MATLAB: How to find a less-correlated part of two signals

correlationMATLABnoisesignal processing

Greetings,

Say, I have two signals, s1 and s2, where s2 is simply a "noisier" version of s1. What would be the best way to find the less-correlated portion (due to high noise level) between s1 and s2?

Thanks.

Best Answer

The Signal Processing Toolbox has two recent functions that could help you: alignsignals and findsignal.

Related Solutions

MATLAB: Signal Normalization for cross correlation

Hi Mustafa,

Since in your case you are doing a cross-correlation between a signal and its own shifted version, to simplify things, you would just need to ensure that the signal and its shifted version have the same scaling. To ensure this, you can use something like:

% s1 = some signal
% s2 = time shifted s1
s1 = s1/max(abs(s1));
s2 = s2/max(abs(s2));
scorr = xcorr(s1,s2);

You can also see the examples in the documentation for xcorr .

Hope this helps,

Cheers!

Abhiram

MATLAB: How to use cross-correlation as a tool to align two signals in MATLAB

Correlation reaches a maximum value at the point where two signals are maximally correlated. You may therefore use this value to align them. The following code provides an example of how this can be done using the XCORR function in the Signal Processing Toolbox (see attached file variable_file.mat):

load variable_file
X1=xcorr(s1,s2);   %compute cross-correlation between vectors s1 and s2
[m,d]=max(X1);      %find value and index of maximum value of cross-correlation amplitude
delay=d-max(length(s1),length(s2));   %shift index d, as length(X1)=2*N-1; where N is the length of the signals
figure,plot(s1)                                     %Plot signal s1
hold,plot([delay+1:length(s2)+delay],s2,'r');   %Delay signal s2 by delay in order to align them
grid on

Best Answer

Related Solutions

MATLAB: Signal Normalization for cross correlation

MATLAB: How to use cross-correlation as a tool to align two signals in MATLAB

Related Question