The algorithm of noise subtraction for cochlear implant based on improved auditory perception wavelet packet
jingru huang,guanyu tian,guoqiang qi,shuzhong bai,shibin du,lan tian
DOI: https://doi.org/10.2991/iccnce.2013.77
2013-01-01
Abstract:Under the noisy background environment, the performance of cochlear implant (CI) will decrease rapidly. Wavelet transform is the effective analyzing tool for the complicated speech signal. But the traditional wavelet transform has the disadvantage of 'over threshold' that leads some effective speech parts lost in the enhanced speech. In order to raise the anti-noise ability of CI and retain richer and clearer speech signal, an improved speech enhancement algorithm is presented and applied into CI. In this method, by using the perceptual wavelet packet transform, the noisy speech signal was decomposed into wavelet packet node coefficients and the time adaptive threshold (TAT) was introduced in the process of noise subtraction. For each frame signal, based on the estimated speech-presence probability, the TAT of subtracting noise was adjusted along with time so that the noise signal can be effectively suppressed while much more useful speech component was retained in the recovered signal, especially for the endpoint parts of speech signal. Based on the ACE processing strategy, the enhanced speech signal was analyzed and synthesized in the simulation system of CI. The simulation experiment results show that the proposed algorithm can produce clearer and better synthesized speech in CI than that of the traditional wavelet transform algorithm.