Mathematicians Solve the 'Cocktail Party Problem'

Aug 22, 2006

Officials at the CIA and scientists around the world have pondered the "cocktail party problem" for decades. How could they separate one sound - perhaps a voice - from a group of other recorded sounds, perhaps a multitude of voices at a cocktail party? Now, two researchers at the University of Missouri-Columbia have found a mathematical solution to this problem.

"Theoretically, our solution says you should be able to pick up voices on a squeaky old microphone and then separate them all out so that you can hear what each person is saying in his or her own voice," said Peter Casazza, professor of mathematics in MU's College of Arts and Science. "This is a very old problem, and we have the first mathematical solution to it."

Casazza and Dan Edidin, also a professor of mathematics at MU, worked with Radu Balan of Siemens Corporate Research to solve the problem. Their solution shows that it is possible to separate voices and still retain vocal characteristics. Researchers had previously found a solution for separating and reconstructing voices, but they were only able to reconstruct the words spoken, not the characteristics of the voice itself.

"Our solution is called 'signal reconstruction without noisy phase,'" Edidin said. "In speech recognition technology, a 'signal' could be a recording of 25 people in a room talking at the same time. Our solution shows that we can pull out each voice individually, not just with the words, but with the voice characteristics of each individual. We showed that this 'cocktail party problem' is mathematically solvable."

Although Casazza, Edidin and Balan do not have a computer program that can do this automatically, they hope to find a way to develop one. Currently, their solution runs on a computer, but the process cannot be easily replicated or distributed.

"The computer we use is doing the work without an algorithmic program. It uses a system called a neural net, which is designed for the computer to teach itself. Basically, it works on trial and error," Casazza said. "This isn't consistent and cannot be duplicated easily. We need to find a way to design an implementable algorithm that could do this consistently and quickly."

Casazza said that there are already programs that can separate and reconstruct voices, but they are not completely reliable. For example, such programs have difficulty separating voices with similar pitch characteristics. A program using the researchers' solution would be more exact.

Source: University of Missouri

Explore further: Math journal puts Rauzy fractcal image on the cover

add to favorites email to friend print save as pdf

Related Stories

We are all made of stars

Sep 02, 2014

Astronomers spend most of their time contemplating the universe, quite comfortable in the knowledge that we are just a speck among billions of planets, stars and galaxies. But last week, the Australian astronomical ...

WEF unveils 'crowdsourcing' push on how to run the Web

Aug 28, 2014

The World Economic Forum unveiled a project on Thursday aimed at connecting governments, businesses, academia, technicians and civil society worldwide to brainstorm the best ways to govern the Internet.

Lenovo's smart glasses prototype has battery at neck

Jul 28, 2014

China's PC giant Lenovo last week offered a peek at its Google Glass-competing smart glass prototype, further details of which are to be announced in October. Lenovo's glasses prototype is not an extreme ...

Recommended for you

King Richard III died painfully on battlefield

2 hours ago

England's King Richard III might well have lost his kingdom for a horse. The reviled king suffered nearly a dozen injuries on the battlefield, but the fatal blows were probably only sustained after he had to abandon his horse, ...

History books spark latest Texas classroom battle

6 hours ago

As Texas mulls new history textbooks for its 5-plus million public school students, some academics are decrying lessons they say exaggerate the influence of Christian values on America's Founding Fathers.

Flatow, 'Science Friday' settle claims over grant

6 hours ago

Federal prosecutors say radio host Ira Flatow and his "Science Friday" show that airs on many National Public Radio stations have settled civil claims that they misused money from a nearly $1 million federal ...

'Hidden Treasure of Rome' project unveiled

7 hours ago

For more than a century, hundreds of thousands of historical artifacts dating back to before the founding of Rome have been stored in crates in the Capitoline Museums of Rome, where they have remained mostly untouched. Now, ...

Poverty rate drops for the first time since 2006

8 hours ago

The poverty rate in the United States has dropped for the first time since 2006, bringing a bit of encouraging news about the nation's economy as President Barack Obama and Congress gear up for the November elections.

User comments : 0