The study of speech and the breath considering the learning stage of pronunciation

考虑发音学习阶段的言语和呼吸的研究

• 批准号: 04680294
• 负责人: AOYAMA Kensei
• 金额: $ 1.15万
• 依托单位: Tsukuba College of Technology
• 依托单位国家: 日本
• 项目类别: Grant-in-Aid for General Scientific Research (C)
• 财政年份: 1992
• 资助国家: 日本
• 起止时间: 1992 至 1993
• 项目状态: 已结题
• 来源: https://kaken.nii.ac.jp/en/grant/KAKENHI-PROJECT-04680294/
• 关键词: pronunciation; Learning stage; accoustic parameter; velocity of breath; 修得度

This research studies acoustic parameters of speech and the velocity wave form of breath considering the learning stage of pronunciation. Three normal and three hearing impaired subjects pronounce explosive VCV syllables. The VCV syllable speech is snalyzed, and acoustic parameters are extracted. The parameters are pitch, the maximum partial auto-correlation coefficient(PARCOR), the amplitude of the sound source for linear predictive analysis, speech power, cepstrum distance, sound spectrogram, linear prediction cording(LPC)spectrum, and the line spectrum pair(LSP)coefficient. First and second and third format locus are esimated from sound spectrogram and LPC spectrum. Its transition from the explosive consonant to the following vowel shows whether pronunciation of explosives is correct or not. And the learnig stage of vowel is found in the first and second format. Then the learning stage of prosody and voiced(unvoiced)sound is found in those acoustic parameters. Next breath discharged with VCV syllable speech is studied for hearing impaired subjects considering their learning stage of pronunciation. The learning stage of explosives is found in the maximum velocity of breath(U_<pm2>), the direction(rheta_<m2>)where U_<pm2> is measured, and the duration between the beginning of the following vowel and the portion of U_<pm2> at consonant-vowel transition. And then the learning stage of the breath control is found.

本研究从语音学习阶段出发，研究了语音的声学参数和呼吸的速度波形。三名正常人和三名听力受损的受试者发出爆炸性VCV音节。对VCV音节语音进行分析，提取声学参数。这些参数是音调、最大偏自相关系数（PARCOR）、用于线性预测分析的声源的幅度、语音功率、倒谱距离、声谱图、线性预测编码（LPC）谱和线性谱对（LSP）系数。从声谱图和LPC谱中估计出第一、第二和第三格式轨迹。爆破音由爆破辅音向其后元音的过渡，是爆破音发音正确与否的标志。第一、二格式中存在元音的学习阶段。然后在这些声学参数中找到韵律和清（浊）音的学习阶段。针对听力障碍者的语音学习阶段，研究了以VCV方式发出的下一次呼吸音节语音。爆破音的学习阶段主要体现在最大呼吸速度（U_<pm2>）、测量U_的方向（rheta_<m2>）<pm2>以及从下一个元音开始到U_的子元音过渡段之间的时间<pm2>。然后找到呼吸控制的学习阶段。