Klatt D H
J Speech Hear Res. 1975 Dec;18(4):686-706. doi: 10.1044/jshr.1804.686.
The voice onset time (VOT) and the duration of the burst of frication noise at the release of a plosive consonant were measured from spectrograms of word-initial consonant clusters. Mean data from three speakers reading English words in a sentence frame indicated that the VOT changed as a function of the place of articulation of the plosive and as a function of the identity of the following vowel or sonorant consonant. Burst durations varied in a similar way such that the remaining interval of aspiration in /p, t, k/ was nearly the same duration in comparable phonetic environments. The VOT was longer before sonorants and high vowels than before mid- and low vowels. Aspiration was also seen in an /s/-sonorant cluster. To explain these regularities, production strategies and perceptual cues to a voicing decision for English plosives are considered. Variations in VOT are explained in terms of articulatory mechanisms, perceptual constraints, and phonological rules. Some VOT data obtained from a connected discourse were also analyzed and organized into a set of rules for predicting voice onset time in any sentence context.
从词首辅音丛的语图中测量了爆破辅音释放时的嗓音起始时间（VOT）和摩擦噪音爆发的时长。三位讲英语者在句子框架中朗读英语单词的平均数据表明，VOT会随着爆破音的发音部位以及后续元音或响音辅音的特性而变化。爆发时长的变化方式类似，以至于在可比的语音环境中，/p、t、k/的剩余送气时长几乎相同。VOT在响音和高元音之前比在中元音和低元音之前更长。在/s/ - 响音丛中也能看到送气现象。为了解释这些规律，我们考虑了英语爆破音发声决策的发音策略和感知线索。VOT的变化是根据发音机制、感知限制和音系规则来解释的。还分析了从连贯话语中获得的一些VOT数据，并将其整理成一组规则，用于预测任何句子语境中的嗓音起始时间。
