How to extract audio segment?

Hello,

I have several speech files (between 20 and 100s), and I want to segment them for further analysis. How can I cut them in frames and then extract them? I looked into audioread and audiowrite, and I also tried a Mirsegment(...,'RMS') segmentation, however even if the segmentation works, it doesn't extract them into separate audio files.

Thank you very much,

C

[Geoff Hayes]

Cyrielle - how does Mirsegment work? What segmentation does it do to the audio stream? I.e. what structure is the output data. It may be that you will need to use audiowrite to create the separate audio files.

[Cy]

Hi Geoff, Thanks for your answer. I used an RMS segmentation (from the Mirtoolbox), so it's using the peaks and silences to make slices. Here is a pic of the outcome. I wrote that with audiowrite, but it doesn't extract every segment. I think I would need a "for" loop to use every segment and extract them separately, but I don't see how.

sourceDir = ('D:\CHAP');

input = audioread('D:\CHAP\Social1A.wav');
destDir = ('D:\CHAP\frames');

% segment the sound

sg = mirsegment(input, 'RMS');
mirsegment(sg);

audiowrite(sg, 'mysave1',destDir) 

disp('there you go!')

[Cy]