A translation method for the spherical harmonics expansion coefficients of a sound field using plane wave expansions is proposed. It is based on the decomposition of a plane wave in the spherical harmonics domain and without the use of the spherical harmonics addition theorem, thus is very computationally efficient. Simulations are conducted for validation. The stabilities of the translations are compared with the conventional method in terms of matrix condition numbers. The proposed method is demonstrated to be more robust as the frequency increases and when upscaling the coefficients. Besides, the computation is much faster when high truncated orders are solved.

A spherical microphone array has proved effective in reconstructing an enclosed sound field by a superposition of spherical wave functions in Fourier domain. It allows successful reconstructions surrounding the array, but the accuracy will be degraded at a distance. In order to extend the effective reconstruction to the entire cavity, a plane-wave basis in space domain is used owing to its non-decaying propagating characteristic and compared with the conventional spherical wave function method in a low frequency sound field within a cylindrical cavity. The sensitivity to measurement noise, the effects of the numbers of plane waves, and measurement positions are discussed. Simulations show that under the same measurement conditions, the plane wave function method is superior in terms of reconstruction accuracy and data processing efficiency, that is, the entire sound field imaging can be achieved by only one time calculation instead of translations of local sets of coefficients with respect to every measurement position into a global one. An experiment was conducted inside an aircraft cabin mock-up for validation. Additionally, this method provides an alternative possibility to recover the coefficients of high order spherical wave functions in a global coordinate system without coordinate translations with respect to local origins.

Active control of low frequency sound radiation using planar secondary sources is theoretically investigated in this paper. The primary sound field originates from a vibrating panel and the planar sources are modeled as simply supported rectangular panels in an infinite baffle. The sound power of the primary and secondary panels are calculated using a near field approach, and then a series of formulas are derived to obtain the optimum reduction in sound power based on minimization of the total radiate sound power. Finally, active reduction for a number of secondary panel arrangements is examined and it is concluded that when the modal distribution of the secondary panel does not coincide with that of the primary panel, one secondary panel is sufficient. Otherwise four secondary panels can guarantee considerable reduction in sound power over entire frequency range of interest.

The application of sound quality in household appliances has gradually increased in recent years. In addition to modeling algorithms, appropriate acoustic metrics that characterize product sounds also play an important role in developing models. In this study, an artificial neural network based sound quality model for range hood noise was established with the combination of prior metric selection by multidimensional scaling (MDS) analysis of perceptual dissimilarities. First, sounds in different environments, speeds, and positions were recorded, and their annoyance was evaluated by grouped anchor semantic differential subjective jury testing. Then, the timbre space underlying dissimilarity judgments were analyzed by CLASCAL, an accurate MDS algorithm. Each dimension of the space was well explained by some metrics through stepwise regression. Finally, a sound quality model was established based on a back propagation neural network (BPNN). Results show that the combination of BPNN and CLASCAL can address the interpretation of the sound quality model and the ability to model nonlinearity for high accuracy. In addition, the application of noise control on range hoods showed that passive and active noise control (ANC) measures improve sound quality, especially ANC systems.

Thanks to the development of deep learning, various sound source separation networks have been proposed and made significant progress. However, the study on the underlying separation mechanisms is still in its infancy. In this study, deep networks are explained from the perspective of auditory perception mechanisms. For separating two arbitrary sound sources from monaural recordings, three different networks with different parameters are trained and achieve excellent performances. The networks’ output can obtain an average scale-invariant signal-to-distortion ratio improvement (SI-SDRi) higher than 10 dB, comparable with the human performance to separate natural sources. More importantly, the most intuitive principle—proximity—is explored through simultaneous and sequential organization experiments. Results show that regardless of network structures and parameters, the proximity principle is learned spontaneously by all networks. If components are proximate in frequency or time, they are not easily separated by networks. Moreover, the frequency resolution at low frequencies is better than at high frequencies. These behavior characteristics of all three networks are highly consistent with those of the human auditory system, which implies that the learned proximity principle is not accidental, but the optimal strategy selected by networks and humans when facing the same task. The emergence of the auditory-like separation mechanisms provides the possibility to develop a universal system that can be adapted to all sources and scenes.