Voice Audio Data and Pitch Extraction

I want to extract numerical data from raw voice audio files(.wav)- take out the numeric pitch (Hz) and do specturm analysis in a voice audio file. However, I want to have the pitch throughout the audio file with time stamps to a later to be converted to a .csv or .txt, and for other analysis like spectrum and stats.

How do I do that? I am looking for succesful code. I am new to Matlab, but I have the audio packages already installed. So far I seen parts of what I want to do on the internet, but they never work. As for now the data is not noisy, but I might need to clean some with extra sound in the background.

Thanks for your help in advance.

Answer 1

[jibrahim]

You can use the pitch function to get pitch values along with the time stamps

[x,fs] = audioread("SingingAMajor-16-mono-18secs.ogg");

winLength = round(0.05*fs);
overlapLength = round(0.045*fs);
[f0,idx] = pitch(x,fs,Method="SRH",WindowLength=winLength,OverlapLength=overlapLength);
tf0 = idx/fs;

[jcclar12]

Thank you. How this is not getting me the time stamp for the whole audio file

[Walter Roberson]

The internal code uses

hopLength = params.WindowLength - params.OverlapLength;
numHops = floor((size(x,1)-params.WindowLength)/hopLength);


varargout{2} = cast(((0:numHops)*hopLength + params.WindowLength)','like',x);

So it is more or less creating windows, and you are getting an index for each window.

[jcclar12]

Thank you!