Accession Number : ADA291270

Title :   Pattern-Analysis Based Models of Masking by Spatially Separated Sounds.

Descriptive Note : Final rept. 15 May 91-30 Sep 94,

Corporate Author : WRIGHT STATE UNIV DAYTON OH

Personal Author(s) : Gilkey, Robert H.

PDF Url : ADA291270

Report Date : 09 FEB 1995

Pagination or Media Count : 25

Abstract : Work is presented on masked detection, sound localization, neural networks, and the sense of presence. Both free-field and headphone- based studies of masking indicate that traditional models of binaural interaction may be inadequate to explain the reductions in masking that often occur with dichotic, as opposed to diotic, stimulation. The accuracy of localization judgments in the presence of a masker is determined by both the signal-to-noise ratio and the spatial location of the masker. Localization judgments for speech stimuli are in general less accurate than those for some nonspeech stimuli. Neural - network models of sound localization can achieve performance similar to human observers based on monaural information alone or based on interaural information alone. The reports of suddenly-deafened adults suggest that realistic auditory stimulation may be critical for determining the sense of presence in virtual environments.

Descriptors :   *MASKING, *HEARING, *AUDITORY ACUITY, *AUDITORY PERCEPTION, OBSERVERS, SPATIAL DISTRIBUTION, POSITION(LOCATION), STIMULATION(PHYSIOLOGY), NEURAL NETS, DETECTION, MODELS, HUMANS, INTERACTIONS, SIGNAL TO NOISE RATIO, ACCURACY, SOUND, ACOUSTIC SIGNALS, SEPARATION, SPEECH, STIMULI, ADULTS, PSYCHOACOUSTICS, EARPHONES, DEAFNESS.

Subject Categories : Anatomy and Physiology
      Psychology

Distribution Statement : APPROVED FOR PUBLIC RELEASE