Spatial Hearing, Attention And Informational Masking In Speech Identification
Abstract
The work has two primary objectives: to understand the nature of the "masking" or interference one source causes on another source and how masking can be reduced by spatially separating the sound sources, and to understand how humans use predictability and context to overcome the masking that occurs among sequences of sounds such as streams of speech. The PI will conduct a series of human speech recognition experiments that test hypotheses about the factors that provide a release from "informational masking" when sound sources are spatially separated. The foundation upon which this line of work is built is a theory of auditory masking that posits two separate components: peripherally-based "energetic masking" (EM) and centrally-based "informational masking". The goal is to advance understanding of the theory of masking and how binaural/spatial cues are used by human listeners to segregate and attend to a target talker in competition with masking talkers.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Dec 05, 2016
- Source ID
- FA95501610372
Entities
People
- Gerald Kidd
Organizations
- Air Force Office of Scientific Research
- Boston University
- United States Air Force