Continuous Spectrogram from multiple wav files.

Hello all,

I have been scouring the existing online help but cant seem to find exactly what I’m after. I have a folder with hundreds of 4 min wav files and want to produce a single spectrogram which contains them all. My initial thinking was to read in all wav files and somehow join them into one large file to later put into a spectrogram. Any help would be great appreciated!

Thanks,

Thomas

Answer 1

[Stijn Haenen]

Something like this:

data_tot=[];
>> for i=1:numel(files)
    data=audioread(files(i));
    data_tot=[data_tot ; data];
end
audiowrite('file.wav',data_tot,Fs);

[Thomas Webber]

Hi Stijn,

Thanks for this, would you know of anyway to account for the length of files not being exactly equal? The exact number of samples in each wav file differ slightly, which is throwing up the error:

"Error using horzcat

Dimensions of arrays being concatenated are not consistent."

Thanks,

[Stijn Haenen]

you have to make colomns from your data, you can add some lines to do this:

data_tot=[];
>> for i=1:numel(files)
    data=audioread(files(i));
    if size(data,1)<2
        data=data';
    end
    data_tot=[data_tot ; data];
end
audiowrite('file.wav',data_tot,Fs);

and you have to check whether all your files are single audio files or double (left and right audio).

[Thomas Webber]

Thats wonderful, thankyou for your help!