adding of signals

Hi, When we add two images, we need to maintain same dimension to add. I need to add two wave signals. How to alter their dimensions, so that i can add.

Answer 1

[Image Analyst]

The imresize in your Image Processing Toolbox will also work on one dimensional signals.

% Make wave 2 the same length as wave 1.
resizedWave2 = imresize(wave2, length(wave1));

[Sivakumaran Chandrasekaran]

Hi Image Analyst, Just now I checked your coding and its not working for my application.

Regards,

Siva

[Image Analyst]

OK, what does this say

length(wave1)

length(wave2)

length(resizedWave2)

Tell me what those lengths are, or else tell me what "not working" means to you (such as an error message or something).

[Sivakumaran Chandrasekaran]

Hi Image Analyst,

My question is ,

x1=wavread('speech.wav');

x2=wavread('speech_noise.wav');

v=x1-x2

I am getting the error as

??? Error using ==> minus

Matrix dimensions must agree.

Error in ==> Untitled at 3

v=x1-x2

[Image Analyst]

Yes, I know -- they're different lengths so you can't subtract them. But show me where you tried the code I gave you to make them the same length.

[Walter Roberson]

You do not appear to have used imresize() there as Image Analyst indicated.

[Sivakumaran Chandrasekaran]

Hi walter,

are you saying that i should not use the command "imresize" to change the dimension of a signal. Am i right

[Walter Roberson]

No, I am saying you omitted imresize() where you needed to have it.

[Walter Roberson]

newsize = min(size(x1), size(2));

x1 = imresize(x1, newsize);

x2 = imresize(x2, newsize);

v = x1 - x2;

What will you do in the places where the result would be out of range for sounds? For example, -32768 for x1 and +32767 for x2 would try to subtract to -65535 but that is probably going to be out of range.

[Sivakumaran Chandrasekaran]

Thanks walter. I shall check it and let you know.

[Image Analyst]

It IS possible though that resizing doesn't make sense. Maybe it does for you because they are so close in length, like if one was 30 seconds long and the other was 31 seconds long. But would it make sense to subtract a sound signal that was originally 15 minutes long from one that is only 10 seconds long? What is the interpretation/meaning of that?

[meziane madani]

hello; i tried to do what you said, but i have this error :imresize: IM must be a grayscale or RGB image, so i don't think that we can applicate it from a song

[Image Analyst]

It can be a song. Here's proof for two stereo waveforms:

[y1, fs1] = audioread('C:\Users\Public\Music\Sample Music\Sleep Away.mp3'); % Stereo
[y2, fs2] = audioread('C:\Users\Public\Music\Sample Music\Maid with the Flaxen Hair.mp3'); % Stereo
% [y2, fs2] = audioread('guitartune.wav'); % Mono signal
whos y1 fs1
% Name            Size                Bytes  Class     Attributes
% 
%   fs1             1x1                     8  double              
%   y1        8845056x2             141520896  double  
whos y2 fs2
%   Name            Size                Bytes  Class     Attributes
% 
%   fs2             1x1                     8  double              
%   y2        7483391x2             119734256  double     
[rows1, columns1] = size(y1)
[rows2, columns2] = size(y2)
% Make wave 2 the same length as wave 1.  Code if both are stereo.
resizedWave2 = imresize(y1, [rows2, columns2]);
whos resizedWave2 % Proves y1 same size as y2.

It can get a little tricky if one is stereo and one is mono. Is that what you had?

Answer 2

[Richard Brown]

Also you can use interp1 if you don't have the image processing toolbox

Answer 3

[Sk Group]

Signal Addition function in MATLAB

MATLAB CODE:

function [y n] = sigadd(x1,n1,x2,n2)

if n1(1)< n2(1)

a = n1(1)

x1 = [zeros(1,abs(n1(1)-n2(1))) x1]

else

a = n2(1)

x1 = [zeros(1,abs(n1(1)-n2(1))) x1]

end

if n1(end)>n2(end)

b = n1(end)

x2 = [x2 zeros(1,abs(n1(end)-n2(end)))]

else

b = n2(end)

x2 = [x2 zeros(1,abs(n1(end)-n2(end)))]

end

n = a:b;

y = x1+x2;

Signal Addition function (Alternate Method) in MATLAB

MATLAB CODE:

function [y n] = sigadd_another_method(x1,n1,x2,n2)

n = min(min(n1),min(n2)):max(max(n1),max(n2));

y1 = zeros(1,length(n));

y2 = y1;

y1(find((n>=min(n1))&(n<=max(n1))==1))=x1;

y2(find((n>=min(n2))&(n<=max(n2))==1))=x2;

y = y1+y2;

Answer 4

[GULLA]

function [y n] = sigadd(x1,n1,x2,n2)

if n1(1)< n2(1)

a = n1(1)

x1 = [zeros(1,abs(n1(1)-n2(1))) x1]

else

a = n2(1)

x1 = [zeros(1,abs(n1(1)-n2(1))) x1]

end

if n1(end)>n2(end)

b = n1(end)

x2 = [x2 zeros(1,abs(n1(end)-n2(end)))]

else

b = n2(end)

x2 = [x2 zeros(1,abs(n1(end)-n2(end)))]

end

n = a:b;

y = x1+x2;

[Image Analyst]

You could do that simpler by simply setting the last element to zero to make it match the size of the longer vector.

% Add two vectors together, padding the shorter one 
% with 0's, if needed, to make them the same size.
function [y, n] = sigadd(x1, n1, x2, n2)
if n1 > n2 % x1 is longer so pad x2
    x2(n1) = 0; % Pad out with zeros.
elseif n1 < n2 % x2 is longer so pad x1
    x1(n2) = 0;
end
y = x1 + x2;
n = numel(y);

This assumes both x1 and x2 are vectors (monochrome audio signals, not 2-D stereo audio signals). See this Tricks and Tips item:

https://www.mathworks.com/matlabcentral/discussions/tips/845271-trick-to-enlarge-a-matrix-by-padding-it-with-zeros