How can I resolve the subscripted assignment dimension mismatch error (for the cwt plot)?

Hello,

I have an issue with the syntax used for the wavelet plot. The syntax used is:

[cfs(:,i), f(:,i)] = cwt(signal(:,i),fs(:,i)).

I have placed this syntax inside a for loop, where 'i' is the Excel sheet name. However, I am facing an error saying, "Subscripted assignment dimension mismatch." Can you please guide me on how to fix this error? I would appreciate any help you can provide.

Answer 1

[Christopher McCausland]

Hi Navid,

Can you please post so information in relation to both your signal and fs variable? Or better yet attach them?

cfs, should probably be a three dimetional array, so:

cfs(:,:,i) 

as the cwt will return a two dimentional image if this code is sitting in a for loop. However, until you give us a bit more information this is hard to solve, for example your input signal could be in columns, or may not be.

[Navid]

Dear Christopher McCausland,

I want to express my gratitude for your help. I have attached the Excel file containing the data, as requested. I appreciate your assistance in advance.

Thank you, and kind regards,

Navid

[Walter Roberson]

filename = "C1.xlsx";
S = sheetnames(filename);
for K = 1:numel(S)
    s = S(K);
    T = readtable(filename, 'sheet', s, 'VariableNamingRule','preserve');
    [s, T.Properties.VariableNames]
end

ans = 1×3 string array
    "CP1"    "Time(s)"    "Signal"

ans = 1×3 string array
    "CP2"    "Time(s)"    "Signal"

ans = 1×3 string array
    "CP3"    "Time(s)"    "Signal"

ans = 1×3 string array
    "CP4"    "Time(s)"    "Signal"

ans = 1×3 string array
    "CP5"    "Time(s)"    "Signal"

ans = 1×3 string array
    "CP6"    "Time(s)"    "Signal"

ans = 1×3 string array
    "CP7"    "Time(s)"    "Signal"

It is not clear to me how these inputs are converted into signal and fs ?

[Christopher McCausland]

Taking a bit of a guess here, we can work out the sampling frequency with f = 1/T. T appears to be 0.01 so f must be 100 Hz (do be careful of Nyquest etc.). And assuming that each sheet is a different signal. (If either if these statments are not true, then this code will need tweaked).

We can modifiy Walters code to add in the cwt command like so:

filename = "C1.xlsx";
S = sheetnames(filename);
for K = 1:numel(S)
    s = S(K);
    T = readtable(filename, 'sheet', s, 'VariableNamingRule','preserve');
    [s, T.Properties.VariableNames]
    [cfs(:,:,K), f(:,K)] = cwt(T.Signal(:,:),100);
end

% We can also have a look at the generated values 
imagesc(abs(cfs(:,:,1)))

You should also note that fs is a fixed value, it doesn't appear to change in your code, I am not 100% sure how you were generating it, as incrementing though the time vector would give very wrong answers.

Hopefuly this helps!

Christopher

[Navid]

Dear Christopher McCausland,

Unfortunately, your code was accompanied by the following error:

"Undefined function or variable 'sheetnames'."

I have used the following code:

filename = 'C1.xlsx';

% Get the sheet names
[~, sheetNames] = xlsfinfo(filename);

for i = 1:numel(sheetNames)
    % Read the data from the sheet
    data = xlsread(filename, i);
    
    % Extract the time and signal columns
    t (:,i) = data(:, 1);
    signal (:,i) = data(:, 2);
    dt = t(2) - t(1); % sampling time
    fs = 1/dt; % freq
end

[Walter Roberson]

Sorry, I did not notice the R2016b; sheetnames() was added in R2019b.

filename = 'C1.xlsx';

% Get the sheet names
[~, sheetNames] = xlsfinfo(filename);

for i = 1:numel(sheetNames)
    % Read the data from the sheet
    data = xlsread(filename, i);
    
    % Extract the time and signal columns
    t(:,i) = data(:, 1);
    signal(:,i) = data(:, 2);
    dt = mean(diff(t(:,i))); % sampling time
    fs(i) = 1/dt; % freq
    [temp_cfs, temp_f] = cwt(signal(:,i), fs(i));
    whos temp_cfs temp_f
    cfs(:,:,i) = temp_cfs;
    f(:,i) = temp_f;
end

  Name            Size                  Bytes  Class     Attributes

  temp_cfs      104x10015            16664960  double    complex   
  temp_f        104x1                     832  double              

  Name            Size                  Bytes  Class     Attributes

  temp_cfs      104x10015            16664960  double    complex   
  temp_f        104x1                     832  double              

  Name            Size                  Bytes  Class     Attributes

  temp_cfs      104x10015            16664960  double    complex   
  temp_f        104x1                     832  double              

  Name            Size                  Bytes  Class     Attributes

  temp_cfs      104x10015            16664960  double    complex   
  temp_f        104x1                     832  double              

  Name            Size                  Bytes  Class     Attributes

  temp_cfs      104x10015            16664960  double    complex   
  temp_f        104x1                     832  double              

  Name            Size                  Bytes  Class     Attributes

  temp_cfs      104x10015            16664960  double    complex   
  temp_f        104x1                     832  double              

  Name            Size                  Bytes  Class     Attributes

  temp_cfs      104x10015            16664960  double    complex   
  temp_f        104x1                     832  double              

[Navid]

Dear Walter Roberson,

Thank you very much.

Coefficients were successfully obtained.

It is better to mention that in the next step, I have the following code:

figure; contour(t,f,abs(cfs).^2)

But The following error is displayed:

"Error using contour

Input arguments for contourc must have at most 2 dimensions."

Maybe it would have been better to first mention the contour plot issue in my question.

Thank you anyway.

[Walter Roberson]

Your cfs is 3D because you have one layer for each sheet.

What would a 3D contour plot be like for you?

Perhaps isosurface would work for you... but I suspect not.

[Navid]

Thank you for your time and consideration.

I am trying to generate the wavelet contour for each sheet in an Excel file. The code provided in the MATLAB documentation efficiently plots this contour for data in a single sheet. Please consider the following example used in the MATLAB documentation:

Fs = 1000;

t = 0:1/Fs:1-1/Fs;

x = 1.5*cos(2*pi*100*t).*(t<0.25)+1.5*cos(2*pi*50*t).*(t>0.5 & t<=0.75);

x = x+0.05*randn(size(t));

[cfs,f] = cwt(x,Fs);

contour(t,f,abs(cfs).^2);

axis tight;

grid on;

xlabel('Time');

ylabel('Approximate Frequency (Hz)');

title('CWT with Time vs Frequency');

Utilizing a loop would enable automatic plotting for multiple sheets. However, I cannot generalize the code for my specific requirement if implementing a for loop.

[Christopher McCausland]

Hi Navid,

Can I ask why you want to do this?

While we can include plots in for loops, its generally considered bad practice. I am also not convinced about using contour as I do not think it convays the information well.

Regardless, please see attached contour as requested, and in addition via mesh() which I personally think convays the information better. The y-axis has been set to log to reflect the nature of the CWT. You should also note that this information is the same as a 2D cwt plot with a colour scale.

Christopher

filename = 'C1.xlsx';
% Get the sheet names
[~, sheetNames] = xlsfinfo(filename);
for i = 1:numel(sheetNames)
    % Read the data from the sheet
    data = xlsread(filename, i);
    
    % Extract the time and signal columns
    t(:,i) = data(:, 1);
    signal(:,i) = data(:, 2);
    dt = mean(diff(t(:,i))); % sampling time
    fs(i) = 1/dt; % freq
    [temp_cfs, temp_f] = cwt(signal(:,i), fs(i));
    whos temp_cfs temp_f
    cfs(:,:,i) = temp_cfs;
    f(:,i) = temp_f;
    
    
    % Begin plotting here, typically we would not do so in a for loop. 
    figure()
    contour(t(:,i),f(:,i),abs(cfs(:,:,i)).^2);
    set(gca, 'YScale', 'log');
    
    % Here is the same information presented with Mesh, which convays the data trends
    % better in my opinion.
    figure()
    mesh(t(:,i),f(:,i),abs(cfs(:,:,i)).^2)
    view(-30,60)
    set(gca, 'YScale', 'log');
end

[Walter Roberson]

Of course in the production code you can omit the

    whos temp_cfs temp_f

which was added for debugging purposes,

[Navid]

Dear Christopher,

I wanted to express my gratitude for the valuable guidance you provided me. This part of the code:

mesh(t(:,i),f(:,i),abs(cfs(:,:,i)).^2)

truthfully gives a deeper understanding of the signal.

However, sometimes, a 2D cwt plot may be more suitable for monitoring specific frequency ranges of a signal.

Thank you for taking the time to share your knowledge with me. Your guidance has been instrumental in helping me improve my work, and I greatly appreciate it.

Best regards,

Navid

[Navid]

Dear Walter Roberson, Your support is greatly appreciated. Thank you so much!