Research Of Adaptive Echo Cancellation Technology In Digital Heating Aid

Digital hearing aid helps patients with hearing impairment improve hearing, but as the volume of hearing aid becomes smaller gradually, and the ‘open ear’ fitting mode is more and more popular, echo phenomenon is more and more evident. Echo seriously affects the speech intelligibility and articulation. In order to eliminate echo interference on the speech signals, in this paper, adaptive echo cancellation algorithm was researched.After deeply understanding traditional adaptive algorithm, completed variable step size LMS(VSS_LMS) algorithm, the purpose was to solve the problem of slow convergence in normalized LMS(NLMS) algorithm.This paper described the design of loop detector, which was used for detecting the existing of the echo and controling the echo cancellation module, saved the system resource. The presence of environmental noise, the nonlinearity of A/D and other components lead some residual echo left in the speech signal after echo processing. In order to reduce the residual, the design increased the nonlinear processing module. This paper compared the VSS_LMS with the NLMS algorithm, from the objective judgment, variable step size algorithm has faster convergence speed and smaller echo residue, from the subjective judgment, variable step size algorithm has better clarity and comfort.NBLMS(normalized block LMS) algorithm effectively changes the point by point way to update the filter coefficients in NLMS and RLS algorithm, solved the problem of wasting system resources, but the NBLMS algorithm has low convergence, large residue, the poor sound quality, this paper made some improvement on the NBLMS algorithm, NBLMS_M-K algorithm was proposed.After the implementation of the NBLMS_M-K algorithm, completed the simulation on PC. Based on the results of subjective and objective judgments, compared to NBLMS algorithm, NBLMS_M-K algorithm has high echo processing ability, better performance.