Noise signals can be sorted by computers
Hear, hear: new technology improves hearing aids
A digital hearing aid that filters out background noise and focuses on key audio signals is being refined at the University of Guelph.
Prof. Bob Dony, School of Engineering, and masters student Edward Chau are using computer technology to better understand speech and audio signal processing, and improve digital-based hearing aids. Digital hearing aids are more powerful and adaptable to real-life situations than most standard hearing aids because they're outfitted with a special microchip - a mini computer - that can be programmed to improve sound enhancement and corrections to incoming signals.
"Noisy environments are not comfortable for most hearing aid wearers," says Chau. "There's a lot of background noise like speech babble that make it harder for hearing aid users to understand what's going on. The digital hearing aid can be programmed to get rid of some of that noise."
Dony and Chau are investigating ways to reduce background noise.
They're also finding ways to help hearing aids focus on one person's speech. The microchip they're using was developed specifically for low-power, miniature applications - such as hearing aids - by dpsfactory, a Waterloo company specializing in digital signal processing.
The chip can be programmed to manipulate what is being heard to make it sound better. But before the microchip can be programmed, the processes involved in the hearing and sorting of speech signals must be intricately understood.
To compute and analyze audio signals, Dony and Chau are using artificial neural networks that mimic the structure of the brain. These artificial networks are loosely based on how neurons (brain cells) work and can provide insight into the brain's neuron activity on a smaller level when processing signals.
The researchers hope that insight can be used to develop better methods for processing audio signals.
"If we can understand the intelligence of how the brain naturally processes signals, then we can come up with a better computer-based system for sorting speech signals," says Dony.
By using speech and audio signals recorded from a microphone or hearing aid configuration, the researchers are evaluating different speech processes. Their experiments involve several microphones and two computers, one of which produces audio signals with different noise scenarios while the other analyzes and processes the microphone signals.
The information obtained from the processing of these various audio signals will be used as the basis for programming the microchip so that it will hear and process sounds in a similar manner.
Other applications based on programmable microchips include hearing aids that can pick up and process sounds originating from different directions, and portable voice recognition devices.
This research is sponsored by the Natural Sciences and Engineering Research Council and dspfactory.
SPARKplugs articles appear every Sunday. They are produced by student writers in the Office of Research, University of Guelph.
University of Guelph
Digital Sound - Edward Chau, left, and Bob Dony are using computer technology to improve digital-based hearing aids.
Length: Medium, 415 words
© 1999 Guelph Mercury. All rights reserved.
Doc. : news·19991226·GU·0024
|This material is
copyrighted. All rights reserved.|
© 2001 CEDROM-SNi