Does audioread not return a vector?

I'm having a small issue with understanding why this conv wont pass through.

Fs = 33600;%sampling frequency
T = 1;
samples = [1,T*Fs];%this is my sample range, for seconds recorded change the multiplier
[filename,filepath] = uigetfile('*.*', 'Please choose audio data'); 

B = audioread(strcat(filepath,filename),samples);
A = zeros(1,100);
op = conv(A, B);

When I run this and select the file that I want to read, I get this Error:

Error using conv (line 28)
A and B must be vectors.

Error in convcheck (line 8)
op = conv(A, B);

I was under the impression that Audioread did return a vector, but if i'm wrong then there has to be another way to convolute audio signals that I'm unaware of?

thank you for your time.

[Stephen23]

"I was under the impression that Audioread did return a vector..."

The audioread documentation clearly states that the first output is "returned as an m-by-n matrix, where m is the number of audio samples read and n is the number of audio channels in the file. " So the answer to your question is no if there are two or more channels in the file.

It is more reliable to read the documentation than to rely on an "impression" .

Answer 1

[Star Strider]

It depends on your audio file. A stereophonic signal will return an (N x 2) matrix, with the first column being the left channel, and the second column being the right channel. using:

NrChan = size(B,2);

will tell you this.

[Logan Betts]

I see, so if I want to convolve the audio file should I then try and make the file mono. Thank you for your help.

[Star Strider]

As always, my pleasure.

However to ‘make the file mono’ will necessarily remove or change the data in the file (unless the same information is present in both channels).

Answer 2

[Walter Roberson]

No, often audioread() does not return a vector. audioread() normally returns a 2D array with one column per channel.

Note: you should think about providing an option such as 'valid' or 'same' to conv()

Note by the way that convolving with 0 is always going to return all zeros. You could skip the conv() call and just do

op = zeros(length(B) + length(A) - 1, 1);

[Logan Betts]

Ok, thanks you. I know that a convolution with 0s will always be zero, I should have specified that it was only a test to ensure the culprit was the audio file.