News Bureau | University of Illinois

New computer application integrates composition and sound synthesis

Melissa Mitchell, Arts Editor
217-333-5491; melissa@illinois.edu

Click photo to enlarge Photo by Kwame Ross Sever Tipei continues the tradition of the computer music pioneers at the University of Illinois School of Music with his latest collaboration.

6/20/05

CHAMPAIGN, Ill. — For nearly a half century, composers at the University of Illinois at Urbana-Champaign – often in collaboration with scientists and engineers – have been making music and music history within the soundproof walls of the School of Music’s Experimental Music Studios.

That history dates to 1956, the year Lejaren Hiller, a chemistry professor working on a master’s degree in composition, and research associate Leonard Isaacson completed their “Illiac Suite” for string quartet – the first musical composition generated by a computer. What began at Illinois as a novel exercise prompted by intellectual curiosity gave rise through the years to a new form of artistic expression commonly referred to as computer music.

That tradition of experimentation – which expanded from computer-assisted composition to include computer sound synthesis, visualization of music and scientific sonification – lives on at Illinois, where composers and music theorists still trade notes long into the night with computer programmers, mathematicians and other techie types. Except in today’s WiFi world, of course, the gurus of the studios’ Computer Music Project no longer crunch numbers and bend sound waves in the shadows of tall towers of mainframe computers; the studios’ minimalist surroundings now include just a few desktop machines and speakers. And current collaborations sometimes occur from a distance.

That was the modus operandi for Sever Tipei and Hans G. Kaper, the brains behind the CMP’s latest killer app: the Digital Instrument for Sound Synthesis and Composition, better known by its nickname, DISSCO.

Tipei is a U. of I. professor of composition and manager of the CMP; Kaper is a senior mathematician at the U.S. Department of Energy’s Argonne National Laboratory, and has an adjunct appointment in the U. of I. School of Music. Together with graduate students from Tipei’s advanced computer music seminar, the musician and the mathematician have worked on the development and testing of the software since 1993. It recently was made available free to the public under terms of the GNU General License on SourceForge.net.

So, what makes DISSCO such a hot property among those drawn to both rhythm and algorithm?

Click photo to enlarge John Cage, left, and Lejaren Hiller collaborated in the Computer Music Studio in the late 1950s.

“With DISSCO, composition and sound synthesis are integrated in one seamless process, and share a common formal approach,” Tipei said. In other words, on the composition side, the software functions as an assistant – or collaborator of sorts – to the composer, who still maintains the ultimate creative license.

On the synthesis side, “DISSCO is a tool for the composer – like a piano,” Tipei said. “The beauty of this is that we have absolute, detailed control over everything.”

Using DISSCO, the composer determines the formal structure of a composition – selecting basic elements, such as scale factors, tone values, pitch and timbre, as well as “modifiers” such as vibrato, tremolo and glissando. At the same time, the composer plugs in variables to create different types of sound combinations – in effect, creating the instruments to play the composition.

“What we’re doing is creating sounds completely from scratch,” said Tipei, who noted that unique features of the program include precise control over perceived loudness of sounds and the ability to produce a finished piece of music that requires no further processing.

And even though the composer controls much of the output by punching in a host of variables, DISSCO is capable of inserting some degree of randomness in the patterns and sounds that are generated.

“The randomness is more of an aesthetic choice than anything else,” Tipei said. “Lots of people – for good reason – complain that electronic music is boring. When presented in concert, there’s nothing to look at on stage and is always exactly the same. One solution is to add live performers.” Another, he said, “is to add degrees of randomness in the composition.”

“For my own compositions, I ask performers not to play the same version twice. Unlike in a gallery, where a painting always appears the same, my work is going to be different each time.”

Tipei said that he expects DISSCO will be used to make music for any number of applications – by composers such as himself, who create electronic art music for performance and recording, as well as those who produce sound accompaniment for digital media, films and television. He also has used it as a valuable tool for teaching music-composition and theory in his classes.

In addition to its use as a music-composition tool, DISSCO has the ability to translate complex computer data into sound waves – using a process known as “additive synthesis,” which builds sounds from simple sine waves – making it useful to scientists and mathematicians. Like visualization techniques that morph data into images that can assist scientists in distinguishing patterns or aberrations in the data, DISSCO can be used to reveal similar characteristics through sonification – the faithful rendition of scientific data in aural images.

“Scientists can use this instrument to explore scientific data by rendering them in a sound file,” Kaper added. “The data are used to define the characteristics of the sound wave, such as the way it is tuned, its loudness, its spatial distribution, and the amount of reverberation. In all, there are more than a dozen useful degrees of freedom that we can build into a sound – more than enough for most physical or computational experiments.”

During DISSCO’s development stage, Tipei said the researchers tested its sonification capabilities using data from a computational chemistry experiment.

“We looked at two interacting molecules, and measured energy levels before and after a reaction,” Tipei said. “The sounds indicated when things ‘happened’ in the molecules.”

When used for scientific applications, “the software probably will be most useful when used in connection with visualization,” Tipei said.

Although DISSCO version 1.0 has been released on SourceForge.net, the U. of I. professor said the fine-tuning phase continues.

“It’s by no means finished; it’s a work in progress,” he said.

Meanwhile, Tipei and Kaper plan to travel to Barcelona in September to present a paper on DISSCO at the 2005 conference of the International Computer Music Association. And later this year, two of Tipei’s DISSCO-generated compositions will be featured among other selections on a Centaur recording.

Samples of music created with the program are available.