Cochlear Implant Recipients’s Ability Analyzed Of Speech Recognition And Voice

Cochlear implant (CI) is the only medical prosthesis which can help severe deafpeople to regain the hearing currently. With the rapid development of integrated circuitand speech signal processing technology, the performance of CI was also graduallyimproved and perfected. More and more hearing loss patients get better auditory qualitywith the help of CI. CI is a device to replace the feeling of damaged hair cells.The CIprocesses the speech signal first, then it send the transmitting electrical stimulations tothe implanted electrodes.In recent years Chinese hearing-impaired people also began tobenefit from cochlear along with the economic development of China. Although thewestern countries had a mature theoretical basis and clinical experience in the cochlearimplant research already, it’s not suitable for our country. the pitch has a very importantinfluence for the semantic. So it is more difficult to process and coding the speechsignal in China then western developed countries. Another unavoidable interference forthe CI patients’ speech perception in daily life is the noise. Above all, study in theinfluence factors of the mandarin speech intelligibility in CI and the ability of thepatients’ production is very important for the advantage and improvement of CI in ourcountry. This paper the influence factors of mandarin speech sentence and pitchrecognition, so as the pitch detection algorithm. The pitch information is a necessaryreference for the tone analyses.First, the present study discussed the affecting factors of cochlear speechrecognition in the noisy environment by a simulation experiment. Let the speech signalplus to the noise in different types, first add the different frequency noise with the+10db SNR. The second is the inclusion all kinds of SNR white Gaussian noise.,letpatients have a listening training with these two synthetic speech, Analysis of theexperimental results then get the result.As the primary means of tone characteristic discriminant, the fundamentalfrequency detection algorithm was improved. This paper presented one improved pitchdetection method based on classic fundamental frequency extraction method. Themethod combined the average magnitude difference function (AMDF) with the autocorrelation function (ACF), so that the peak point at the fundamental frequency cycle turned more prominent and sharper to improve the noise immunity,then evaluate thevocal ability of cochlear implants recipients. Evaluation method is to extract theparameters of patients. Conclusion cochlear implant pronunciation characteristics andthe insufficiency. Then make the rehabilitation training and implement.At last, analyze the cochlear platform of Australian. Mainly analyzes thecomposition of the hardware and the use of the Matlab toolbox. And sums up theadvantages and Rehabilitation training plan and implement of the platform. Real-timeand simplicity is his greatest characteristics. In the future will develop into a portabledevice, Can be used by patients in the home. Make the test scenarios is the real lifeenvironment. Can provide more accurate data base for the new algorithm proposed.