Function Ouput Array With Repeated Elemnts

Hello everyone, i have the following function :

function [ph]=Inst_Phase_Extraction( Filt_Signal )

ph=zeros(size(Filt_Signal,1),size(Filt_Signal,2));

for i = 1:size(Filt_Signal,1)

      %Hilbert Transform

      hil=zeros(1,size(Filt_Signal,2));

      hil(1,:)=hilbert(Filt_Signal(i,:));

      ph(i,:)=atan2(imag(hil(1,:)), real(hil(1,:)));

end

end

The ouput i get (ph) has the 1st row repeated size(Filt_Signal,1) times...

[Guillaume]

There's a lot of unnecessary indexing in your code.

function [ph]=Inst_Phase_Extraction( Filt_Signal )
    ph = zeros(size(Filt_Signal));   %I'am assuming Filt_Signal is 2d
    for row = 1:size(Filt_Signal, 1)
        hil = hilbert(Filt_Signal(row, :));  %no need to preallocate hil
        ph(row, :) = atan2(imag(hil), real(hil));  %hil(1, :) is the same as hil 
    end
end

If that code produces the same values for all rows, then that's because the hilbert transform or atan2 produces the same result for the rows of your input. Considering that the documentation of hilbert states that The imaginary part is a version of the original real sequence with a 90° phase shift, I wouldn't be surprised that atan2 produces the same result for all inputs. But I don't know anything about hilbert transforms.

[Chris Shortgeorge]

Unfortunately i dont think this is the case because if i calculate ph with the following code, outside the function, i get different results (not the first row repeated)

for row = 1:size(Filt_Signal, 1)

hil = hilbert(Filt_Signal(row, :));

ph(row, :) = atan2(imag(hil), real(hil));

end

Yes Filt_Signal is 2d

[Guillaume]

Right, then show us the inputs and the exact way you call the function with these inputs, so we can see what the problem is ourselves.

[Chris Shortgeorge]

load('signal.mat')

%% Signal Filtering

sf=input('Define Sampling Frequency (Hz) : sf = '); %Sampling Frequency

filter_order=input('Define Filter Order : ' );

Freq_bounds=input('Choose Frequency Bounds (use brackets i.e [fmin fmax]) : ');

Filt_Signal=Filter(signal,Freq_bounds(1,1),Freq_bounds(1,2),sf,filter_order );

ph=Inst_Phase_Extraction(Filt_Signal);

where function Filter is :

[Filt_Signal]=Filter(signal,lower_limit,upper_limit,sf,filter_order )

Nyquist_freq = sf/2; % Use Of Nyquist Frequency In Order To Ensure The Validity of Shannon's Theory

lowcut = lower_limit/Nyquist_freq;

highcut = upper_limit/Nyquist_freq;

passband = [lowcut highcut]; % [fmin fmax]

[Bc Ac] = butter(filter_order, passband); % Creation Of Butterworth Filter

Filt_Signal=zeros(size(signal,1),size(signal,2));

for i=1:size(signal,1)

Filt_Signal(i,:) = filtfilt(Bc, Ac, signal(i,:));

end

end

Hope it helps

[Chris Shortgeorge]

i just noticed that with following code in the main program produces the right results:

ph=zeros(size(Filt_Signal,1),size(Filt_Signal,2));

for i=1:size(Filt_Signal,1)

ph(i,:)=Inst_Phase_Extraction(Filt_Signal(i,:));

end

where :

 function [ph]=Inst_Phase_Extraction( Filt_Signal )

    hil=hilbert(Filt_Signal(1i,:));

    ph=atan2(imag(hil), real(hil));

 end

But i would prefer the for loop to be in the function as i mentioned in the first question

[Guillaume]

Please, remove all these extra blank lines you've added, then select all the code and press the {}Code button with the code selected.

This will make your posts a lot more readable.

[Chris Shortgeorge]

load('signal.mat')
%%Signal Filtering
sf=input('Define Sampling Frequency (Hz) : sf = '); %Sampling Frequency
filter_order=input('Define Filter Order : ' );
Freq_bounds=input('Choose Frequency Bounds (use brackets i.e [fmin fmax]) : ');
Filt_Signal=Filter(signal,Freq_bounds(1,1),Freq_bounds(1,2),sf,filter_order );
ph=Inst_Phase_Extraction(Filt_Signal);

where function Filter is :
[Filt_Signal]=Filter(signal,lower_limit,upper_limit,sf,filter_order )
Nyquist_freq = sf/2; % Use Of Nyquist Frequency In Order To Ensure The Validity of Shannon's Theory
lowcut = lower_limit/Nyquist_freq;
highcut = upper_limit/Nyquist_freq;
passband = [lowcut highcut]; % [fmin fmax]
[Bc Ac] = butter(filter_order, passband); % Creation Of Butterworth Filter
Filt_Signal=zeros(size(signal,1),size(signal,2));
for i=1:size(signal,1)
Filt_Signal(i,:) = filtfilt(Bc, Ac, signal(i,:));
end
end