| 摘要 |
PURPOSE:To realize the analysis based on a waveform model of a vocal cord sound source by a small operation quantity by providing an FIR filter having a reverse characteristic of a frequency characteristic which a vocal cord sound source wave has. CONSTITUTION:This device is provided with an input buffer 2 for storing temporarily an input sound signal, an FIR filter 8 having a reverse characteristic of a frequency characteristic which a vocal cord sound source wave has, a coefficient generating means for generating as coefficient of the FIR filter 8 having the reverse characteristic of the frequency characteristic which various vocal cord sound source waves have, spectrum analyzing means 6, 7 for analyzing a spectrum of an output signal of the FIR filter 8, and an error comparing means 5 for comparing each analytic error obtained successively by the spectrum analyzing means. In this state, when the input sound signal is inputted to the FIR filter 8 having the coefficient of the FIR filter 8 to a candidate of various vocal cord sound source waves, the candidate of the vocal cord sound source wave at the time when the analytic error obtained by the error comparing means 5 becomes minimum is set as a result of analysis of the vocal cord sound source wave, and the control is executed so that a result of the spectrum analysis at that time is outputted as a result of analysis of a vocal meatus characteristic. In such a way, the analysis based on a waveform model of the vocal cord sound source can be realized by a small operation quantity. |
