Spatial audio systems are commonly utilized for sound-field analysis and reconstruction or signal enhancement algorithm. In order to provide realism and immersion with such systems it is crucial to improve the amount of information that is delivered to a listener and enhance the naturalness of the reproduced sound field. The acquisition, enhancement and reproduction of a sound field is performed with microphone arrays. Microphone array signal processing techniques provide an essential tool in multichannel signal processing applications and have gained more attention over the past years due to the advancement in computational power and the potential applications. Traditional enhancement algorithms are the beamformers which optimally combine the sensor signals from a microphone array and minimize the noise level, while retaining the signal arriving from a desired direction. Another class of signal enhancement algorithms, the post filters can further improve the performance of such beamformers, in terms of noise suppression. Post filters provide an additional noise reduction but they rely on the performance of the beamformer for spatial filtering.
A novel technique that provides additional beamforming capabilities and further noise reduction is the cross-pattern coherence (CroPaC). CroPaC is a parametric spatial filtering technique that utilizes coherence-based measures between signals from microphones of different orders as a criterion for focusing in specific directions. A spatial filter is calculated from a set of directional microphones with their positive phases and maximum magnitude response towards the same direction and it assigns attenuation values to the output of a robust beamformer. An instantaneous signal-to-noise ratio based adjustment of the smoothing parameters and spectral floor adjustments are utilized for the spatial filter to mitigate modulation artifacts such as musical noise. The effectiveness of the algorithm is demonstrated with instrumental measures, listening tests and sound examples with a prototype microphone array in a multi-talker scenario with several background noise levels.
Symeon Delikaris-Manias received his B.Sc. degree in mathematics from the University of Crete, Heraklion, Greece in 2006 and his M.Sc. degree in sound and vibration from the Institute of Sound and Vibration Research (ISVR), Southampton, UK, in 2008 with a thesis on inverse-filtering methods and cross-talk cancellation systems. Between 2008-2010, he was employed by GSacoustics as an acoustic consultant, and was responsible for acoustic modeling, development of sound reproduction techniques and auralization. In
2010-2011 he was at the Center for Virtual Reality, Brest, France developing and evaluating sound-field recording and reproduction techniques for ship simulators. He is currently pursuing his doctoral degree in Electrical Engineering at Aalto University, Espoo, Finland and his research focuses on parametric spatial audio coding and microphone array signal processing techniques.