remove noise from an audio signal

I need to take a specific part from a signal, which represents a one-word sound, how I can use that to any entered signal.

the figures below illustrate the part that I need to keep.

any suggestions?

Answer 1

[Image Analyst]

I've done something like this a few times before here. Here is one:

https://www.mathworks.com/matlabcentral/answers/229678-how-to-divide-a-signal-into-groups-of-peaks#answer_185935

What I did was to take the absolute value of the signal and threshold it to get quiet parts. Then use bwareaopen() to throw out short quiet parts, and invert the result to get a binary signal of where the words are.

Or you might try experimenting with movmax() and then thresholding.

Attach your signal and code if you have trouble. On the other hand, I'm sure there is literature out there for extracting individual spoken words so there is undoubtedly something better than my algorithms.

Or try this:

https://www.mathworks.com/matlabcentral/fileexchange/28826-silence-removal-in-speech-signals

[Image Analyst]

abdallah, did that work for you? Let me know.

[abdallah Bahrawi]

unfortunately, this did not work for me, maybe I didn't understand it very well, for more specification I'm reading the signal using [y,fs] = audioread('here is the directory of the audio file '); , and when I plot the audio the result will be as the figures.

s=load('signal.mat')

signal = s.coarse_d;

I don't know what exactly is coarse_d

I am a new user of Matlab, I'm looking for an easy way to do the job.

___

I've tried something else instead, I divided the signal into equal frames and for each frame take the maximum value and remove it if it as an example less than 0.2, this will work with the second and third signals in the figures, still looking for an efficient way.

[Image Analyst]

Please read this link first

TUTORIAL: How to ask a question (on Answers) and get a fast answer

then attach signal.mat with the paper clip icon. Maybe you didn't adapt my other code correctly.

[abdallah Bahrawi]

silence removal was very helpful thanks for your support