| Summary: | The speech-to-noise ratio required under noisy conditions so that intelligibility is comparable
to that in quiet is significantly higher for a hearing-impaired individual than a normal-hearing
person. In this research a simple, efficient, real-time implementation of a single-input speechenhancement
scheme, designed to increase the intelligibility of speech corrupted by additive noise
using adaptive noise cancellation (ANC) techniques was evaluated with speech-discrimination tests
using hearing-impaired subjects. Although many single-input speech-enhancement schemes have
been shown to provide insignificant gains for the normal-hearing person, this work found that a
hearing-impaired person’s ability to understand the speech processed using ANC can improve
noticeably, as compared to results obtained with unprocessed speech.
The algorithm takes advantage of a novel speech-detection algorithm recently developed at
the University of British Columbia to classify individual signal segments as silence or speech; the
coefficients of an adaptive filter are updated during signal periods classified as ‘silence’. Two
variants of the least-mean square (LMS) algorithm, the leaky- and normalized-LMS, are integrated
in the weight-update process to produce a noise-reduction scheme which adapts itself to the power
of the input signal and also corrects for ill-conditioned inputs. The resulting speech-enhancement
scheme is most effective with narrowband, quasi-stationary noises and robust under changing
noise conditions. Although often used individually, to our knowledge the incorporation of both
LMS variants into a single algorithm is unique to this research.
Speech Perception in Noise (SPIN) tests were used on eight hearing-impaired subjects
to evaluate the algorithm’s performance. Two types of noise background were used. Of the
four subjects who heard motor noise, raw scores increased by 20 percent for the enhanced
speech. These results were shown to be statistically significant using a two-way analysis of
variance with repeated measures. However for those who heard babble or ‘cafeteria’ noise, the
speech-enhancement process evidenced little improvement over scores from unprocessed speech.
Further clinical evaluation is necessary, with a larger sample base, since the limited size of the
study sample precluded assessing the algorithm with respect to specific subject factors such as
degree or type of hearing loss. === Applied Science, Faculty of === Electrical and Computer Engineering, Department of === Graduate
|
|---|
