How to attenuate the signal in frequency domain without disturbing the noise floor?

3 次查看（过去 30 天）

Praveen kumar 2014-2-20

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/116970-how-to-attenuate-the-signal-in-frequency-domain-without-disturbing-the-noise-floor

编辑： Wayne King 2014-2-20

I have written a code to calculate the power spectrum of a 200Hz sinusoidal signal. In frequency domain, the strength of the 200Hz component is 24dbm and the noise floor is at -15dbm. How to attenuate the signal with out affecting the noise floor at -15dbm?

close all;
clear all;
clc;
fs = 1e3;                      % sample freq
f = 2e2;
D = 0.3;                         % pulse delay times
StopTime = 1;
t = (0 : 1/fs : StopTime-1/fs)';         % signal evaluation time
N = size(t,1);
w = 0.2;                      % width of each pulse
yp = pulstran(t,D,@rectpuls,w);
x = (10)*sin(2*pi*f*t);
display(size(yp));
display(size(x));
z = yp.*x;
noisy_z = z + 0.5*randn(size(z));
y = xcorr(noisy_z,'biased');
   display(size(y));
   subplot(2,1,1);
   plot(t*(1/(f*StopTime)),z);
   title('Pulsed sine signal');
   subplot(2,1,2);
   plot(t*(1/(f*StopTime)),noisy_z);
   title('Noisy signal');
   l = length(t);
     %%Fourier Transform:
     X1 = fft(y,l);
     X2 = X1(1:l/2);
     X = abs(X2)/N;
     %%Frequency specifications:
     dF = fs/N;                      % hertz
     f1 = 0:dF:fs/2-dF;           % hertz
     display(size(f1));
     display(size(X));
     %%Plot the spectrum:
     figure;
     plot(f1,(10*log10((X).^2*1e3)));
     xlabel('Frequency (in hertz)');
     ylabel('Power (in dbm)');
     title('Autocorrelation spectrum');
     grid on;

1 个评论
显示 -1更早的评论隐藏 -1更早的评论

Wayne King 2014-2-20

编辑：Wayne King 2014-2-20

在 MATLAB Online 中打开

Well, right out of the gate, how are you doing this?

fs = 1e3;                      % sample freq
f = 2e2;

A sinusoid with a period of 1/2000 sampled at 1/1000 is going to be aliased to be DC (all zeros for sin with zero phase and all ones for a cosine with zero phase.)

请先登录，再进行评论。

请先登录，再回答此问题。