Error using horzcat Dimensions of matrices being concatenated are not consistent.

Signal=[zeros(1,((Lmax-SignalLength-1)/2)),(Obj.CompCellList{CellNr}{iFrame}.(signal)),zeros(1,((Lmax-SignalLength+1)/2))];

I want to add zeros to both sides of the signal so that its length is equal to the value of Lmax. I have two more conditions where Lmax and SignalLength are either even or odd, but when one of them is even and the other is odd, I need to add different amounts of zeros to either side of the signal. When I add equal amounts of zeros to both sides in the case where both values ​​are either odd or even, I don't have this problem.

Lmax=77

SignalLength=20

SignalLength=length(Obj.CompCellList{CellNr}{iFrame}.(signal))

Answer 1

[Voss]

Here's how you can add zeros to the beginning and end of your signal (with the extra zero going to the end in case the number of zeros needed is odd - as you appear to have it):

% Signal = Obj.CompCellList{CellNr}{iFrame}.(signal);
Signal = rand(1,4);

Lmax = 9;
SignalLength = length(Signal);
Nzeros = Lmax-SignalLength;

if mod(Nzeros,2)
    Signal = [zeros(1,(Nzeros-1)/2), Signal, zeros(1,(Nzeros+1)/2)];
else
    Signal = [zeros(1,Nzeros/2), Signal, zeros(1,Nzeros/2)];
end

disp(Signal);

         0         0    0.8021    0.0218    0.9967    0.1352         0         0         0

Or, using a variable Nzeros_mod2 = mod(Nzeros,2) instead of the if/else construction:

Signal = rand(1,4);

Lmax = 9;
SignalLength = length(Signal);
Nzeros = Lmax-SignalLength;
Nzeros_mod2 = mod(Nzeros,2);

Signal = [zeros(1,(Nzeros-Nzeros_mod2)/2), Signal, zeros(1,(Nzeros+Nzeros_mod2)/2)];

disp(Signal);

         0         0    0.0974    0.4638    0.7127    0.5242         0         0         0

Another example, where the number of zeros is even:

Signal = rand(1,4);

Lmax = 8;
SignalLength = length(Signal);
Nzeros = Lmax-SignalLength;
Nzeros_mod2 = mod(Nzeros,2);

Signal = [zeros(1,(Nzeros-Nzeros_mod2)/2), Signal, zeros(1,(Nzeros+Nzeros_mod2)/2)];

disp(Signal);

         0         0    0.7752    0.4344    0.5274    0.7130         0         0

Notice that your expression will work only when the number of zeros is odd, i.e., when one of Lmax and SignalLength is odd and the other is even. In case they are both even or both odd, then their difference is even and you'd get an error using the zeros function because (Lmax-SignalLength+1)/2 is not an integer:

Signal = rand(1,4);

Lmax = 8;
SignalLength = length(Signal);

try
    zeros(1,(Lmax-SignalLength+1)/2) % error here
catch ME
    disp(ME.message); % report the error message
end

Size inputs must be integers.

However, this is not the error message you report in the title of the question. The error message, "Dimensions of matrices being concatenated are not consistent." in this context would suggest that your signal is not of size 1 in dimension 1 most likely:

Signal = rand(2,4); % size 2 in dimension 1, i.e., Signal has two rows

Lmax = 8;
SignalLength = length(Signal);
Nzeros = Lmax-SignalLength;
Nzeros_mod2 = mod(Nzeros,2);

try
    Signal = [zeros(1,(Nzeros-Nzeros_mod2)/2), Signal, zeros(1,(Nzeros+Nzeros_mod2)/2)]; % error here
catch ME
    disp(ME.message); % report the error message
end

Dimensions of arrays being concatenated are not consistent.

To add zeros to the beginning and end of a signal with size other than 1 in dimension 1, take that into account in calling the zeros function:

Signal = rand(2,4);
Nrows = size(Signal,1); % number of rows of Signal

Lmax = 8;
SignalLength = size(Signal,2); % use size(_,2) instead of length() for robustness
Nzeros = Lmax-SignalLength;
Nzeros_mod2 = mod(Nzeros,2);

Signal = [zeros(Nrows,(Nzeros-Nzeros_mod2)/2), Signal, zeros(Nrows,(Nzeros+Nzeros_mod2)/2)];

disp(Signal);

         0         0    0.4438    0.8402    0.1166    0.8648         0         0
         0         0    0.2582    0.4988    0.6103    0.5558         0         0