Simple speech segmentation in Matlab
The two dataset are Set-C and Set-G, where each set consists of 10 wav files and 10 segmentation files. For Set-C files, all of them are from a female speaker, while the Set-G files are from a male speaker.
Experimental setup and Result of Proposed Algorithm
Experiment setup :
• Experimental Variable : Time tolerance, Threshold and win size
• Time tolerance: t = {0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.10 }
• Threshold : Thres = { 1.4 , 2.4, 3.4}
• Win size : W = {256}
The proposed algorithm adds a end point of the soundwave.
The proposed algorithm also adds a moving average filter (‘movmean’) in order to reduce noise and fluctuations. This is to smoothen out the energy curve.
By implementing the smoothing effect of the moving average filter and the inclusion of end points in the proposed algorithm, it offers an improved approach for speech segmentation, enabling a more accurate identification of segment boundaries and potentially enhancing subsequent analysis or processing tasks on the segmented portions of the speech.