Multi-modality Database Of Speech Production

This paper discusses how to build a Multi-modality database.of the proverb.Modern physiological techniques explore the issue of building a Multi-modality speech production database for languages.Chapter 1 is the introduction.It introduces the relevant situation of the Awa dialect,summarizes the domestic and foreign research of Multi-modality database,explains the definition of Multi-modality related terms,and introduces the definition.Chapter 2 proposes the construction of multimodal database of ethnic languages.The idea of using five modes to combine the simultaneous acquisition of speech physiological signal data is to collect speech and arpeggio data;to collect voice and dynamic electronic clamp data;Acoustic and aerodynamic data acquisition;voice arpeggio,aerodynamics,tongue image data acquisition,voice voice,dynamic electronic position,tongue image,aerodynamic data collection,and database speaker information For the five different experiments,the corresponding corpus was designed,and the experimental equipment was briefly introduced to sort out the experimental process.Chapter 3 explains the five configurations one by one,shows the experimental process of synchronous acquisition of speech physiological signals,records the slang corpus,collects the dynamic data of slang speech,and carries out data processing work,which is often encountered in speech experiments.The problem is given the corresponding solution,and the proverbial Multi-modality database is constructed.The fourth chapter analyzes the acoustic parameters such as formant,fundamental frequency and duration with the elastic vowel of Buolke dialect in the proverb.It concludes that the tight vowel tongue is slightly lower than the high vowel.An aerodynamic study of the nasal consonants of the Awa dialect was also made.It was found that the nasal consonants of the Awa dialect are complex consonants.Chapter 5 summarizes the full text and reflects on some of the shortcomings of this article.