Temporal and spectral determinants of informational masking in tonal patterns

Publisher Website
Google Scholar
Abstract
A form of “informational masking” is observed when listeners attempt to detect single target components of word-length tonal patterns [J. Acoust. Soc. Am. Suppl. 1 64, S39 (1978)]. These masking-like effects, sometimes as large as 50–60 dB, occur under conditions of high stimulus uncertainty, when contextual patterns and the temporal and spectral positions of target tones are varied from trial-to-trial, but are absent under conditions of minimal psychophysical uncertainty. Stimulus determinants of informational masking were investigated in three experiments, by (a) varying the duration of the contextual patterns from 0 to 400 ms, (b) varying the duration of the target tones from 20 to 160 ms, and (c) varying the spectral range for the contextual tones from 1 to 2700 Hz. Individual differences in asymptotic detection thresholds, and in training times required to approach those thresholds, are unusually large in these tasks. However, the data are generally characterized by the following rules. Informational masking approaches asymptotic levels for context durations in excess of 160 ms. Little or no masking occurs for target tones longer than 60–80 ms. Similar average amounts of masking were observed for each of the spectral ranges of the context tones. [Work supported by NIH.]

This publication has 0 references indexed in Scilit: