Howell P, Rosen S
J Acoust Soc Am. 1983 Mar;73(3):976-84. doi: 10.1121/1.389023.
Rise times of voiceless affricates and fricatives were measured when the test material occurred in sentences, in isolated words, and in isolated nonsense syllables. As has been reported previously the rise times of affricates were significantly shorter than those of fricatives. Rise times varied with the type of test material and for all types of material were significantly longer than those reported by L.J. Gerstman [unpublished doctoral dissertation, New York University (1957)]. Differences in the way the previous measurements were made do not appear to explain the discrepancy. Because rise time varies with type of test material, no auditory sensitivity at a single rise time value can be responsible for the perceptual distinction between voiceless affricates and fricatives. This conclusion was also supported by the results of a perceptual experiment in which listeners were asked to label speech or nonspeech stimuli drawn from a continuum varying in rise time. Substantially different estimates of the category boundary were obtained when the range of rise times covered by the stimuli differed.
当测试材料出现在句子、孤立单词和孤立无意义音节中时，对清塞擦音和擦音的上升时间进行了测量。如先前报道，塞擦音的上升时间明显短于擦音。上升时间随测试材料类型而变化，并且对于所有类型的材料，其上升时间均明显长于L.J. 格斯特曼[纽约大学未发表的博士论文（1957年）]所报道的时间。先前测量方式的差异似乎无法解释这种差异。由于上升时间随测试材料类型而变化，单一上升时间值下的听觉敏感度不可能是清塞擦音和擦音在感知上产生区别的原因。这一结论也得到了一项感知实验结果的支持，在该实验中，要求听众对从上升时间连续变化的范围内抽取的语音或非语音刺激进行分类标注。当刺激所涵盖的上升时间范围不同时，获得了对类别边界的显著不同估计。
