picking the signal from PSD

i have a signal and it is corrupted with noise !!! so how can i separate the signal frequency component from the noise using PSD .........i already know the main theme but the thing that is confusing is ...with the plotting of PSD ....how does it works ??i mean the plot(f,pxx)....how the f along the x axis is defined here can anyone help me out of this ......for ur kind information um a newbie in DSP....so pls explain me as easily as possible!

Answer 1

[Wayne King]

The PSD shows how the power in the signal is distributed as a function of frequency. The x-axis, f, are the frequencies. If you use the sampling frequency as an input to the PSD routine you are using, then f is in cycles/sec.

How you separate noise from signal depends on what your signal and what your noise are like. If your signal is sinusoidal, it shows up as peaks in the PSD estimate.

[newbie traveller]

thank you sir very much ...but i want to know how the f along the x axis is defined ...more precisely u have to define f by a row vector right ???i just want to know tht part only!!

how the f is vectorize along the x axis of the psd plot??--------i am expecting an answer from you sir!

Answer 2

[Wayne King]

If you specify a sampling frequency, then f is in increments of Fs/N where Fs is the sampling frequency and N is the length of the signal

    Fs = 1000;
    t = 0:0.001:1-0.001;
    x = cos(2*pi*100*t)+randn(size(t));
    [Pxx,F] = periodogram(x,[],length(x),Fs);

Notice that in this case, Fs = 1000 and N = 1000 also, so we would expect the frequency vector, F, to increment by Fs/N=1

If you enter

    F(2)-F(1)

Or more generally

   plot(diff(F))

You'll see that is the case.