首页 | 本学科首页   官方微博 | 高级检索  
     

汉语声调在言语可懂度中的重要作用
引用本文:张家騄, 齐士钤, 宋美珍, 刘全祥. 汉语声调在言语可懂度中的重要作用[J]. 声学学报, 1981, 6(4): 237-241. DOI: 10.15949/j.cnki.0371-0025.1981.04.004
作者姓名:张家騄  齐士钤  宋美珍  刘全祥
作者单位:中国科学院声学研究所(张家騄,齐士钤,宋美珍),中国科学院声学研究所(刘全祥)
摘    要:在四种失真(高、低通滤波并混有非线性失真)条件下,进行言语清晰度试验,证明了失真条件下,声调具有很强的抗干扰能力,声调清晰度几乎不受影响。
采用以下不同的激励源(1.白噪声、2.基频固定的三角波、3.基频随声带基频变化的三角波、4.条件1.加条件3.)产生的合成言语,进行言语清晰度试验,证明:声调信息大大有助于提高言语可懂度,声调信息主要存在于基频随时间的变化中,强度变化对声调信息有补偿作用,以及清辅音的存在与否对声调清晰度是有一定影响的。

收稿时间:1980-02-29

ON THE IMPORTANT ROLE OF CHINESE TONES IN SPEECH INTELLIGIBILITY
ZHANG Jia-lu, QI Shi-qian, SONG Mei-zhen, LIU Quan-xiang. ON THE IMPORTANT ROLE OF CHINESE TONES IN SPEECH INTELLIGIBILITY[J]. ACTA ACUSTICA, 1981, 6(4): 237-241. DOI: 10.15949/j.cnki.0371-0025.1981.04.004
Authors:ZHANG Jia-lu  QI Shi-qian  SONG Mei-zhen  LIU Quan-xiang
Abstract:Chinese is a tone language. The role of Chinese tones played in the speech perception were investigeted by speech articulation tests. The first part of these tests was carried out under four conditions: I. Speech via low-pass filter, cutoff frequency at 1600 Hz, S/N = 10 dB, II. speech via high-pass filter, cutoff frequency at 1600 Hz, S/N=20 dB, III. speech via low-pass filter with some nonlinear distortions, cutoff frequency at 750 Hz, S/N=25 dB, and IV. speech via high-pass filter with some nonlinear distortions, cutoff frequency at 2500 Hz, S/N = 25dB. It is shown that the tone articulation essentially does not change when the syllable articulation is changed from 77% to 28%, but at the sametime the initial articulation is changed from 87.2% to 46.7% and the final articulation from 89% to 39.5%. So the tones possess strong anti-interference ability.
In order to separate the tones from the speech sounds, a 2400 bits/s digital channel vocoder was used in the second part of the articulation tests under four different conditions of excitation source: 1. the excitation source was only a noise source, 2. the excitation source was only a triangular pulse train source with constant fundamental frequency F0=127 Hz, 3. the excitation source was only a triangular pulse train source with fundamental frequency controlled by the voice, and 4. both noise source and pulse source were used and the fundamental frequency controlled by the voice. It is shown that I. the rank order of the tone articulation for the four testing conditions is 4, 3, 1, 2; II. the fundamental frequency is the most important cue for tones; III. the intensity variation in noise spectrum (condition 1) with the variation of fundamental frequency of the speech input may provide some acoustical cues for tone perception; IV. the unvoiced sounds may contribute somewhat to the tone perception (the tone articulation in condition 4 is higher than condition 3).
Keywords:
本文献已被 CNKI 等数据库收录!
点击此处可从《声学学报》浏览原始摘要信息
点击此处可从《声学学报》下载全文
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号