopes to start a similar collaboration in Canada, by attracting composers and students with unique ideas. “They will come in with ideas that are different from mine,” he says. “Together, we can take a direction with music research that has not been done before.”

Learn More

Learn about PWGL, a web-based program which allows Sandred to collaborate with other composers.