宋词呼吸韵律结构模型研究

Research on the Structure Model of Respiratory and Rhythm for Song-poems

本文选择宋词为实验语料,使用PowerLab生物电采集器、MLT1132压电呼吸带传感器等仪器录制呼吸信号和语音信号,对语音信号的韵律等级进行切分,通过判断呼吸周期的波峰和波谷来标记吸气相和呼气相,揭示了不同语音韵律等级时长和两个常见的呼吸结构模型之间的相关性。在语音多模态研究中,可以为语音多模态的合成提供可靠的数据,提高言语模型的自然度。准确的数据模本还可以为治疗言语障碍提供帮助,为语言教学和二语习得提供标准。

Objective： To design a forensic automation platform for DNA samples preprocessing and test the extraction effect. Methods： In this paper, Song-poems was chosen as record corpus, PowerLab was used as bioelectricity collector, and respiratory signal and speech signal were recorded with MLT1132 piezo respiratory belt transducer. The inspiratory phase and expiratory phase were marked through distinguishing rhythm rank and respiratory cycle, which revealed the relationship between time of different rhythm rank and the two common respiratory structure models. Our research provided reliable data for synthesis of speech multimodal, improving the naturalness of speech model, which could be helpful for the treatment of speech disorder and provide a standard for the language teaching and second language acquisition.