Nittrouer S, Whalen D H
Haskins Laboratories, New Haven, Connecticut 06511.
J Acoust Soc Am. 1989 Oct;86(4):1266-76. doi: 10.1121/1.398741.
Earlier work [Nittrouer et al., J. Speech Hear. Res. 32, 120-132 (1989)] demonstrated greater evidence of coarticulation in the fricative-vowel syllables of children than in those of adults when measured by anticipatory vowel effects on the resonant frequency of the fricative back cavity. In the present study, three experiments showed that this increased coarticulation led to improved vowel recognition from the fricative noise alone: Vowel identification by adult listeners was better overall for children's productions and was successful earlier in the fricative noise. This enhanced vowel recognition for children's samples was obtained in spite of the fact that children's and adults' samples were randomized together, therefore indicating that listeners were able to normalize the vowel information within a fricative noise where there often was acoustic evidence of only one formant associated primarily with the vowel. Correct vowel judgments were found to be largely independent of fricative identification. However, when another coarticulatory effect, the lowering of the main spectral prominence of the fricative noise for /u/ versus /i/, was taken into account, vowel judgments were found to interact with fricative identification. The results show that listeners are sensitive to the greater coarticulation in children's fricative-vowel syllables, and that, in some circumstances, they do not need to make a correct identification of the most prominently specified phone in order to make a correct identification of a coarticulated one.
早期的研究工作[尼特鲁尔等人,《言语与听觉研究杂志》32卷,120 - 132页(1989年)]表明,当通过前元音对擦音后腔共振频率的影响来衡量时,儿童擦音 - 元音音节中的协同发音证据比成人的更多。在本研究中,三个实验表明,这种增加的协同发音导致仅从擦音噪声中就能提高元音识别率:成年听众对儿童发音的元音识别总体上更好,并且在擦音噪声中更早就能成功识别。尽管儿童和成人的样本是随机混合的,但仍获得了儿童样本的这种增强的元音识别能力,因此表明听众能够在擦音噪声中对元音信息进行归一化处理。在这种擦音噪声中,通常只有一个主要与元音相关的共振峰的声学证据。发现正确的元音判断在很大程度上与擦音识别无关。然而,当考虑到另一种协同发音效应,即/u/与/i/相比擦音噪声的主要频谱峰值降低时,发现元音判断与擦音识别相互作用。结果表明,听众对儿童擦音 - 元音音节中更大的协同发音很敏感,并且在某些情况下,他们不需要正确识别最突出指定的音素就能正确识别协同发音的音素。