How can I fix a "Subscripted assignment dimension mismatch."

hi, my project is about atrial fibrillation I want to use butterworth method to filter out the signal, however when i use filtfilt it keep saying "Subscripted assignment dimension mismatch" which didnt occur in the first few runs

Can someone please help me

Thanks

 %%Signals with no fitler
 Virtuals = loadmat('LA_D_001_PRE_subtracted_VIR.mat');
 Virtuals = Virtuals';

 subplot(3,1,1);
 plot(Virtuals);
 title('Signals with no filter');
 xlabel('Time');

 %%Applying with Butterworth filters
 X = Virtuals;
 %First order
 [b, a]= butter(1, [4/256 10/256]);
 %%Filtfilt
 for i=1:2048
    Y2(i,:)=filtfilt(b,a,X(i,:));
 end

 subplot(3,1,2);
 gplot(Y2,b);
 hold on;
 gplot(Y2,a);
 hold off;
 title('signals with Butterworth filter - First order');
xlabel('Time');

 %%phase
                    A= Y2';
 %                     waitbar(6/8)
 %    [r c]=size(A);
 %    Phase=zeros(size(A));
         for j=1:size(A,1)
            s=squeeze(A(j,:));
            s=s-mean(s);
            ha=imag(hilbert(s));
            Phase(j,:)=-atan2(ha,s);
         end
           clear A

[Geoff Hayes]

Tram - use the debugger and step through the code to see which oune is said using the error. Once you find that line, look at the inputs and question whether they make sense.

[Thorsten]

Please provide the LA_D_001_PRE_subtracted_VIR.mat file such that we can have a look.

[Star Strider]

Is ‘Virtuals’ a cell array? What size it is? At what value of ‘i’ does the error occur?

(It is not advisable to use i or j as variables. MATLAB uses them as its imaginary operators, so using them as variables could cause confusion.)

[tram nguyen]

geoff, i've tried to used debugger but it says error and cannot be used

ive tried cell array but it didnt seem to work

it is a very large size, 2048x148765

the file is too big that i cannot upload it here :(

Answer 1

[dpb]

First, eliminate all the spurious code that has no bearing on the problem...the plotting and such is of no use; is simply distraction.

Second, paste the actual error message in context from the window with the offending line; don't make us guess which of the possible lines it actually is that is the offender.

Third, it appears given the code that the likely culprit is that you change the definition of the input data from one case to a previous and didn't redefine or clear Y2 between those and so the failing length of the array isn't the same as that of the first.

Insert a

clear Y2

before the loop and see if that doesn't fix the problem.

It would be better if you were to preallocate Y2 rather than dynamically resizing it each pass...

There are at least a couple of other anomalies/possible problems here--

1) What's the magic number 2048; where did it come from?

2) You plotted(*) the unfiltered signals by column since you transposed the input array but you're processing them/applying the filter by row, not column.  That doesn't make any sense.

%%Applying with Butterworth filters
[b, a]= butter(1, [4 10]/256);

Virtuals = loadmat('LA_D_001_PRE_subtracted_VIR.mat').';
Y2=zeroes(size(Virtuals));  % allocate output same size for filtered
for i=1:size(Virtuals,2)    % filter each column
  Y2(:,i)=filtfilt(b,a,Virtuals(:,i));
end

 (*) So, I guess in that context the *plot* _does_ add something in that it seems to demonstrate the orientation of the data in the array.  Of course, if it is actually the other way 'round then the plot is wrong and the filtered signal could be right.  We can't know that; only you know for sure which way the data are stored in the file (or at least hopefully you do or they're such that it's easy to verify which is which).  But, the key is we can't tell from here...