The discrimination of samples of noise in monotic, diotic, and dichotic conditions.

Listeners' ability to discriminate between independent tokens of 800-Hz-wide bands of noise (centered at 1200 Hz) was measured in monaural, diotic, and several dichotic conditions. In a two-cue, two-alternative, temporal forced-choice task, listeners indicated whether the second or third interval contained a token of noise that was different from the sample of noise that was presented in each of the remaining three intervals. Samples of noise were selected on a trial-by-trial basis from a set of 50 independent noises for each hourly session. Six durations of noise, ranging from 10 to 328 ms, were utilized, and all were presented at an overall level of 79 dB. The baseline condition for comparisons of performance was a monaural presentation of the noises. As expected, discriminability improved when the baseline condition was presented diotically. The next best performance was obtained when the baseline stimuli was presented to both ears with the proviso that the noises were interaurally uncorrelated. Other dichotic conditions, all of which also utilized interaurally uncorrelated noises, produced performance that was worse than that obtained in the monaural condition. Large degradations in performance were observed when the left ear received a different sample of noise in each of the four intervals of a trial while the right ear continued to receive the baseline stimuli. This finding indicates that listeners were effectively unable to ignore the waveforms presented to one ear while performing a task which could be solved with the information presented to the other ear.(ABSTRACT TRUNCATED AT 250 WORDS)