25th SIGNAL PROCESSING AND COMMUNICATIONS APPLICATIONS CONFERENCE, 2017
15-18 MAY 2017
MARITIM PINE BEACH RESORT BELEK • ANTALYA / TURKEY
SUMMARY:
Spherical microphone arrays provide distinct advantages in 3D acoustic sensing and analysis problems owing to the spherical symmetry they possess. There are two types of spherical microphone arrays. The first type, rigid spherical microphone arrays, consist of microphones positioned on a rigid spherical surface. The second type, open spherical microphone arrays, consist of microphones suspended in free field on an open spherical shell. Past decade has seen progress in the development of algorithms for both types of arrays, and solutions provided to problems in sound source localisation, sound source separation and sound field analysis.
Sound field recorded by spherical microphone arrays can be analysed using two different ways. The first method, which is applicable to both open and rigid microphone arrays is based on the spherical harmonic decomposition of the sound field. The second method, applicable mainly to open spherical microphone arrays is based on the expression of the sound field inside the volume enclosed by the spherical surface using the Kirchhoff-Helmholtz (KH) integral. These methods differ in the way that they model the sound field. The former, aims to sample and reconstruct the sound field on the rigid sphere and the latter is based on the prediction of the sound field inside a spherical volume using its samples on the open sphere.
This tutorial aims to instruct its audience on the theoretical basis that underlies spherical microphone arrays and the application of acoustical signal processing methods on signals obtained from such arrays. The topics to be covered include: i) the fundamentals of energetic and spectral analysis of acoustic fields, ii) sound source localisation and source separation using spherical microphone arrays, and iii) physical limitations of spherical microphone arrays.
NOVELTY OF THE TOPICS COVERED:
The spherical symmetry affords spherical microphone arrays distinct advantages compared to linear microphone arrays. The spherical geometry allows simple representations of sound fields using the KH integral and spherical harmonic decomposition. The compact representations afforded by these arrays resulted in the design and development of spherical arrays both in academic and in commercial contexts. The uses of spherical microphone arrays encompass areas as diverse as sound source separation to architectural acoustics, from robot navigation to teleconferencing systems and from recording technologies to gunshot localisation.
TOPICS TO BE COVERED:
SHORT BIO OF THE SPEAKER:
Assoc. Prof. Dr Huseyin Hacihabiboglu graduated from METU Electrical and Electronic Engineering in 2000, got his M.Sc. from University of Bristol (UK) in 2001 on Electrical and Electronic Engineering, and his Ph.D. on Computer Science from Queen's University Belfast in 2004. Between 2004 and 2008, he was a postdoctoral research fellow at Centre for Communication Systems Research (CCSR) at University of Surrey and between 2008-2010 a senior postdoctoral research associate at Centre for Digital Signal Processing Research (CDSPR) at King's College London. Since 2012 he has been a member of staff at Graduate School of Informatics at METU. Dr Hacihabiboglu has been working on audio and acoustical signal processing for the past 16 years. Particularly, during the last decade he has worked on microphone array signal processing. He has more than 50 conference and journal publications and three patents in the audio signal processing area. He is the founder and the principal investigator of METU Spatial Audio Research Group (METU SPARG). He is a senior member of IEEE since 2012.