Isolating the informational component of speech-on-speech masking

Abstract

Speech perception in the presence of a competing talker involves both energetic (EM) and informational masking (IM). This study aimed to isolate the informational component of speech-on-speech masking. EM can be eliminated by presenting the target and masker to opposite ears (i.e., dichotically). However, this also dramatically reduces the effects of IM by providing listeners with lateralization cues. Previous research using tonal sequences has shown that IM can be isolated by presenting the target and masker dichotically while rapidly switching the two streams across the ears. The question remains whether this technique can also be used for speech materials. Speech reception thresholds (SRTs) were measured for sentences produced by a female talker in the presence of a competing male talker under three conditions: diotic, dichotic, and switching. In the switching condition, target and masker were presented dichotically, but their lateralization was switched after every word in the target sentence. Preliminary data suggests that SRTs for the switching condition are higher than for the dichotic condition but lower than for the diotic condition. This suggests that, contrary to findings for tonal sequences, rapidly switching the target and masker speech across the ears does not fully reintroduce IM. [Work supported by NIH.]