Normalized sub-band adaptive echo elimination method based on M estimation

Abstract

The invention relates to a normalized sub-band adaptive echo elimination method based on M estimation. The method comprises steps that A, a sub-band signal is acquired; B, adaptive filtering, an mth filter output value ym(k) at the sampling time n=kM is acquired through an adaptive filter on the basis of an mth adaptive filter input vector Xm(k) at the sampling time n=kM, and an equation described as the specification is acquired; C, echo cancellation, a kth time segment extraction value d<->m(k) of an mth near-end sub-band signal subtracts the mth filter output value ym(k) at the sampling time n=kM to acquire an mth sub-band residual error em(k) at the sampling time n=kM, and an equation described as the specification is acquired; D, a filter tap weight vector is updated, a normalized sub-band method based on M estimation is utilized, a tap weight vector w(k+1) of the adaptive filter at the time n=(k+1)M is acquired, and an equation described as the specification is acquired; and E, k=k+1 is made, the steps of B, C and D are repeated till conversation is over. Through the method, the acoustic echo elimination effect for a communication system is good, a convergence speed is fast, and a stable state error is small.

Description

technical field [0001] The invention relates to a normalized subband adaptive echo cancellation method based on M estimation, and belongs to the technical field of communication echo cancellation. Background technique [0002] Today's video and teleconferencing under the full hands-free function mode has become the most effective way of communication. In this full hands-free function mode, since there is no good separation between the speaker and the microphone, the sound output from the speaker is picked up by the microphone directly or reflected by obstacles (such as walls, ceilings, etc.), and transmitted through the system. Back to the remote user, so that the remote user can hear his voice again, we call this sound echo. In addition, for car drivers, a completely hands-free car phone system will bring them more convenience. But in the car, the external environment changes greatly, which makes the acoustic coupling between the speaker and the microphone more serious, s...

AI Technical Summary

Problems solved by technology

[0008] Reference 2 "Novel sign subband adaptive filter algorithms within individual weighting factors" (Y.Yu and H.Zhao, Signal Process., vol.122, pp.14–23, 2016.) uses the same method as the traditional sign subband method Different weighting factors, this method uses different weighting factors for different sub-bands, which further improves the convergence speed of the algorithm, but the convergence speed of this method is still slow

Images