The Research On Speech Enhancement And Loudness Compensation Of Digital Hearing Aids

Due to the existence of a variety of noise sources such as reverberation and reverberation, the performance of digital hearing aids is affected, so it is very important to suppress noise and improve speech intelligibility. This paper mainly analyzes the key technology in digital hearing aid, and focuses on the microphone array speech enhancement algorithm and multichannel loudness compensation algorithm, based on proposed some improvement and optimization.The main work of this paper is as follows:Firstly, the basic theory of speech signal and the working principle of digital hearing aids are studied, and the principle of speech sound is analyzed, and the mathematical model of speech signal is analyzed. On Key Technology in digital hearing aids, such as sound source localization, speech enhancement, loudness compensation and echo elimination were studied, and analyzes the their application in digital hearing aids.Secondly, the speech enhancement algorithm based on the microphone array is studied deeply.The generalized sidelobe canceller(GSC) algorithm analysis, it is found that the method will cause the speech leakage and residual noise, on the basis of this, put forward a kind of improved GSC structure method, using a voice activity detection(VAD) will be fixed beam forming for generating the speech into the frame of speech and noise and to enhance the speech frame, to attenuate the noise frame, while protecting the formant. Finally, the other modules of the speech signal input GSC for subsequent processing.Finally, in-depth study of loudness compensation algorithm. The commonly used methods of loudness compensation using the amplification and compression technology, but this method may make part of the high strength in the process of amplification and compression becomes too large,thus affecting the effectiveness of loudness compensation. In this paper, wide dynamic compression technique is applied to loudness compensation algorithm, the high-frequency part mapping to listen to the range of hearing loss in patients with, and use of wavelet threshold denoising technique is used to remove noise, further helping patients to improve the speech recognition rate.