Revisiting binaural release from energetic and informational masking

Binaural release from energetic versus informational masking is typicallydescribed as relying on different cues. Whereas speech-in-noise recognition is thought to benefit from low-level binaural difference cues, speech-in-speech recognition is said to improve when differences in perceived location facilitate selective attention to the target. This suggests that binaural masking release might be maximal under different conditions for maskers that are dominated by energetic versus informational masking. Data were collected from normal-hearing adults to test this expectation. Sentence recognition was evaluated in speech-shaped noise and in two-talker speech, with binaural cues created by manipulating interaural phase (0- or π-phase) or interaural time (0 or 660 μs). The results replicate previous studies showing that a robust masking release is achieved for a variety of interaural configurations. The magnitude of the release for speech-in-speech recognition does not appear to be maximal for configurations producing well-localized sound images, however. This result raises questions about the fundamental differences between binaural release from energetic and informational masking.