Volume 134, Issue 3, September 2013
Index of content:
- ACOUSTIC SIGNAL PROCESSING 
Optimal irregular microphone distributions with enhanced beamforming performance in immersive environments134(2013); http://dx.doi.org/10.1121/1.4816540View Description Hide Description
Complex relationships between array gain patterns and microphone distributions limit the application of optimization algorithms on irregular arrays. This paper proposes a Genetic Algorithm (GA) for microphone array optimization in immersive (near-field) environments. Geometric descriptors for irregular arrays are proposed for use as objective functions to reduce optimization time by circumventing the need for direct array gain computations. In addition, probabilistic descriptions of acoustic scenes are introduced for incorporating prior knowledge of the source distribution. To verify the effectiveness of the proposed optimization, signal-to-noise ratios are compared for GA-optimized arrays, regular arrays, and arrays optimized through direct exhaustive simulations. Results show enhancements for GA-optimized arrays over arbitrary randomly generated arrays and regular arrays, especially at low microphone densities where placement becomes critical. Design parameters for the GA are identified for improving optimization robustness for different applications. The rapid convergence and acceptable processing times observed during the experiments establish the feasibility of this approach for optimizing array geometries in immersive environments where rapid deployment is required with limited knowledge of the acoustic scene, such as in mobile platforms and audio surveillance applications.
134(2013); http://dx.doi.org/10.1121/1.4816545View Description Hide Description
During the last decade, the aeroacoustic community has examined various methods based on deconvolution to improve the visualization of acoustic fields scanned with planar sparse arrays of microphones. These methods assume that the beamforming map in an observation plane can be approximated by a convolution of the distribution of the actual sources and the beamformer's point-spread function, defined as the beamformer's response to a point source. By deconvolving the resulting map, the resolution is improved, and the side-lobes effect is reduced or even eliminated compared to conventional beamforming. Even though these methods were originally designed for planar sparse arrays, in the present study, they are adapted to uniform circular arrays for mapping the sound over 360°. This geometry has the advantage that the beamforming output is practically independent of the focusing direction, meaning that the beamformer's point-spread function is shift-invariant. This makes it possible to apply computationally efficient deconvolution algorithms that consist of spectral procedures in the entire region of interest, such as the deconvolution approach for the mapping of the acoustic sources 2, the Fourier-based non-negative least squares, and the Richardson–Lucy. This investigation examines the matter with computer simulations and measurements.
134(2013); http://dx.doi.org/10.1121/1.4817835View Description Hide Description
Distributed underwater sensors are expected to provide oceanographic monitoring over large areas. As fabrication technology advances, low cost sensors will be available for many uses. The sensors communicate to each other and are networked using acoustic communications. This paper first studies the performance of such systems for current measurements using tomographic inversion approaches to compare with that of a conventional system which distributes the sensors on the periphery of the area of interest. It then proposes two simple signal processing methods for ocean current mapping (using distributed networked sensors) aimed at real-time in-buoy processing. Tomographic inversion generally requires solving a challenging high dimensional inverse problem, involving substantial computations. Given distributed sensors, currents can be constructed locally based on data from neighboring sensors. It is shown using simulated data that similar results are obtained using distributed processing as using conventional tomographic approaches. The advantage for distributed systems is that by increasing the number of nodes, one gains a much more improved performance. Furthermore, distributed systems use much less energy than a conventional tomographic system for the same area coverage. Experimental data from an acoustic communication and networking experiment are used to demonstrate the feasibility of acoustic current mapping.
134(2013); http://dx.doi.org/10.1121/1.4817876View Description Hide Description
For the derivation of 2.5-dimensional operator in wave field synthesis, a virtual source is assumed to be positioned far from a loudspeaker array. However, such far-field approximation inevitably results in a reproduction error when the virtual source is placed adjacent to an array. In this paper, a method is proposed to generate a virtual source close to and behind a continuous line array of loudspeakers. A driving function is derived by reducing a surface integral (Rayleigh integral) to a line integral based on the near-field assumption. The solution is then combined with the far-field formula of wave field synthesis by introducing a weighting function that can adjust the near- and far-field contribution of each driving function. This enables production of a virtual source anywhere in relation to the array. Simulations show the proposed method can reduce the reproduction error to below −18 dB, regardless of the virtual source position.