Auditory Spectro-Temporal Pattern Analysis

Abstract

The present research is a continuation of an AFOSR-sponsored program that has supported our research on CMR essentially since the discovery of the phenomenon. Over the past six years, the research program has revealed several important findings related to CMR. 1) The phenomenon is based primarily upon an across-frequency analysis of amplitude envelope. 2) All portions of the envelope do not appear to contribute equally to CMR: envelope dips carry the critical information; 3) CMR increases with the number of comodulated noise bands present, with diminishing returns after three bands are present; 4) The presence of non-comodulated bands within a set of comodulated bands can substantially decrease CMR; 5) When auditory grouping principles are applied to segregate the noncomodulated bands from the comodulated bands, CMR can be restored to a significant extent; 6) CMR occurs for FM, but a much smaller magnitude than for AM. 7) Part of the MLD appears to be based upon a process that is essentially identical to CMR.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Dec 31, 1993
Accession Number
ADA273554

Entities

People

  • Joseph W. Hall

Organizations

  • University of North Carolina at Chapel Hill

Tags

Communities of Interest

  • Materials and Manufacturing Processes

DTIC Thesaurus Topics

  • Abstracts
  • Amplitude
  • Background Noise
  • Classification
  • Detection
  • Frequency
  • Monitoring
  • Narrowband
  • Noise
  • North Carolina
  • Numbers
  • Security
  • Square Roots

Readers

  • Speech Processing/Speech Recognition.
  • Theoretical Analysis.