Spectral Analysis of Familiar Human Voice Based On Hilbert-Huang Transform

摘要

Spectral analysis of human voice signals is important to reveal hidden information when is not available in the time-domain. Extracting spectral information from those voice signals will enhance our knowledge in understanding the nature and characteristic of the voice. It concerned with the decomposition method of voice signals into simpler components in frequency and time. The frequency analysis tools are also give beneficial for describing the spectral distribution in a voice signal, very often the methods used by the tools have limitations that restrict us to interpret the data properly. This paper describes a powerful data analysis method called the Hilbert-Huang transform (HHT), which can be used to extract audio frequency components from nonlinear and nonstationary human voice signals. It can describe the audio frequency components locally and adaptively for nearly any oscillating signal. This makes it very extremely versatile to be used for analysing familiar human voices.