July 22, 1996 in Nation/World

Noise Busters Research Team At Sirti Using Mathematics And Computers To Wipe Out Bad Audio

By The Spokesman-Review
 

Professor Steve Simmons and his students are on a mission to stamp out bad sound.

They’re using mathematics and computers to clean up the pops, snaps and hiss found on old vinyl records.

Their research may also spawn technology capable of cleaning up the sound from “live” sources, such as intercoms and public-address systems.

“Bad audio is essentially everywhere,” said Simmons, who teaches computer science at Eastern Washington University.

XN Technologies of Cheney and the Washington Technology Center are putting up most of the money for the project.

This is the kind of scientific endeavor intended at the Spokane Intercollegiate Research and Technology Institute (SIRTI), built to stimulate high-tech industry in Spokane. Simmons’ team is headquartered there.

Simmons said his initial idea was to come up with a way to improve and preserve vinyl-record collections stored in the nation’s libraries.

Using mathematical formulas, the records are played into a computer. A software program then finds and eliminates the noise.

Simmons, who once built a home-stereo amplifier from a do-it-yourself kit, thinks in sine waves and algorithms. He’s a “techie,” and he’s gathered around him a group of like minds.

Graduate students Keith Harland and Mike Wahlstedt are researching and writing algorithms, the mathematical recipes that will do the work of eliminating noise without harming the real sound.

Harland is an electrical engineer who once worked in the defense industry but is back in college taking advanced computer-science classes.

“I love the challenge of it,” Harland said of the research.

Wahlstedt is a retired Air Force navigator launching a second career.

“We are constantly trying to improve what we’ve got,” Wahlstedt said.

To demonstrate, they played an abused vinyl recording of the Beatles’ “I’m So Tired.”

The group’s software analyzed and identified the sound-wave patterns created by the popping of a deep scratch and - voila! - the song was replayed on a computer with no pops.

“Garbage in, good stuff out, you could say,” Simmons said.

During the next year, the research team hopes to discover a way to repair bad sound as it is being received in “real time.”

For example, microwave transmissions are frequently plagued with bad audio, yet higher education is increasingly going to two-way electronic hookups to expand its reach.

There’s no guarantee the research will succeed. “We are taking a measured risk,” Simmons said.

But that’s part of SIRTI’s mission. The institute provides a home for scientific projects that are too speculative for private companies to take up on their own, but still have promise.

XN Technologies is paying about 30 percent of the lab expenses, estimated at $135,000 a year, and would share any profits with the university if the research sires commercial products. The state technology center is contributing 60 percent, with EWU chipping in the other 10 percent.

Under this kind of arrangement, technological breakthroughs come relatively cheap. The graduate students receive a $600-a-month stipend, and Simmons’ salary is paid by the university.

As the project continues, Simmons is learning a greater appreciation for the miracle of hearing. The human ear can easily detect a sound wave that’s no larger than a single atom, he said.

“Your ear is such a fabulous amplifier,” he said.

, DataTimes


Thoughts and opinions on this story? Click here to comment >>

Get stories like this in a free daily email