Abstract
Recent research suggests that some types of emotional speech are more intelligible than others when presented in background noise (Dupuis & Pichora-Fuller, 2015). Yet, emotional speech can often be present as the background noise itself. Attentional mechanisms and auditory stream segregation capabilities likely impact target word recognition in these situations with emotional speech maskers. The present study examined calm target sentences in masking background combinations of four emotions (calm, sad, happy, and angry) and two masker types (2-talker babble and speech-shaped noise). The emotion categories differ from one another on perceptual dimensions of activation (low to high) and pleasantness (unpleasant to pleasant). Speech-shaped noise maskers were spectrally and temporally similar to the 2-talker babble maskers. Performance was compared between the two masker type conditions to quantify the amount of informational masking induced by different emotional speech maskers. Furthermore, the number of target-masker confusions were identified for each masker emotion and subsequently analyzed. Results showed lowest target recognition with calm speech maskers, consistent with auditory stream segregation framework; however, confusions and informational masking were high for angry and happy speech maskers, which suggests that auditory attention to high-arousal stimuli may have influenced performance.