how can i write a matlab code for lms algorithm for audio signal?

2 次查看（过去 30 天）

lipsa tripathy 2017-2-26

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/326970-how-can-i-write-a-matlab-code-for-lms-algorithm-for-audio-signal

评论： Walter Roberson 2025-5-18

clc 
clear all
close all
[x,Fs] = audioread('test11.wav');
sound(x,Fs);
pause(10);
delay1 = 0.6; % delay
delay2=0.9;
delay3=0.6;
alpha1 = 0.65;% echo strength
alpha2=.40;
alpha3=.5; 
t=0:1/Fs:(length(x)-1)/Fs;
subplot(3,2,1)
plot(t,x)
title('original signal')
D1 = delay1*Fs; 
D2=delay2*Fs;
D3=delay3*Fs; 
y = zeros(size(x));
y(1:D1) = x(1:D1);
snr=4;
for i=D1+40000:length(x) 
    y(i) = x(i)+alpha1*x(i-D1)+alpha2*x(i-D2)+alpha3*x(i-D3);
 end 
   sound(y,Fs);
   t1=0:1/Fs:(length(y)-1)/Fs;
   subplot(3,2,2)
   plot(t1,y)
   title('original+echo');
   % ----------------------------------------------------------------- 
   load nearspeech
   load farspeech
   N=100000;
   L=1000;
   mu=0.00003;
   h1=fir1(2*L-1,0.5); 
   h=10*h1(L+1:2*L);
   % ---------------------------------------------------------------------
   wlms=zeros(L,1); 
   xlms=zeros(L,1); 
   for m=1:N
    xlms=[ y(m);xlms(1:L-1) ];
    ylms= h * xlms;
    elms(m)= d(m) - ylms;
    wlms= wlms + (mu * xlms * elms(m));
   end
   subplot(3,2,4)
   plot(elms,'m')
   title('Recovered Signal')

here only the first element of y is passing not the whole y signal.how can i make the whole y signal pass through the filter without any dimension missmatch?