心理科学
心理科學
심이과학
Psychological Science
2013年
2期
306~310
,共null页
语音频谱 艾森克人格问卷说谎量表 说谎
語音頻譜 艾森剋人格問捲說謊量錶 說謊
어음빈보 애삼극인격문권설황량표 설황
speech spectrum, lie scale of Eysenck personality questionnaire, lying
本研究探索艾森克人格问卷说谎量表上说谎与不说谎的语音频谱差别。我们录下被试在做艾森克人格问卷时回答“是”或“不是”的声音。使用Praat语音分析软件对／bu／、／sh／、／i／的前三个共振峰频率和波段（F1、F2、F3、B1、B2、B3）进行分析。结果发现：说“是”时说谎和说真话在／sh／的F1、F3、B1上存在差异，说“不是”时在／sh／的F1、F2、F3上存在差异。本研究表明说谎和说真话在语音频谱上存在差异，同时说谎在不同音素上的效应可能不一样。
本研究探索艾森剋人格問捲說謊量錶上說謊與不說謊的語音頻譜差彆。我們錄下被試在做艾森剋人格問捲時迴答“是”或“不是”的聲音。使用Praat語音分析軟件對／bu／、／sh／、／i／的前三箇共振峰頻率和波段（F1、F2、F3、B1、B2、B3）進行分析。結果髮現：說“是”時說謊和說真話在／sh／的F1、F3、B1上存在差異，說“不是”時在／sh／的F1、F2、F3上存在差異。本研究錶明說謊和說真話在語音頻譜上存在差異，同時說謊在不同音素上的效應可能不一樣。
본연구탐색애삼극인격문권설황량표상설황여불설황적어음빈보차별。아문록하피시재주애삼극인격문권시회답“시”혹“불시”적성음。사용Praat어음분석연건대／bu／、／sh／、／i／적전삼개공진봉빈솔화파단（F1、F2、F3、B1、B2、B3）진행분석。결과발현：설“시”시설황화설진화재／sh／적F1、F3、B1상존재차이，설“불시”시재／sh／적F1、F2、F3상존재차이。본연구표명설황화설진화재어음빈보상존재차이，동시설황재불동음소상적효응가능불일양。
This research aimed at exploring the differences of speech spectrum of different phonemes （ vowels or consonants） between telling lies and telling truths in the lie scale of Eysenck personality questionnaire（ L scale）. Because the voice can reflect a man＇s emo- tional and physical state and telling lies may evoke physical arousal, we hypothesized that the speech spectrum during lying and truth - telling would be different. Forty-six subjects completed the revised Eysenck Personality Questionnaire Short Scale for Chinese edited by Zhonggeng Chen, which has been demonstrated to be reliable and valid for Chinese participants. The inventory contains four personality trait scales： Psy- choticism （P）, Extraversion （E）, Neuroticism （N）, and Social Desirability （L）. The L scale contains questions on which individuals tend to lie for social desirability, and the scoring standard is the same for everyone. If a participant received a score of one instead of ze- ro on an L scale question, then this response was recorded as a lie. They were instructed to verbally answer the questions shown on a computer screen by E-Prime 2.0 with ＂yes＂ or ＂no＂ （/shi/or/bushi/） in Chinese and their responses were recorded by the computer. Praat 5.1.30 software in a Windows 2003 operating system was used to analyze participants＇recorded voice samples. Measured speech spectrum parameters included frequency of the former three formants （ F1, F2, F3 ）, and bandwidth of the first three formants （ B1, B2, B3）. We analyzed the speech spectrum parameters of both the consonant/sh/and the vowel/i/in the present study because it has been previously demonstrated that the position and motion of oral cavity, larynx and tongue are different between consonants and vowels when they are pronounced. The responses of ＂yes＂ and ＂no＂ were also analyzed separately in order to control the influences of/bu/to/sh/ and/i/. Consistent with our hypothesis, the paired sample t -test showed that when participants responded with ＂yes＂ , the effects of lying were significant or marginally significant on F1, F3, and B1 of the consonant/sh/ （p =. 074 ,p =. 032 ,p =. 069, respectively） ; when participants responded with ＂no＂, the effects of lying were significant on F1, F2, and F3 of the consonant/sh/ （p =. 037, p =. 020,p =. 015, respectively） ; There was not any significant effect of lying on the speech spectrum of/bu/ and the vowel/i/ in whatever kind of responses （all ps 〉 . 10）. The results of this research indicated that it was effective to distinguish telling lies from telling truths through speech spectrum analysis. In addition, the effects of lying on the speech spectrums of different phonemes may be varied.
