This paper summarizes some results obtained in the last few years for the modeling of nonlinear vibrating instruments such as gongs and cymbals. Linear, weakly nonlinear and chaotic regimes are successively examined. A theoretical mechanical model is presented, based on the nonlinear von Kármán equations for thin shallow spherical shells. Modal projection and Nonlinear Normal Mode (NNM) formulation leads to a subset of coupled nonlinear oscillators. Current developments are aimed at using this subset for sound synthesis purpose.
A method of feedback control that minimizes the total acoustic potential energy in a sound field excited by an unknown disturbance is introduced. A state-space description of an acoustic plant is derived using common acoustical pole and zero modeling of room transfer functions, which we consider an effective scheme for experimental system identification. The feedback controller was designed using H∞ control theory to achieve both robust performance and high stability. Computer simulations verified that the resonant peaks in the frequency spectrum of the total acoustic potential energy could be attenuated in the low-frequency range involved in the nominal model of the plant without exciting the residual dynamics in the high-frequency range. A practical control system was built for experimental verification of the theory. Experimental results showed that the proposed control method provided the desired performance in real-time.
This paper proposes an algorithm for the blind dereverberation of speech signals based on multi-channel linear prediction. Traditional dereverberation methods usually perform well when the input signal is white noise. However, when dealing with colored signals generated by an autoregressive process such as speech, the generating autoregressive process is deconvolved causing excessive whitening of the signal. We overcome this whitening problem by estimating the generating autoregressive process based on multichannel linear prediction and applying this estimated process to the whitened signal so that the input signal can be recovered. Simulation results show the good potential of the proposed method.
The statistical distribution of normal hearing thresholds for pure tones of frontal incidence under binaural, free-field listening conditions was estimated as a function of frequency. First, the form of threshold distribution was investigated with threshold measurement data of the present study and those of other studies. Analytical results indicate that the threshold distribution has a form of normal distribution for the frequency range from 25 Hz to 16 kHz. Second, under the assumption of normality, standard deviations of thresholds were calculated for the frequency range from 25 Hz to 18 kHz by combining available threshold data of different studies. A supplementary experiment showed that thresholds at frequencies above 16 kHz were measurable with high reliability. These results illustrate the profile of our auditory sensitivity more accurately because they account for individual differences.