Enhancement Pattern Analysis Technique for Voiced/Unvoiced Classification

Requirement to deciding whether a given frame of a speech waveform should be classified as voiced speech or unvoiced speech arises in many speech analysis systems. Several approaches have been described in the literature for making this decision. In this article presents four enhancement pattern analysis techniques to classify voiced and unvoiced based on the linear predictive coefficients. Those techniques are also compared the performance with the prosodic technique. Ten minutes of speech signal are collected to be input speech. The results show that the 4th technique provides the best performance of the quality of V/UV classification at 89.29% with 19,356 voiced frame. This technique can apply to vector quantization technique for speech compression and speech recognition which usually uses the LP coefficients as the speech feature.