Finding the main frequency after fft

Hi,

I've got these blocks in simulink:

Sine Wave -> Buffer -> Matrix sum -> Matlab function.

I have to find the frequency of the input, so I wrote some code in Matlab function (by the http://www.mathworks.com/help/techdoc/ref/fft.html) :

{

function [freq, amp] = fcn(u)

%#codegen

Fs=2000;

T=1/Fs;

L=length(u);

t = (0:L-1)*T;

NFFT = 2^nextpow2(L); % Next power of 2 from length of y

Y = fft(u,NFFT);

f = Fs/2*linspace(0,1,NFFT/2+1);

amp = abs(max(Y));

}

It's working properly, I'm getting the maximum value, but I can't get from code frequency of maxium value :/ how Can I get it? Thank You for response.

Best regards, Pawel

Answer 1

[Wayne King]

Hi, You have to find the corresponding index in a frequency vector. For example:

    Fs = 1e3;
    t = 0:.001:1-0.001;
    x = cos(2*pi*100*t)+randn(size(t));
    xdft = fft(x);
    [~,index] = max(abs(xdft(1:length(x)/2+1)));
    freq = 0:(Fs/length(x)):Fs/2;

    % freq(index)
    fprintf('Maximum occurs at %2.3f Hz\n',freq(index))

Wayne

Answer 2

[Pawel Burzynski]

It's excatly what I was looking for :) I had problems with those way of notation. I'm getting some mismatch, but I can handle this.

Thank You Wayne very much! :)

[Wayne King]

Glad it's working. Your mismatch is likely due to your frequencies not falling directly on a DFT bin.

Wayne

[Pawel Burzynski]

I've got the same problem, mainly i can't get the real frequency of the input event if it is pure sinusoidal source... There is always +/- 15% error. Can someone can help me?

Best regards

[Wayne King]

Can you please provide code? It is hard to say the source of the error without a coded example.