I have recorded audio files for the English letters, each file includes 26 letters. I have split each letter into a separate audio file. Now I want to put similar audio letters into one folder. I can do it manually but it will take time. Is there a classifier method to this?
4
-
$\begingroup$ Why do you want to put the files in a folder? For classification all you need to do is to get each audio input and the associated label (A-Z) in this case $\endgroup$– Jon NordbyCommented Feb 20, 2021 at 9:37
-
$\begingroup$ @jonnor I want to cluster the data into a separate folder so that I can use them to train a DL algorithm for detecting the correct pronunciation of a letter. $\endgroup$– Fatimah MohmmedCommented Feb 21, 2021 at 12:33
-
$\begingroup$ Unless you are locked to a very particular DL training setup, it is not needed to split into folders up front. The existing audio can be split based on labels on-demand during training. $\endgroup$– Jon NordbyCommented Feb 21, 2021 at 13:04
-
$\begingroup$ In any case, I have suggested an approach for you in the answers below. $\endgroup$– Jon NordbyCommented Feb 21, 2021 at 13:05
Add a comment
|
1 Answer
$\begingroup$
$\endgroup$
1
If in each file the letters are spoken separately, with silence in between, and always in the same order (A,B...,Z) then one can try to automate finding each section and its label. Use a Voice Activity Detection (VAD) module to detect each spoken character. Then assign A to the first voiced area, B to the next etc. If you have very clean and uniform audio, then it may work to use volume instead of voice activity detector.
Note that this process is a bit vulnerable to missed sections, because then all the subsequent labels will be shifted. It is also vulnerable to speakers saying characters out of order, or repeated.
An alternative would be to use a Speech Recognition system. That should be able to handle arbitrary order of utterances.
-
$\begingroup$ I have already separated each letter into a single audio file using the splitting tool based on silence in between the letters. Now I am trying to combine/cluster these audio files into groups. $\endgroup$ Commented Feb 21, 2021 at 13:05