Inside -for loop- store a single value and a matrix. How to?

Hello guys,

I have a specific code issue. My raw data (x) is a matrix where each column is a measurement (trial 1, trial2, etc.).

Problem: I created a -for loop- portion in the code below. I need to record the NCP outcome (thus a matrix containing the NCP for each iteration resulting in a matrix of n columns as many as the raw data) and the Fc variable for each iteration. Otherwise the previous value/matrix is ovewritten and I will get only the last calculation. I think for the Fc(x) works (I get a row vector with as many columns as the number of iterations) but I can't get this done for the NCP matrix.

Any help would help me to fix the problems, and to better understand how matlab works. Thank you

x=data; %give to x the name of the raw data variable in the workspace

ncol = size(x,2);
Fs=1000; NFFT=4096;

for k = 1:ncol

%--------------------------------------------

L=length(x(:,k));
X=fftshift(fft(x(:,k),NFFT));
Px=X.*conj(X)/(NFFT*L); %Power of each freq components
fVals=Fs*(-NFFT/2:NFFT/2-1)/NFFT;
fVals=fVals((NFFT/2+1):end);
Px=Px((NFFT/2+1):end);

CP = cumsum(Px) ;
NCP = CP / CP(end)*100;
figure1 = figure('Color',[1 1 1]);
axes1 = axes('Parent',figure1,'FontSize',14);
box(axes1,'on');
hold(axes1,'all');
plot(fVals,NCP,'MarkerSize',1,'LineWidth',2,'Color',[0 0 0]);
title('Cumulative Power');
xlabel('Frequency (Hz)','LineWidth',1,'FontWeight','bold','FontSize',14);
ylabel('Power (%)','FontWeight','bold','FontSize',14);

optimal_Fc=find(NCP>99);

Fc(k)=fVals(1,(optimal_Fc(1))); %76.179Hz

end

Answer 1

[Adam Danz]

You correctly stored single values in Fc(k) which becomes a vector. When you store vectors such as NCP, they will become matrices. Matrices become 3D arrays and so on (always adding a dimension).

change

   NCP = CP / CP(end)*100;

to

NCP(:,k) = CP / CP(end)*100;

Also, prior to your loop, allocate the arrays.

NCP = nan(q, ncol);  %where q is the length of each vector - you can figure that one out.
Fc = nan(1, ncol);

[Giuseppe]

Thanks Adam, tha was very helpful. The code works and the problem is fixed. q was basically NFFT/2 in that specific case. Cheers