This paper proposes real-time sound source orientation estimation based on orientation-extended amplitude beamforming (OE-ABF). To recognize a sound source orientation (such as face orientation) is an important function for a robot who can achieve natural human-robot interaction because the function is required to distinguish the human target from a robot or another person. We developed a sound source orientation system using orientation-extended beamforming (OE-BF) and showed the system worked properly at least under a specific controlled environment. However, in practical use, this system does not work properly because the system doesn't take into account the differences between the supposed model in OE-BF and in practical situations. For example, the system model supposes that there is neither noise nor reverberation, however, this is not a realistic assumption. To solve this assumption mismatch problem, we propose sound source orientation estimation based on OE-ABF, and constructed a real-time sound source orientation estimation system with the proposed method using a 96ch microphone array. Evaluation results of our proposed system show that the average error of estimated angles is lower than 5°, while the error of our previously reported system was greater than 20°. With this system, the robot is able to distinguish that the utterance target of a person standing 1m in front is itself or another person standing 0.2m to the left of the robot. This is valuable for human-robot interaction.

The sound data for open evaluation is necessary for studies such as sound source localization, sound retrieval, sound recognition and hands-free speech recognition in real acoustic environments. This paper reports on our project for acoustic data collection. There are many kinds of sound scenes in real environments. The sound scene is specified by sound sources and room acoustics. The number of combinations of the sound sources, source positions and rooms is huge in real acoustic environments. We assumed that the sound in the environments can be simulated by convolution of the isolated sound sources and impulse responses. As an isolated sound source, hundred kinds of environment sounds and speech sounds are collected. The impulse responses are collected in various acoustic environments. Additionally we collected sounds from a moving source. In this paper, progress of our sound scene database collection project and application to environment sound recognition and hands-free speech recognition are described.

A method is proposed for modeling a room transfer function (RTF) by using common acoustical poles and their residues. The common acoustical poles correspond to the resonance frequencies (eigenfrequencies) of the room, so they are independent of the source and receiver positions. The residues correspond to the eigenfunctions of the room. Therefore, the residue, which is a function of the source and receiver positions, can be expressed using simple analytical functions for rooms with a simple geometry such as rectangular. That is, the proposed model can describe RTF variations using simple residue functions. Based on the proposed common-acoustical-pole and residue model, methods are also proposed for spatially interpolating and extrapolating RTFs. Because the common acoustical poles are invariant in a given room, the interpolation or extrapolation of RTFs is reformulated as a problem of interpolating or extrapolating residue values. The experimental results for a rectangular room, in which the residue values are interpolated or extrapolated by using a cosine function or a linear prediction method, demonstrate that unknown RTFs can be well estimated at low frequencies from known (measured) RTFs by using the proposed methods.

Use of a common-acoustical-pole and zero model is proposed for modeling head-related transfer functions (HRTFs) for various directions of sound incidence. The HRTFs are expressed using the common acoustical poles, which do not depend on the source directions, and the zeros, which do. The common acoustical poles are estimated as they are common to HRTFs for various source directions; the estimated values of the poles agree well with the resonance frequencies of the ear canal. Because this model uses only the zeros to express the HRTF variations due to changes in source direction, it requires fewer parameters (the order of the zeros) that depend on the source direction than do the conventional all-zero or pole/zero models. Furthermore, the proposed model can extract the zeros that are missed in the conventional models because of pole-zero cancellation. As a result, the directional dependence of the zeros can be traced well. Analysis of the zeros for HRTFs on the horizontal plane showed that the nonminimum-phase zero variation was well formulated using a simple pinna-reflection model. The common-acoustical-pole and zero (CAPZ) model is thus effective for modeling and analyzing HRTF's.

A new configuration for a stereo echo canceller with nonlinear pre-processing is proposed. The pre-processor which adds uncorrelated components to the original received stereo signals improves the adaptive filter convergence even in the conventional configuration. However, because of the inaudibility restriction, the preprocessed signals still include a large amount of the original stereo signals which are often highly cross-correlated. Therefore, the improvement is limited. To overcome this, our new stereo echo canceller includes exclusive adaptive filters whose inputs are the uncorrelated signals generated in the pre-processor. These exclusive adaptive filters converge to true solutions without suffering from cross-correlation between the original stereo signals. This is demonstrated through computer simulation results.

A new blind deconvolution method is proposed for recovering an unknown source signal, which is observed through two unknown channels characterized by non-minimum phase impulse response filters. Conventional methods cannot estimate the non-minimum phase parts. Our method is based on computing the eigenvector corresponding to the smallest eigenvalue of the input correlation matrix and using a cost function to determine the order of the impulse response filter model. Multi-channel inverse filtering with the estimated impulse responses is used to recover the unknown source signal. Sub-band processing is also used to reduce the complexity of dealing with long impulse responses such as room impulse responses. Computer simulation shows that the effectiveness of our method.

This paper proposes a microphone array system which realizes the following important functions for speech recognition: (i) SNR improvement, (ii) flat spectrum response for an arbitrary speaker position, and (iii) speech period detection in noisy speech. This microphone array system features time delay estimation using prewhitening signal processing, an optimally weighted delay-and-sum array, and speech period detection (called MLD) based on the level difference between signals before and after array processing. Word recognition experiments performed in the presence of crowd noise demonstrate that the proposed system has great robustness against noise than does the system with a conventional directional microphone and a speech period detection method.

A multiple-point equalization filter using the common acoustical poles of room transfer functions is proposed. The common acoustical poles correspond to the resonance frequencies, which are independent of source and receiver positions. They are estimated as common autoregressive (AR) coefficients from multiple room transfer functions. The equalization is achieved with a finite impulse response (FIR) filter, which has the inverse characteristics of the common acoustical pole function. Although the proposed filter cannot recover the frequency response dips of the multiple room transfer functions, it can suppress their common peaks due to resonance; it is also less sensitive to changes in receiver position. Evaluation of the proposed equalization filter using measured room transfer functions shows that it can reduce the deviations in the frequency characteristics of multiple room transfer functions better than a conventional multiple-point inverse filter. Experiments show that the proposed filter enables 1-5 dB additional amplifier gain in a public address system without acoustic feedback at multiple receiver positions. Furthermore, the proposed filter reduces the reflected sound in room impulse responses without the pre-echo that occurs with a multiple-point inverse filter. A multiple-point equalization filter using common acoustical poles can thus equalize multiple room transfer functions by suppressing their common peaks.

Of the many adaptive filtering algorithms, the normalized LMS (NLMS) algorithm is generally used in practice because of its simplicity. The computational complexity of the NLMS algorithm is low, however, convergence is very slow and tracking is poor for a colored input signal such as speech. The projection algorithm was proposed as a generalization of the NLMS algorithm. This paper provides a fast projection algorithm and a step size control to obtain the same steady-state excess mean squared error (MSE) for various projection orders. Computer simulations for colored noise and speech input signal confirm the effectiveness of the projection algorithm and the step size control.

The paper proposes a new adaptive algorithm (called the ES-RLS algorithm) with double the convergence speed of the conventional RLS algorithm. Makino et al. (1993) showed that the variation of a room impulse response becomes progressively smaller along the series by the same exponential ratio as the impulse response. The ES-RLS algorithm is derived by incorporating these variation characteristics into the conventional RLS algorithm using Kalman filter theory, which gives physical meaning to the RLS algorithm. The ES-RLS algorithm adjusts coefficients with large errors in large steps and coefficients with small errors in small steps. Computer simulations demonstrated that the new adaptive algorithm converged twice as fast as the conventional RLS algorithm.>

A new model for a room transfer function (RTF) by using common acoustical poles that correspond to resonance properties of a room is proposed. These poles are estimated as the common values of many RTF's corresponding to different source and receiver positions. Since there is one-to-one correspondence between poles and AR coefficients, these poles are calculated as common AR coefficients by two methods: (i) using the least squares method, assuming all the given multiple RTF's have the same AR coefficients and (ii) averaging each set of AR coefficients estimated from each RTF. The estimated poles agree well with the theoretical poles when estimated with the same order as the theoretical pole order. When estimated with a lower order than the theoretical pole order, the estimated poles correspond to the major resonance frequencies, which have high Q factors. Using the estimated common AR coefficients, the proposed method models the RTF's with different MA coefficients. This model is called the common-acoustical-pole and zero (CAPZ) model, and it requires far fewer variable parameters to represent RTF's than the conventional all-zero or pole/zero model. This model was used for an acoustic echo canceller at low frequencies, as one example. The acoustic echo canceller based on the proposed model requires half the variable parameters and converges 1.5 times faster than one based on the all-zero model, confirming the efficiency of the proposed model.>

A normalized least-mean-squares (NLMS) adaptive algorithm with double the convergence speed, at the same computational load, of the conventional NLMS for an acoustic echo canceller is proposed. This algorithm, called the ES (exponentially weighted stepsize) algorithm, uses a different stepsize (feedback constant) for each weight of an adaptive transversal filter. These stepsizes are time-invariant and weighted proportionally to the expected variation of a room impulse response. The algorithm adjusts coefficients with large errors in large steps, and coefficients with small errors in small steps. A transition formula is derived for the mean-squared coefficient error of the algorithm. The mean stepsize determines the convergence condition, the convergence speed, and the final excess mean-squared error. Modified for a practical multiple DSP structure, the algorithm requires only the same amount of computation as the conventional NLMS. The algorithm is implemented in a commercial acoustic echo canceller, and its fast convergence is demonstrated.>

A method for modeling a room transfer function (RTF) by using estimated common acoustical poles that correspond to resonance properties of a room is proposed. These poles are estimated as common values of the multiple RTFs corresponding to different source and receiver positions. This common-acoustical-pole and zero (CAPZ) model requires far fewer variable parameters to represent RTFs than conventional all-zero or pole/zero models. This model was applied to an acoustic echo canceller and to head-related transfer functions. At low frequencies, the acoustic echo canceller based on this model converges 1.5 times faster than the one based on the all-zero model. Head-related transfer functions that have resonance characteristics of the external ear are also successfully modeled by the proposed model.>

A NLMS (normalized LMS) adaptive algorithm is proposed for an acoustic echo canceller with double the convergence speed of but the same computational load as the conventional NLMS. This algorithm, called the ES (exponential step) algorithm, implements a different value of step gain (feedback factor) for each tap coefficient of a canceller. These step gains are weighted exponentially along the series by the same exponential ratio as the expected variation in a room impulse response. This algorithm is implemented in a commercial subband echo canceller, and its superiority to the conventional algorithm is demonstrated.>

A novel theorem is proposed for realizing exact inverse-control of a linear finite-impulse-response (FIR) system. It is named the multiple-input/output inverse theorem (MINT). It is impossible to realize exact inverse-control of a nonminimum-phase system with previously reported methods. However, even in such a case, inverse-control can be realized by using the MINT. According to this theorem, the exact inverse is constructed from multiple FIR filters (transversal filters) by setting extra signal-transmission channels to the linear system. The coefficients of these FIR filters can be computed by the well-known rules of matrix algebra. The MINT can be applied to room acoustics by using multiple loudspeakers or microphones. An experiment is made for realizing inverse-control in a sound-field. It is ascertained that The MINT is grately superior to previous methods which use only one signal-transmission channel.