Relative importance of temporal envelope and fine structure in lexical-tone perception

The relative importance of temporal envelope and fine structure in speech and music perception was investigated by Smith et al. [Nature (London) 416, 87-90 (2002)] using "auditory chimera" in which the envelope from one sound was paired with the fine structure of another. Smith et al. found that, when 4 to 16 frequency bands were used, recognition of English speech was dominated by the envelope, whereas recognition of melody was dominated by the fine structure. In the present study, Mandarin Chinese monosyllables were divided into 4, 8, or 16 frequency bands and the tine structure and envelope of one tone pattern were exchanged with those of another tone pattern of the same monosyllable. Five normal-hearing native Mandarin Chinese speakers completed a four-alternative forced-choice tone-identification task. In the vast majority of trials, subjects based their identification of the monosyllables on the fine structure rather than the envelope. Thus, the relative importance of envelope and fine structure for lexical-tone perception resembled that for melody recognition rather than that for English speech recognition. Delivering fine-structure information in cochlear implant stimulation could be particularly beneficial for lexical-tone perception. (C) 2003 Acoustical Society of America.