Auditory backward recognition masking with well-practiced listeners.

Five adult listeners who had extensive previous practice on a pitch discrimination task participated in an auditory backward recognition masking (ABRM) experiment. S identified the relative pitch "higher"-"lower" of a test tone. A standard tone of 20 msec duration at 800 c/s was followed after 500 msec by a similar 20-msec test tone at 800 +/- delta f. delta f varied from 3.3-11.4 c/s across Ss, adjusted for each S to yield approximately 75%-correct performance. The test tone was followed, after a silent interval of 20-250 msec, by a 100-msec mask, either a tone of 800 c/s or white noise. All stimuli were approximately equal to 77 dbA. All Ss exhibited substantial ABRM when the backward mask was a tone but substantially less when it was white noise. Evidently it is not necessary to use relatively naive listeners to demonstrate ABRM. The performance of most Ss was reduced by the tone masker (i.e., ABRM was in evidence) to chance or nearly chance levels (i.e., little or no ABRM) until the ISI exceeded 80 msec, and for 3 of the Ss ABRM with the tone masker was still evident after 250 msec. It was concluded that a backward mask interferes with test-tone information held in a multichannel preperceptual store and that such interference depends, in part, on the degree to which the properties of the test tone and mask overlap.