| Elderly sensorineural deafness has become a worldwide problem, and being equippedwith a pair of digital hearing aids is the best solution for these deaf patients.With the healthproblems of deafness and hearing being focused on widely in society, the various functions ofadvanced digital hearing aids have broad application prospects.It was a fundamental objectiveto improve the auditory comfort and enhance the rate of speech recognition. A variety oftargeted speech processing algorithms that it uses will help the wearer significantly improvethe level of hearing.It brought gospel to deaf patients that they would recover normal verbalcommunication. However, whether the voice processing algorithms which digital hearing aidsrely on has a optimized parameter and a stable performance or not, will directly affect thefurther enhance of function. Therefore,the key to researching digital hearing aids that hearingaids algorithm and its system implementation, has more important social significance andapplication values.This article focused on digital hearing aid loudness compensation algorithm and itssystem implementation. Starting with human auditory physiological characteristics， itdeveloped a intelligent loudness compensation algorithm which accurately fitted with injuredears’ hearing loss curve, and designed software and hardware DSP system solutions foralgorithm implementation. The main jobs of this thesis are as followed:1.Elaborating physiological functions of the human auditory system and analyzing themanifestations of the hearing loss. And mentioned the scientific method that the technologicaladvances of the hearing aids provided more and more convenient to the hearing recovery ofhearing-impaired patients.The analysis and summary of the various voice processingalgorithm principle applied to digital hearing aids are made in detail further.2.Understanding the basic characteristics of the speech signal in depth, analysing traits ofhearing loss in the frequency domain of elderly sensorineural hearing-impaired patients in detail. Elaborating that voice input sound intensity in which the frequency of the differentvalues is the key factor affects the correctness of the calculated gain compensationfactors.According to the distribution of pure tone hearing test data in the frequency domain,summarizing algorithm principle and implementation method of wide dynamic compressionloudness compensation. An algorithm improved way based on formant estimation wasproposed,which avoided voice nonlinear distortion caused by neglecting the formants whenspectral gains,improved the auditory comfort after the application of loudness compensationalgorithm.3.Ultilizing powerful algorithms simulation capabilities of Matlab, simulation validatedthe input and output voice data loudness compensation algorithm step-by-step in detail.Inaddition, analysing and comparing carefully the data over the same period of the twoalgorithms in implementing voice loudness compensation. Simulation results show that: Theimproved algorithm based on formant estimation, on the basis of real-time compensation ofpatients’hearing loss of loudness in the frequency domain, effectively reduced the nonlineardistortion of the formants;4.From two aspects of system hardware and software,carrying out detailed discussions ofdigital hearing aids system implementation respectively. And the key links in the systemdesign process and the treatment of the specific details will be elaborated... |
PDF Full Text Request