processing and saving wav files in a loop

Hello. I'm attempting to change the sample rate of a directory of .wav files. This is what I've tried so far, and it's not working. I've tried similar code with a single file, but it is not working with my loop.

directory = dir('/data/*.wav');
numfiles = length (directory);


for i=1:numfiles

[yOld, FsOld] = audioread('directory(i)');
FsNew = 22050;
yNew = resample(yOld,FsNew, FsOld);
filename = sprintf('%s%d.wav',i)
audiowrite(filename, yNew, FsNew);

end

Thank you for the help in advanced.

Answer 1

[Image Analyst]

directory(i) is NOT the filename. directory(i).name is the base file name.

See the FAQ

You'll want something like

baseFileName = theFiles(k).name;
thisFolder = theFiles(k).folder;
fullFileName = fullfile(thisFolder, baseFileName);

Then do audioread() and audiowrite().

[Julisa Maria Ricart]

Hi Image Analyst,

Would I then do audioread() and audiowrite() on the fullFileName or is that just for writing? Would I use directory(i).name for the read?

Thank you for your help!

[Image Analyst]

You'd do

[yOld, FsOld] = audioread(fullFileName);

for reading. Then for writing, create a new filename

[~, baseFilenameNoExt, ext] = fileparts(theFiles(k).name);
baseFileName = sprintf('%s_resampled%s', baseFilenameNoExt, ext)
fullOutputFileName = fullfile(theFiles(k).folder, baseFileName);
audiowrite(......

[Julisa Maria Ricart]

Ok, this still isn't working for me... my code now looks like:

theFiles = dir('/data/*.wav');
numfiles = length (directory);

for i=1:numfiles

baseFileName = theFiles(i).name;
thisFolder = theFiles(i).folder;
fullFileName = fullfile(thisFolder,baseFileName)

[yOld, FsOld] = audioread(fullFileName);

FsNew = 22050;
yNew = resample(yOld,FsNew, FsOld);
[~, baseFilenameNoExt, ext] = fileparts(theFiles(i).name);
baseFileName = sprintf('%s_resampled%s', baseFilenameNoExt, ext)
fullOutputFileName = fullfile(theFiles(i).folder, baseFileName);
audiowrite(fullOutputFileName, yNew, FsNew);

end

[Image Analyst]

Try

folder = 'D:\data';
filePattern = fullfile(folder, '\*.wav')
theFiles = dir(filePattern);
numfiles = length (theFiles);
FsNew = 22050;
for k=1:numfiles
	if theFiles(k).isdir
		% Skip . and ..
		continue;
	end
	baseFileName = theFiles(k).name;
	thisFolder = theFiles(k).folder;
	fullFileName = fullfile(thisFolder,baseFileName);
	fprintf('Processing %s...\n', fullFileName);
	[yOld, FsOld] = audioread(fullFileName);
	yNew = resample(yOld, FsNew, FsOld);
	[~, baseFilenameNoExt, ext] = fileparts(theFiles(k).name);
	baseFileName = sprintf('%s_resampled%s\n', baseFilenameNoExt, ext);
	fullOutputFileName = fullfile(theFiles(k).folder, baseFileName);
	fprintf('    Created %s...', fullOutputFileName);
	audiowrite(fullOutputFileName, yNew, FsNew);
end

[Julisa Maria Ricart]

hmm.. still not working. I'm not getting any output.

[Image Analyst]

Sorry, get rid of the \n. Try this:

baseFileName = sprintf('%s_resampled%s', baseFilenameNoExt, ext);

[Marena]

Hello @Image Analyst,

I am having an issue using this code for looping and trimming my wav files. I am trying to loop through a large number of wav files and trip the first 4(ish) seconds and then re-write the file as a different name. This is my code and I am getting an error regarding the audiowrite function. Thanks in advance!

% create directory with folder containing data & get length
wav_folder = 'D:\Test Data'; 
folder_pat = fullfile(wav_folder, '\*.wav');
files = dir(folder_pat); 
num_files = length (files);

% for loop to run through files and save new
for i = 1:num_files
    if files(i).isdir
        continue
    end
    base_files = files(i).name; % file name
    this_folder = files(i).folder; % file folder
    file_name = fullfile(this_folder, base_files); % file names in file folder
    fprintf('Processing %s...\n', file_name); % print progress
    [y_old, Fs] = audioread(file_name); % read file in
    y_trim = y_old(1732790:69049200, 1); % clip after first 5s
    [~, base_files_noext, ext] = fileparts(files(i).name); % create new filename with no ext
    base_files = sprintf('%s_b', base_files_noext, ext); % print progress
    full_output_file_name = fullfile(files(i).folder, base_files); % create new name with ext
    fprintf('Created %s...', full_output_file_name); % print new filename
    audiowrite(full_output_file_name, y_trim, Fs); % write new wav file with new filename
end

[Image Analyst]

@Marena You may have waveforms that are less than 69049200 elements long. Try this

numElements = 5 * Fs; % 5 seconds.
lastIndex = max([numElements, size(y_old, 1)]);
if lastIndex < numElements
     % The sound was not 5 seconds long.
     firstIndex = 1;
else
    % Sound is at least 5 seconds long.
    firstIndex = lastIndex - numElements + 1;
end
y_trim = y_old(firstIndex : lastIndex, :); % clip after first 5s

If youi have any more questions, please start your own discussionn thread (not here) and attach a file that worked and one that threw the error, and tell me what the error is.

[Chali Dadi]

I have also the same problem please help for the problem below. thank you in advance

I am using emperical wavelet transform for signal decomposition for feature extraction from audio signal. the problem is not decomposing signal using empirical wavelet transform, but it is calculating the energy of all the signal decomposed and finding the maximum energy.then extracting feature from signal with maximum energy.

so simply what I want in loop in one is

1.decompose signal using ewt

2.calculate the energy of the decomposed signal

3. find the signal with maximum energy from all the decomposed signal(mra).

4.finally extracting audio feature from the signal with maximum energy.

Answer 2

[Chali Dadi]

I have also the same problem please help for the problem below. thank you in advance

I am using emperical wavelet transform for signal decomposition for feature extraction from audio signal. the problem is not decomposing signal using empirical wavelet transform, but it is calculating the energy of all the signal decomposed and finding the maximum energy.then extracting feature from signal with maximum energy.

so simply what I want in loop in one is

1.decompose signal using ewt

2.calculate the energy of the decomposed signal

3. find the signal with maximum energy from all the decomposed signal(mra).

4.finally extracting audio feature from the signal with maximum energy