Detection of Dynamic Structures of Speech Fundamental Frequency in Tonal Languages

An approach is proposed specially for capturing fine dynamic structures of speech fundamental frequency F₀ that may vary in such a nonmonotonic way as those of the third tones in Chinese speech. It first estimates the rough trend of variation of a F₀ contour by means of the cepstrum technique, and then, utilizes the trend as a reference to track the variation and calculates the detailed contour from a few of intrinsic mode functions that are decomposed by the ensemble empirical mode decomposition. Intensive evaluation and direct comparisons with existing methods are conducted with the standard Chinese Mandarin database, showing the effectiveness of the proposed method in acquiring accurate and reliable F₀ contours from speech signals even heavily contaminated with noise.