How to plot the coherence function?

For my two signals, x and y, I have calculated the auto- and cross-spectral density functions from the fft. So my spectral densities are Sx, Sy and Sxy.

I want to find the coherence-function, but when I just enter the formula for the coherence function directly i a get a problem with the matrix dimensions.

coh=abs(Sxy)/(Sx*Sy)

Iam sure there is a much more elegant way to do it? any suggestions? I want the unit of the frequency axis to be in rad/s.

Andre

Answer 1

[Wayne King]

If you have the Signal Processing Toolbox, use mscohere()

n = 0:599;
x = cos(pi/4*n)+randn(size(n));
y = 1/2*sin(pi/4*n)+randn(size(n));
[Cxy,W] = mscohere(x,y,160,120,160);
plot(W,Cxy)

[Andre]

So W is in rad/s?

[Wayne King]

It's in radians/sample. If you want it in radians per second, then put in the sampling frequency as an input to mscohere() and then multiple each element of the frequency vector by 2*pi

    Fs = 1000;
    t = 0:0.001:1-0.001;
    x = cos(2*pi*100*t)+randn(size(t));
    y = sin(2*pi*100*t)+randn(size(t));
    [Cxy,F] = mscohere(x,y,200,150,200,Fs);
    plot(F.*(2*pi),Cxy);
    xlabel('Radians/Second')

[Csanad Levente Balogh]

Hi. Dou you have a solution for if you don't have the Signal Processing Toolbox?