how to apply wavelet and extract the features from speech signal

Hi everyone i am doing my project in "Word spotting in continuous speech using wavelet transform" i want matlab code to perform wavelet transform and feature extraction. please any one help me...

Answer 1

[Snehal]

Hi,

I understand that you want to know how to apply wavelet transform and feature extraction on speech signals.

This can be achieved with help of Wavelet toolbox

1. Wavelet Transform: The toolbox provides different functions like ‘wavedec’ for discrete wavelet transform (DWT) and ‘cwt’ for continuous wavelet transform (CWT). For speech processing, DWT is commonly used due to its efficiency and ability to localize features in time.
2. Feature Extraction: Make use of functions like ‘appcoef’ and ‘detcoef’ for extracting the approximation and detail coefficients from wavelet decomposition respectively. These coefficients can then be used to calculate statistical features like energy and entropy.

Refer to the sample code below:

% Performing Discrete Wavelet Transform (DWT) on the speech signal
[c, l] = wavedec(speechSignal, decompLevel, waveletName); 
% decompLevel refers to the level of decomposition of the speech signal   

% using the 'appcoef' and 'detcoef' functions for extracting approximation and detail coefficients
approximation = appcoef(c, l, waveletName, decompLevel); 
details = detcoef(c, l, 1:decompLevel);   

% calculating the energy from coefficients (feature extraction)
energyApproximation = sum(approximation.^2); 
energyDetails = cellfun(@(x) sum(x.^2), details); 

Alternatively, you can also use MFCC features (Audio toolbox) and LSTM (Deep Learning toolbox) to implement word spotting in MATLAB.

Refer to the following documentations for better understanding:

• https://www.mathworks.com/help/wavelet/ref/wavedec.html
• https://www.mathworks.com/help/audio/ug/sequential-feature-selection-for-audio-features.html
• https://www.mathworks.com/help/wavelet/ref/appcoef.html
• https://www.mathworks.com/help/wavelet/ref/detcoef.html
• https://www.mathworks.com/help/audio/ref/mfcc.html
• https://www.mathworks.com/help/matlab/ref/cellfun.html

Hope this helps.