In this paper, we propose an automatic method of designing digital filters for three-dimensional (3-D) sound movement that is dedicated to embedded applications. By this method, read-only memory (ROM) capacity and computational load are reduced with only slight degradation of the 3-D sound effect. For practical applications of 3-D sound, the continuous movement of a virtual sound image is indispensable. To achieve 3-D sound movement, the interpolation of filter coefficients is often required; however, the frequency response of an intermediate filter obtained by interpolation is severely distorted using conventional automatic design methods. The proposed method reduces this distortion by evaluating the response of the intermediate filters in the optimization process. The results of objective evaluation and subjective listening tests show that the proposed method improves the perceptual quality of 3-D sound movement, even with runtime interpolation of filter coefficients.
The Biot-Stoll model is useful for analyzing the acoustic wave propagation in porous marine sediments. However, the practical application of the model is not easy, because 13 physical parameters are required to apply it, and the physical phenomena in this model are difficult to understand. In this study, equivalent circuits for the plane longitudinal wave propagation in porous media are proposed for easy understanding of the physical phenomena. Also, practical approximated equations for the longitudinal wave velocity and attenuation are proposed to simplify practical application of the Biot-Stoll model. Then, simpler models, such as the Wood model, the Gassmann model, and the Biot-Stoll model without frame moduli can be used within limited ranges of frequency and porosity to approximate the exact Biot-Stoll model. Therefore, the ranges of frequency and porosity applicable to some simple acoustic models are obtained to simplify practical application of the Biot-Stoll model.
Listening tests were carried out for investigating the localization judgments of untrained subjects through equalized headphones and with HRTF synthesis. The investigation was made on the basis of the former ‘Graphical User Interface for Blind Persons’ project in order to determine the possibilities of a 2D virtual sound screen and headphone playback. 50 untrained subjects evaluated a virtual audio display in front of the listener using different horizontal and vertical resolutions on a 2D surface. A listening test using white and filtered noise signals was followed by a special investigation using simple high-pass and low-pass filtering of the original sound in order to increase correct vertical localization judgments. The simulation uses high-pass filtering for higher elevations and low-pass filtering for lower elevations in a 5×2 and a 3×3 spatial resolution. Results of the listening test will be presented and the efficiency of the filtering in correct localization judgments will be discussed.
The effects of baseline performance on identification of English /r/ and /l/ in noise by Japanese listeners were examined. Japanese and American listeners’ perception of /r/ and /l/ was measured under various signal-to-noise ratios. Each Japanese listener’s baseline performance was also observed with an identification test with a large set of stimuli. Generally, the signal-to-noise ratio had a similar effect across Japanese listeners of different baseline performances, although the overall accuracy was dominated by the baseline. When fifteen-days of /r/–/l/ identification training in quiet was applied to a group of Japanese listeners, the training effect generalized to identification performance in noise. However, the performance of Japanese listeners did not reach the level that the native listeners exhibited.
The finite-difference time-domain (FDTD) numerical method has the unique characteristic that it can separately calculate the far-field scattered waves and near-field total waves at the same time, where total waves refer to a mixture of incident plane waves and scattered waves. This formulation is called the total-field/scattered-field (TF/SF) formulation and is very effective for the analysis of scattering problems in nondestructive evaluation (NDE). In this report, a TF/SF formulation of the elastic wave fields in solids is described. Specifically, scalar and vector potentials are used for the formulation. Using these potentials is beneficial for two reasons. First, the longitudinal waves and shear waves are initially separated. Therefore, the scattering phenomenon can be clearly recognized. Second, it is facile to use the potentials for setting the absorbing boundaries. The FDTD TF/SF formulation using potentials has been proved effective in the analysis of the scattering of a longitudinal plane wave incident to a square hole.
Formant frequencies are very important speech features in the area of speech perception and practical application. Although some formant estimation methods have been proposed previously, those have been inferior in real-time operation. Previously, we have proposed a new type of formant estimation based on IFC (Inverse Filter Control) method and applied into speech analysis and application systems in our study. Since the IFC method is based on speech waveform processing, the realization of a real-time system has been expected. On the other hand, recently the cheap electric boards which mounted a very high-speed DSP (Digital Signal Processor) have been marketed. We have applied such a DSP board into developing a real-time formant tracker based on the IFC method. By correcting the original C-language program for DSP operation, we have confirmed the real-time operation, which have estimated four formant frequencies from F1 to F4. Moreover, those estimates were almost the same as those by the original PC simulation. In this paper, we describe system modification for DSP realization and discuss effects of software modifications from C-language into DSP program.