Perceptual segregation of competing speech sounds: the role of spatial location

Culling and Summerfield [J. Acoust Soc. Am. 92, 785-797 (1995)] showed that listeners could not use ongoing interaural time differences (ITDs) to achieve source segregation. The present experiments tested a free-field analog of their experiment. The stimuli consisted of narrow bands of noise, pairs of which represented the first and second formants of the whispered vowels "ar," "ee," "er," and "oo." A target noise-band pair (vowel) was presented at various angles on the listeners' left while a complementary distracter was presented on the listeners' right. Listeners correctly identified the target vowel in the free-field well above chance. Performance remained well above chance in headphone experiments that retained spatial cues but eliminated reverberations and head movements. The full range of cues that normally determine perceived spatial location provided sufficient information for segregation. Further experiments, which systematically evaluated the contribution of these cues in isolation and in combination, showed that some listeners, following training, exhibited the ability to segregate based on ongoing ITDs alone. Substantial individual differences were observed. The results show that listeners can use spatial cues to segregate simultaneous sound sources.