Picheny M A, Durlach N I, Braida L D
Acoustic Processing, Thomas J. Watson Research Center, IBM, Yorktown Heights, NY 10598.
J Speech Hear Res. 1989 Sep;32(3):600-3.
Previous studies (Picheny, Durlach, & Braida, 1985, 1986) have demonstrated that substantial intelligibility differences exist for hearing-impaired listeners for speech spoken clearly compared to speech spoken conversationally. This paper presents the results of a probe experiment intended to determine the contribution of speaking rate to the intelligibility differences. Clear sentences were processed to have the durational properties of conversational speech, and conversational sentences were processed to have the durational properties of clear speech. Intelligibility testing with hearing-impaired listeners revealed both sets of materials to be degraded after processing. However, the degradation could not be attributable to processing artifacts because reprocessing the materials to restore their original durations produced intelligibility scores close to those observed for the unprocessed materials. We conclude that the simple processing to alter the relative durations of the speech materials was not adequate to assess the contribution of speaking rate to the intelligibility differences; further studies are proposed to address this question.
先前的研究（皮切尼、迪拉赫和布拉伊达，1985年、1986年）表明，与对话式言语相比，听力受损的听众对于清晰说出的言语存在显著的可懂度差异。本文呈现了一项探索性实验的结果，旨在确定语速对可懂度差异的影响。清晰的句子经过处理后具有对话式言语的时长特性，而对话式句子经过处理后具有清晰言语的时长特性。对听力受损的听众进行的可懂度测试显示，两组材料在处理后都有所退化。然而，这种退化不能归因于处理伪像，因为对材料进行重新处理以恢复其原始时长后，得到的可懂度分数与未处理材料的分数相近。我们得出结论，简单地处理以改变言语材料的相对时长不足以评估语速对可懂度差异的影响；建议进一步开展研究以解决这个问题。
