Real-Time Implementation of an Efficient Speech Enhancement Algorithm for Digital Hearing Aids

Gao Jie,Zhang Hui,Hu Guangshu
DOI: https://doi.org/10.1016/s1007-0214(06)70220-6
2006-01-01
Abstract:In order to remove background noise and improve the quality of speech for digital hearing aids, a single-channel speech enhancement algorithm is proposed. The algorithm is implemented and assessed on a digital hearing aid platform based on the TI DSP TMS320VC5502 chip. Assuming that background noise is stationary or varies slowly, an energy-based voice activity detection algorithm is adopted by adaptively tracking the minima and maxima of the power envelope in noisy speech. The target speech is then enhanced by using a Wiener filter, on the basis of a short-term power spectral estimation. To deal with the distracting musical noise of the processed speech, phase randomization, along with adjacent spectral averaging, is adopted. Objective measures and an informal hearing test both show an improved performance as well as obvious attenuation of residual noise. The low power consumption and high efficiency render the whole algorithm very applicable for use in digital hearing aids.
What problem does this paper attempt to address?