Sequential Organization of Speech in Reverberant Environments by Integrating Monaural Grouping and Binaural Localization

Abstract

Existing binaural approaches to speech segregation place an exclusive burden on location information. These approaches can achieve excellent performance in anechoic conditions but degrade rapidly in realistic environments where room reverberation corrupts localization cues. In this work we propose to integrate monaural and binaural processing to achieve sequential organization and localization of speech in reverberant environments. The proposed approach builds on monaural analysis for simultaneous organization, and combine sit with a novel method for generation of location-based cues in a probabilistic framework that jointly achieves localization and sequential organization. We compare sequential organization performance against a model-based system that uses only monaural cues and an exclusively binaural system, and localization performance against two existing methods that do not utilize monaural grouping. Results suggest that the proposed integration of monaural grouping and binaural localization allows for improved source localization and robust sequential organization performance in environments with considerable reverberation.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Jan 01, 2009
Accession Number
AD1001201

Entities

People

  • DeLiang Wang
  • John Woodruff

Organizations

  • Ohio State University

Tags

Communities of Interest

  • Energy and Power Technologies

DTIC Thesaurus Topics

  • Accuracy
  • Acoustics
  • Algorithms
  • Ambiguity
  • Architectural Acoustics
  • Cognitive Systems Engineering
  • Computer Science
  • Computers
  • Cross Correlation
  • Databases
  • Engineering
  • Free Field
  • Frequency
  • Frequency Bands
  • Speech
  • Time Intervals
  • Training

Fields of Study

  • Computer science

Readers

  • Auditory Neuroscience/Auditory Physiology.
  • Regression Analysis.
  • Systems Analysis and Design