Reconstruction Of The Plenacoustic Function Using Compressive Sampling

Haiying Xia,Rendong Ying,Lingling Zhou
DOI: https://doi.org/10.1117/12.913451
2011-01-01
Abstract:A compressive sampling scheme for reconstruction of the plenacoustic function (PAF) in a closed space is presented, showing that the PAF can be recovered precisely with far less samples. The spatio-temporal PAF is sampled on a circular microphone array. For each fixed sampling angle in the circle, a proper sampling kernel (e.g. kernels which can reproduce polynomials, exponential splines, etc.) leads to precise reproduction of the PAF and dramatically reduction of the number of samples along temporal axis. The decay of spectrum along spatial axis is analyzed, in order to avoid aliasing of spatial reconstruction. A parametric method to find the whole expression of the 2-D PAF is also studied. Finally, experimental results are presented; it shows a low normalized mean-square-error (MSE) on the order of -32dB in practice.
What problem does this paper attempt to address?