Sparse Sound Field Representation Using Complex Orthogonal Matching Pursuit

Spatial audio reproduction and translation for virtual, augmented, and extended reality applications require an efficient representation of the recorded sound fields. In this paper, we investigate the possible sparse representations of the sound field recorded by multiple microphones in reverberant...

Full description

Saved in:
Bibliographic Details
Main Authors: Xu, Shaoheng, Zhang, Jihui Aimee, Abhayapala, Thushara D., Bastine, Amy, Lai, Wei-Ting, Samarasinghe, Prasanga N.
Format: Conference Proceeding
Language:English
Subjects:
Complex Orthogonal Matching Pursuit
Matching pursuit algorithms
Microphone arrays
Recording
Sound Field Reproduction
Sparse approximation
Sparse Representation
Sparse Sound Field Decomposition
Spatial audio
Speech processing
Transforms
Online Access:Request full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Spatial audio reproduction and translation for virtual, augmented, and extended reality applications require an efficient representation of the recorded sound fields. In this paper, we investigate the possible sparse representations of the sound field recorded by multiple microphones in reverberant environments. We combine the Complex Orthogonal Matching Pursuit (COMP) algorithm with the concept of distributed virtual sound sources to propose a sparse sound field representation. The technique uses recordings from a grid of microphones and transforms them into a sparse representation featuring a selected set of active virtual sources. Using simulation, we evaluate the proposed COMP approach with LASSO and IRLS methods in a reverberant room of regular size with a ceiling-mounted microphone array.
ISSN:2379-190X
DOI:10.1109/ICASSP48485.2024.10445940