We propose a sound-timbre control technique that uses estimates of the resonance modes in a room. The key concept introduced here is that the resonance modes in a room where a musical instrument is played can be treated as controllable parameters of that musical instrument. By making full use of these additional parameters, a wide variety of desired sounds can be realized. Simulation results show that rich and expressive effects are achieved by using the proposed system. In particular, each frequency component of a sound can be independently and minutely modified, which leads to an impressive timbre change.
In this paper, we propose a method for repeatedly updating the coefficients of the feedback control filter used for canceling a feedback path from a loudspeaker to a noise detection microphone. The coefficients are usually estimated by feeding a sequence of extra noise to the loudspeaker. However, feeding such extra noise under active noise control is undesirable. The proposed method estimates the coefficients without feeding the extra noise. Concretely the proposed method provides two independent equations using the estimation errors involved in the coefficients of the noise control filter and then estimates the coefficients by solving those equations. One problem is that the feedback component is extremely small in comparison with the primary noise incident on the noise detection microphone. In this case, estimation of the coefficients is difficult. In this work. we solves this problem by estimating first the primary noise and then the feedback path.
In this study, we verify the performance of the simultaneous equations method using an experimental active noise control system. The simultaneous equations method is based on a priciple different from the filtered-x algorithm requiring a filter modeled on a secondary path from a loudspeaker to an error microphone. Instead of the filter, called the secondary path filter, this method uses an auxiliary filter identifying the overall path consisting of a primary path, a noise control filter and the secondary path. As inferred from the configuration of the overall path, the auxiliary filter can provide two independent equations when two different coefficient vectors are given to the noise control filter. The method thereby estimates the coefficient vector of the noise control filter minimizing the output of the error microphone. In this paper, we propose the application of a frequency domain adaptive algorithm to the identification of the overall path. An improvement in the noise reduction speed is thereby expected. In this paper, we also present computer simulation results demonstrating that the simultaneous equations method can automatically recover the noise reduction effect degraded by path changes, and finally, using an experimental system, we indicate that the method successfully works in practical systems.
A multichannel filtered-x least mean square (LMS) algorithm is an efficient feedforward algorithm in an active noise control (ANC) system, whose convergence rate is known to be limited by many factors, such as a secondary path model and the correlation between reference signals. In this paper, we introduce an adaptive blind preprocessing method for reducing the eigenvalue spread of the correlation matrix of reference signals, which is often ignored in a typical multichannel ANC algorithm. Two blind adaptive decorrelation algorithms are derived for different reference path models. Numerical experiments verify the robust performance of the proposed preprocessing methods, including reducing the mean square error and improving the convergence speed.
The effect of noise and presentation level on the perception of English consonants by native listeners and non-native listeners were examined. English words contrasting in /r/–/l/, /b/–/v/ and /s/–/th/ sounds, which are known to be difficult to distinguish for native speakers of Japanese, were presented to both native speakers of American English (AE listeners) and those of Japanese (J) in white noise and in pink noise at sytematically changed signal-to-noise ratios (SNR). Words were also presented at various presentation levels. The effects of noise and presentation level differed by phonetic contrast and language group. The /b/–/v/ and /s/–/th/ contrasts were more affected by noise at high SNRs, while /r/–/l/ was tolerant for noise when the SNR was higher than −3 dB. The presentation level affected AE listeners’ identification of /b/–/v/, but not of other contrasts. J listeners’ perception was affected less than that of AE listeners, possibly because the flooring effect for J listeners’ identification performance was low, even for original stimuli.