Mechanisms of Temporal Pattern Discrimination by Human Observers
Abstract
The project's first component studied how human listeners discriminate two temporal patterns of tones. The stimuli were non-speech, non- musical, tone sequences that conveyed arhythmic or partly rhythmic time patterns. The results indicated that the temporal pattern discrimination process depends on the timing of the sequences and whether they overlap in time. Listeners can perform the task very well when the patterns do not overlap and are presented to separate ears and/or at different tone frequencies. At long time separations, the listener's discrimination mechanism reduces the input information to two lists of intertone times, and the decision is based on the correlation between the lists (a temporal correlation process). At short time separations, the mechanism computes statistics based on the summed envelope of the two input waveforms (a single channel process). The project's second component consisted of analytical and computer modeling of multi-element detection systems. New results were obtained on the performance of statistically optimal detector arrays and on arrays that combine their binary outputs. The project's third component studied how observers process information in multi- element, visual signal detection. The results describe how performance depends on display element reliability, coding and arrangement. Auditory perception, Temporal pattern discrimination group signal detection, Signal detection with multi-element visual displays.
Document Details
- Document Type
- Technical Report
- Publication Date
- Feb 15, 1994
- Accession Number
- ADA277407
Entities
People
- Robert D. Sorkin
Organizations
- University of Florida