How great it would be to be able to talk to someone on a cell phone without having to listen to background noise? This is exactly what Dr. Parham Aarabi, assistant professor in the Department of Electrical and Computer Engineering and founder and director of the Artificial Perception Laboratory, is trying to achieve.
Aarabi began thinking about this problem while working on ways to remove noise from computers. Now, after seven years of work, Aarabi and his colleagues David Halupka, Guangji Shi and Dr. Ali Sheikhaleslami have developed a dual microphone chip that has the ability to remove all kinds of background noise from phone conversations. Aarabi explains that this system involves two microphones and functions in a similar manner to human ears, which focus on a specific person’s voice while filtering out background noise.
If you’re trying to talk to somebody at a very loud party, your ears have the ability to concentrate on one source of sound (e.g. somebody’s voice) and ignore all the other noise around you. Aarabi has built this same ability into a small chip, which will hopefully be used in cellular phones. Background noise is converted into a quiet hum so that it isn’t detectable and doesn’t lead to confusion during conversations. Most other systems that have been developed just attenuate background noise without really tackling it, resulting in often garbled and confused conversations.
This system “can perform sound localization and speech enhancement in real-time, while consuming extremely low power,” explains Aarabi. He plans on making this energy-efficient chip small enough and so that it can be attached to cell phones, and in the future perhaps even built right into phones by the manufacturer. This new technology may be ready for public use in about two years.
In addition to developing this dual microphone chip the Artificial Perception Laboratory is working on several other projects. One of the most notable is a program that has the ability to quickly search and sort through many different images. Aarabi explains that this could be very useful for security systems in airports and casinos, for example. A user would be able to select and sort images of an object or person they would like to investigate according to the best view available in just a few quick user clicks.
