In this paper, the masked threshold of a sinusoidal signal in the presence of a notched-noise masker was measured experimentally for five normal-hearing subjects. The frequencies of sinusoidal signals used in the measurement were 125, 250, 500, 1,000, 2,000, 4,000, and 6,000 Hz. The conditions and procedure in our measurement were the same as those used by Glasberg and Moore (2000), with additional measurements at 125 and 6,000 Hz. Uniformly excited noise (UEN) was not used in our measurements. The measured data was used to estimate the parameters of a double roex auditory filter as presented in Glasberg and Moore (2000). Basically, this filter is the sum of a tip filter and a tail filter, with its gain controlled by a schematic family of input-output functions. The PolyFit procedure was used to fit the filter to the measured data. An individual auditory filter was fitted at each of the signal frequencies in our measurements. The results showed that auditory filter shape varied with level. The gain of the filters centered at frequencies between 125 Hz and 1,000 Hz, increased as the center frequency increased. Above 1,000 Hz, the gain of the filters remained at a constant value. These results are consistent with the results in Baker et al. (1998) and Glasberg and Moore (2000).
Hearing thresholds for pure tones from 2 kHz to 28 kHz were measured. A 2AFC procedure combined with a 3-down 1-up transformed up-down method was employed to obtain threshold values that were less affected by listener’s criterion of judgment. From some listeners, threshold values of 88 dB SPL or higher were obtained for a tone at 24 kHz, whereas thresholds could not be obtained from all participants at 26 kHz and above. Furthermore, thresholds were also measured under masking by a noise low-pass filtered at 20 kHz. At frequencies above 20 kHz, the difference of threshold values between with and without the masking noise was a few decibels, indicating that the tone detection was not affected by subharmonic components that might have appeared in the lower frequency regions. The results of measurement also showed that the threshold increased rather gradually for tones from 20 to 24 kHz whereas it increased sharply from 14 to 20 kHz.
A filtered-reference LMS algorithm is often used in practical active noise control systems. This algorithm is derived under the assumption of a stationary noise signal, and the order of the signal convolution is switched in the derivation process. However, the order of convolution cannot be changed in a real physical process. We examine the differences between these situations, i.e., theoretical formulation compared to practical control process, for various kinds of noise signals. Amplitude-modulated and low-pass noise signals are used as examples of disturbance signals. Results of the numerical simulations indicated that the practical implementation of the algorithm requires severer conditions for the convergence coefficient than those in the theoretical prediction. Analytical examination with the transformed coefficient update procedure also reveals the difference between them. To reduce this difference and achieve robust attenuation under practical conditions, a modified version of a filtered-reference LMS algorithm is introduced. Advantages of this modified version are verified through a series of simulations.
This paper addresses the blind dereverberation problem of a single-input multiple-output acoustic system. Many conventional approaches require a precise order of the transfer functions. In this paper, we propose an equalization algorithm that is less sensitive to the order misadjustment of the transfer functions. First, the transfer functions are estimated using an overestimated order, and the inverse filter set for these estimated transfer functions is calculated. Since the estimated transfer functions contain a common polynomial, the signal processed by the inverse filter set suffers from the effect of this common polynomial. Then, we extract this polynomial to compensate for the distortion. The proposed algorithm recovers input signal as long as the channel is overestimated. Simulation results show that the proposed method works well even when the order is highly overestimated.
Methods of treating initial and boundary conditions when applying the finite difference time domain (FDTD) numerical method to the analysis of elastic wave fields in anisotropic solids have been considered with reference to an electrical circuit analogy. It was found that the staggered lattice of the FDTD analysis of anisotropic solids is expressed by the superposition of two electric circuit networks of identical configurations. One of these circuits is translated relative to the other by one grid length in both the y and z directions, and is then superposed onto the other circuit to yield the final overall circuit. The two circuits are connected by the mutual capacitance between the capacitances loaded at the normal stress nodes of one of the circuits, and the shear stress nodes of the other circuit, which are positioned on the same stress nodes. From the configurations of the circuits, it can be concluded that the initial and boundary conditions should be applied to both the circuit networks. The theory was confirmed by considering two model problems.
Harmonic coding is a very powerful technique for the coding of speech at very low bit rates; and the efficient coding of spectral magnitudes sampled at harmonic frequencies is the key to obtaining good coded-speech quality. This paper presents a weighted vector quantization method for spectral vectors composed of a variable number of harmonic magnitudes. It is based on simple, efficient linear dimension conversion and employs a weighted distortion measure that exploits the human auditory sense. A codebook training algorithm using the weighting matrix is also presented. Finally, a low-complexity VQ codebook search technique based on pre-selection is described that reduces the computational complexity to less than 10% of that of an exhaustive search, without perceptible loss of quality. The proposed quantization scheme is used in Harmonic Vector eXcitation Coding (HVXC), which is a very low-bit-rate speech coding algorithm that combines harmonic and stochastic vector representations of LPC residual signals. Due to the high efficiency of this VQ scheme, HVXC provides good communication-quality speech at bit rates as low as 2–4 kbit/s, and was adopted as the ISO/IEC International Standard for MPEG-4 Audio.