Chaotic and fractal properties of vocal sounds

Recently, it has been discussed that, various fluctuations, which may be involved in biological signals, e.g., vocal sounds and EEG, are not stochastic oscillations such as noise signals or measurement errors, but chaotic fluctuations which can be characterized by chaotic and fractal properties. The chaotic concept based on deterministic dynamics may become an important feature in biological signal analysis. In this report, we shall discuss the chaotic and fractal properties involved in the fluctuations of vocal sounds which may be evaluated from the fractal dimensions, the Lyapunov spectra, and multi-fractal dimensions. Furthermore, to verify nonlinear properties in a time series, we analyze colored random data by using the surrogation method.