The localization of acoustic sound sources is beneficial to signal processing applications of speech enhancement, dereverberation, separation and tracking. Difficulties in position estimation arise in real world environments due to coherent reflections degrading performance of subspace localization techniques. This paper proposes a method of multiple signal classification (MUSIC) subspace localization, which is suitable for reverberant rooms. The method is based on the modal decomposition of a room’s region-to-region transfer function, which is assumed to be known. We perform a numerical simulation of four sound sources in a reverberant room, and show that the localization method exhibits increased spatial resolution and distance focusing abilities when the region-to-region transfer function is incorporated. The proposed method is successful in estimating three-dimensional sound source positions without distortion due to reverberation.

We proposed a spatial coherence-based PSD estimation and source separation technique in [1] using a 32-channel spherical microphone array. While the proposed spherical microphone-based method exhibited a satisfactory performance in separating multiple sound sources in a reverberant environment, the use of a large number of microphones remains an issue for some practical considerations. In this paper, we investigate an alternative array structure to achieve spatial coherence-based source separation using a planar microphone array. This method is particularly useful in separating a limited number of sound sources in a mixed acoustic scene. The simplified array structure we used here can easily be integrated with many commercial acoustical instruments such as smart home devices to achieve better speech enhancements.

The direct-to-reverberant ratio (DRR) is an important characterization of a reverberant environment. This paper presents a novel blind DRR estimation method based on the coherence function between the sound pressure and particle velocity at a point. First, a general expression of coherence function and DRR is derived in the spherical harmonic domain, without imposing assumptions on the reverberation. In this paper, DRR is expressed in terms of the coherence function as well as two parameters that are related to statistical characteristics of the reverberant environment. Then, a method to estimate the values of these two parameters using a microphone system capable of capturing first-order spherical harmonics is proposed, under three assumptions which are more realistic than the diffuse field model. Furthermore, a theoretical analysis on the use of plane wave model for direct path signal and its effect on DRR estimation is presented, and a rule of thumb is provided for determining whether the point source model should be used for the direct path signal. Finally, the ACE challenge dataset is used to validate the proposed DRR estimation method. The results show that the average full band estimation error is within 2 dB, with no clear trend of bias.

The direct-to-reverberant ratio (DRR), which describes the energy ratio between the direct and reverberant component of a soundfield, is an important parameter in many audio applications. In this paper, we present a multichannel algorithm, which utilizes the blind recordings of a spherical microphone array to estimate the DRR of interest. The algorithm is developed based on a spatial correlation model formulated in the spherical harmonics domain. This model expresses the cross correlation matrix of the recorded soundfield coefficients in terms of two spatial correlation matrices, one for direct sound and the other for reverberation. While the direct path arrives from the source, the reverberant path is considered to be a nondiffuse soundfield with varying directional gains. The direct and reverberant sound energies are estimated from the aforementioned spatial correlation model, which then leads to the DRR estimation. The practical feasibility of the proposed algorithm was evaluated using the speech corpus of the acoustic characterization of environments challenge. The experimental results revealed that the proposed method was able to effectively estimate the DRR of a large collection of reverberant speech recordings including various environmental noise types, room types and speakers.