The binaural advantage in reverberation for stimuli with controlled predictability

Abstract

Monaural near‐ear and binaural speech‐recognition performance was assessed under two levels of reverberation (T = 0.0 and 0.8 s), at two signal‐to‐babble ratios (+4 and −2 dB), and at two speaker‐to‐listener orientations (speech at 0° and noise at 180°, and speech at 45° and noise at 225°). Speech‐recognition performance was assessed using the high (PH) and low (PL) predictability sentences from the revised‐SPIN test (Bilger, 1984). The PH and PL items from the SPIN were recorded through a KEMAR manikin under the above conditions and played back to normal‐hearing subjects via earphones. A binaural advantage was observed across all listening conditions for both the PH and PL items. The average binaural advantage for the PH items at T = 0.0 s was about two times greater than at T = 0.8 s. The binaural advantage was approximately the same for the PL items regardless of reverberation time. Significant intersubject variability was observed for the binaural advantage across listening conditions for both types of speech stimuli. In addition, there were several interaction effects between the various acoustic parameters examined for the PH and PL items.