An approach to smooth fundamental frequencies in tone recognition
Jun Liu, Xiaoyan Zhu, Yuping Luo · 2002
Mandarin Chinese is a tonal language. The tones can be recognized separately, primarily by the pitch contour information. The classic approaches, such as curve fitting, linear smoothing and median smoothing, cannot deal with several successive fundamental frequencies which result in errors. Proposed in this paper is a new approach, a searching method to get a preferable accurate pitch contour. This approach is characterized for its simple, reliable and fast performance.