This paper proposes a new auditory filterbank that enables signal resynthesis from dynamic representations produced by a level-dependent auditory filterbank. The filterbank is based on a new IIR implementation of the gammachirp, which has been shown to be an excellent candidate for asymmetric, level-dependent auditory filters. Initially, the gammachirp filter is shown to be decomposed into a combination of a gammatone filter and an asymmetric function. The asymmetric function is excellently simulated with a minimum-phase IIR filter, named the “asymmetric compensation filter”. Then, two filterbank structures are presented each based on the combination of a gammatone filterbank and a bank of asymmetric compensation filters controlled by a signal level estimation mechanism. The inverse filter of the asymmetric compensation filter is always stable because the minimum-phase condition is satisfied. When a bank of inverse filters is utilized after the gammachirp analysis filterbank and the idea of wavelet transform is applied, it is possible to resynthesize signals with small time-invariant errors and achieve a guaranteed precision. This feature has never been accomplished by conventional active auditory filterbanks. The proposed analysis/synthesis gammachirp filterbank is expected to be useful in various applications where human auditory filtering has to be modeled.
Nonuniqueness problem suffered by the normal derivative form (NDF) of a Helmholtz boundary integral equation (HBIE) applied to external Neumann problems is investigated. The NDF equation is useful in solving sound field around open surfaces, but it suffers from nonuniqueness when it is used for solving surface velocity potential at eigenvalues of corresponding internal Neumann problems of a closed surface. We have found that Schenck's CHIEF method reduces the error in the exterior sound field for some eigenvalues of the NDF equation. But the velocity potential on the surface is nonunique by the CHIEF method. Results show that, at internal eigenvalues, the NDF equation is not suitable for calculating sound field in the exterior domain. Whereas, Burton and Miller's equation (BM equation) is effective in estimating the unique surface velocity potential and gives accurate results for both the internal and external sound fields of a closed surface. In this research, a modified NDF equation (MNDF) is developed as a combination of NDF and BM equations. This MNDF is expected to give unique solutions for all wavenumbers for the sound field around open and closed surfaces combined.
Subjective evaluations of a sound field actively controlled by adding reflective sounds in a piano practice room were made in terms of “Reverberance”, “Spatial impression”, “Home practice” and “Hall practice”. The evaluations made by actually playing the piano and made when listening binaurally through headphones were compared. The evaluations were almost the same, but when actually playing the piano, the piano player's preference for reverberance strongly influenced the evaluations and especially influenced the evaluation of “Home practice”. This suggests that it may be neccesary to take into account the sound absorption of a room as well as the piano player's preference for reverberance in the sound field actively controlled by adding reflective sounds using electroacoustic technology.
This paper proposes speaker normalized spectral subband centroids (SSCs) as supplementary features in noise environment speech recognition. SSCs are computed as frequency centroids for each subband from the power spectrum of the speech signal. This feature can be obtained reliably even under noisy conditions because SSC are mainly computed from spectral peaks such as formants whose positions are almost unchanged in a noisy environment. Since the conventional SSCs depend on formant frequencies of a speaker, the distributions of SSCs computed from large amounts of speakers will be highly overlapped between different phones. Therefore, we introduce a speaker normalization technique into SSC computation to reduce the speaker variability. Experimental results on spontaneous speech recognition show that the speaker normalized SSCs are more useful as supplementary features for improving the recognition performance than the conventional SSCs. We observed a significant improvement in error rate by 20.3% and 14.3% at SNR=15dB by adding speaker normalized SSCs to the conventional features and by incorporating a speaker normalized technique into the conventional SSCs, respectively.
Brick/block absorbing walls with openings backed with porous materials are often used for sound absorptive treatments in buildings. This type of brick/block walls have a frequencyselective sound absorption at low frequencies, which is usually explained as the Helmholtz resonance. In addition, they often have peaks of sound absorption at frequencies much different from the Helmholtz resonance frequency. In this paper, the sound absorption mechanism of this type of walls was investigated by experiments. As a result, it has been confirmed that the sound absorption is caused by the effect of open-pipe resonance which happens in the openings. In the case of brick/block sound absorptive constructions made of materials with a thickness of about 10 cm, sound absorption caused by the open-pipe resonance is considerably significant at high frequencies. Therefore, when this type of sound absorption construction is used for room acoustic treatment, sound absorption at high frequencies must be carefully considered in acoustical design.
At the application of an Active Noise Control (ANC) for a duct with high speed flow whose velocity is around 12.0m/s (Reynolds number of 1.0×105), the following resultsfor the use of the meshes, as the simplest method for rectifying the flow, were given. The effect of ANC depends on the disturbance velocity attenuation factor of the rectifying meshes. The meshes should not be installed closer to the detection sensor than half the diameter of the duct. The meshes are more effective than the use of thicker sound absorbent materials on the sensors, although the numerical aperture of duct cross-section can be decreased both. The noise reduction effect by use of the meshes were around 3 to 4dB, without ANC under 200Hz, and 13dB of noise control effect at 500Hz with the ANC system with the meshes (These effects were measured 100mm from the duct outlet).