【语音识别】基于MFCC和SVM的特定人性别识别matlab源码

clc
clear all
close all
 
t=0:0.01:pi;
%%%%测试结果是幅值还原的不太对，频率的分离还可以
%%%%可以调用
x1=10*sin(100*t/(2*pi));
x2=10*sin(400*t/(2*pi));
x3=5*sin(600*t/(2*pi));
figure(1)
for flag=1:6
    x=x1+x2+x3;
    [Am_d1 Fm_d]=lmd(x,flag);
    
    subplot(2,3,flag)
    hold on
    plot(x)
    plot(Fm_d,'r')
    x=x-Am_d1.*Fm_d;
    [Am_d2 Fm_d]=lmd(x,flag);%%%%第二阶
    hold on
    plot(Am_d2.*Fm_d,'g')
    x=x-Am_d2.*Fm_d;
    [Am_d3 Fm_d]=lmd(x,flag);%%%%第三阶
    plot(Am_d3.*Fm_d,'k')
    legend('原始信号','第一阶分解信号','第二阶分解信号','第三阶分解信号')
    switch flag
        case 2%	%极值平均
                       
            psnr1 = PSNR(x1, Am_d1);
            psnr2 = PSNR(x2, Am_d2);
            psnr3 = PSNR(x3, Am_d3);
%             disp(['极值平均端点延拓分解信号平均信噪比psnr=',num2str( psnr1+ psnr2+ psnr3)])
            title(['极值平均端点延拓'])
        case 3
            psnr1 = PSNR(x1, Am_d1);
            psnr2 = PSNR(x2, Am_d2);
            psnr3 = PSNR(x3, Am_d3);
%             disp(['边界局部特征尺度延拓法端点延拓分解信号平均信噪比psnr=',num2str( psnr1+ psnr2+ psnr3)])
            title(['边界局部特征尺度延拓法端点延拓'])
        case 4
              psnr1 = PSNR(x1, Am_d1);
            psnr2 = PSNR(x2, Am_d2);
            psnr3 = PSNR(x3, Am_d3);
%             disp(['极值延拓法端点延拓分解信号平均信噪比'])
        
            title(['极值延拓法端点延拓'])
        case 5
               psnr1 = PSNR(x1, Am_d1);
            psnr2 = PSNR(x2, Am_d2);
            psnr3 = PSNR(x3, Am_d3);
%             disp(['基于ISBM延拓端点延拓分解信号平均信噪比'])
        
            title(['基于ISBM延拓端点延拓'])
        case 6
               psnr1 = PSNR(x1, Am_d1);
            psnr2 = PSNR(x2, Am_d2);
            psnr3 = PSNR(x3, Am_d3);
%             disp(['平行延拓法端点延拓分解信号平均信噪比'])
        
            title(['平行延拓法端点延拓'])
        case 1
             psnr1 = PSNR(x1, Am_d1);
            psnr2 = PSNR(x2, Am_d2);
            psnr3 = PSNR(x3, Am_d3);
%             disp(['改进前分解信号平均信噪比'])
        
            title(['改进前'])
    end
    
    hold off
end