Research On Speech Enhancement In Digital Hearing Aids

Due to the aging of population, environmental noise pollution and some other factors, more and more people's auditory system is impaired. Wearing hearing aids is a most effective method for hearing therapy of impaired person. But the noise pollution problems seriously affect the practical effects of hearing aids. And compared with normal hearing person, the patients with hearing loss have lower speech comprehension in the noisy environment. Speech enhancement technology can effectively eliminate the background noise and improve voice quality, which effectively improve patients' speech comprehension in the noisy environment. Based on deep understanding of existing algorithms, this paper mainly studies the speech enhancement technology for digital hearing aids. The main work can be show as follows:(1) This paper introduces the Wiener Filtering speech enhancement technology, focuses on the speech enhancement based on a priori SNR estimation Wiener Filtering. The method based on priori SNR estimation Wiener Filtering can not effectively improve the speech intelligibility. Some research is done to solve this disadvantage. Firstly, this paper uses the MMSE-based two-step a priori SNR estimate approach to replace the "decision directed" method. Secondly, the gain function is modified because of the problem that the gain function overestimation errors in frequency bins with negative SNR are particularly harmful to speech intelligibility. The comparative experiments with the original method demonstrate the effectiveness of the improved method.(2) In order to reducing the computational complexity of the method based on priori SNR estimation Wiener Filtering, a low computational complexity of sub-band speech enhancement algorithm in digital hearing aids is proposed. The principle and implementation of this algorithm are introduced in detail. This algorithm makes use of calculating signal power in each sub-band to replace the calculation of the power spectrum, so that the transformation between time and frequency domain is omitted, which make the computational complexity of the algorithm greatly reduced. The effectiveness of the proposed algorithm is validated by the simulation experiments.(3) The speech enhancement method based on directional microphones is studied. And this paper focuses on the principle and structure of first-order adaptive directional microphone. There is still some residual noise after enhancing by directional microphone, while sub-band speech enhancement algorithm will cause larger voice distortion when processing signal with low SNR. To address this problem, the two-step speech enhancement which combine with directional microphone and sub-band speech enhancement is proposed. Firstly, directional microphone is used to enhance the voice, and then the sub-band speech enhancement method is used for the further proceesing. The combination of two methods has offset their disadvantage effectively, which finally make the enhanced signal more qualified. The simulation results verify the effectiveness of the proposed algorithm.