Digital hearing-aid signal processing method

Abstract

The invention relates to a digital hearing-aid signal processing method. The digital hearing-aid signal processing method comprises the steps: an audio digital signal is obtained; an analysis filter bank divides the audio digital signal into sub-band signals of multiple channels; a detection module detects sound pressure level data and noise level data of the channels; a calculation module calculates a noise reduction gain value and a compressor amplification gain value of each channel signal; a multiplier performs product on the noise reduction gain value and the compressor amplification gain value of each channel signal and the sub-band signals of multiple channels to obtain product data of the channels; a synthetic filter synthetizes the product data of the channels to form output signals. The digital hearing-aid signal processing method performs noise reduction and compressor amplification on the signals based on an auditory perception model, overcomes the problem that real time realization cannot be performed and phase distortion exists in the prior art, and can achieve human ear frequency resolution by using few channels and further improve the listening comfort level of a patient.

Description

technical field [0001] The invention relates to speech digital signal processing technology, in particular to a digital hearing aid signal processing method. Background technique [0002] For hearing loss, especially sensorineural hearing loss, in addition to the increase in the hearing threshold, the discomfort threshold of most patients remains unchanged or even decreases, resulting in a decrease in the wide dynamic range of hearing. In response to this feature, digital hearing aids usually use Wide Dynamic Range Compression (WDRC), which compresses the wide dynamic range of the input signal to the residual hearing dynamic range of the hearing impaired according to a certain ratio. In addition, hearing-impaired people may have the same perception of sound in a certain frequency range as ordinary people, but they cannot hear sounds with lower sound pressure levels in other frequencies like ordinary people. The way to solve this problem is to carry out multi-channel wide dyn...

AI Technical Summary

Problems solved by technology

However, this method divides the signal based on the WOLA filter bank, and the WOLA filter bank is a uniform filter bank, and its frequency resolution is not suitable for the human auditory system, because the frequency resolution of the human ear is non-uniform, As the frequency increases and decreases, there is a nonlinear relationship between the actual perceived frequency scale (auditory frequency scale) and the common frequency scale, which can be expressed by the Bark frequency scale

Using a uniform filter bank, to make the frequency resolution reach the Bark frequency scale, the channel division is required to be very fine, and this will cause the channel division of the high frequency part to be too fine, wasting resources

Images