Pitch frequency estimation of Chinese multi-syllable words based on wavelet transform

Pitch frequency is an important parameter in many speech applications. In order to estimate the pitch frequency accurately, a voiced regions detection (VRD) algorithm based on wavelet transform and Teager energy operator is firstly proposed. Then an algorithm based on wavelet transform for estimating the pitch frequency in the voiced regions only, is presented. Experiments were conducted to evaluate the performance of the proposed algorithms. Results of the experiments have shown that the VRD algorithm can accurately extract the voiced regions. It is also observed that the proposed pitch frequency estimation algorithm is robust to noise and the obtained pitch frequency curves accord with Chinese tone modes.