• DocumentCode
    2019671
  • Title

    Classification of the pronunciations of the /v/ consonant in the context _/v/V in the Romanian language

  • Author

    Untu, Alina ; Teodorescu, Horia Nicolai

  • Author_Institution
    Telecommun. & Inf. Technol., Gheorghe Asachi Tech. Univ. of Iasi, Iasi, Romania
  • fYear
    2011
  • fDate
    18-21 May 2011
  • Firstpage
    1
  • Lastpage
    9
  • Abstract
    We analyze a set of statistical features of the pronunciations of the /v/ consonant in the Romanian language, in words starting with /v/ followed by a vowel. The analysis reveals several types of pronunciations of /v/. Increasing the statistic from a previous work from 19 to 29 speakers we conclude that beyond the voiced / non-voiced/ fluctuating (voiced - unvoiced) classes, the /v/ in the said contexts fall into distinct subclasses. This classification shows that, while at the perceptive level the Romanian speakers do not distinguish different pronunciations, at the acoustical level the /v/ in the Romanian spoken language is almost as rich as in other languages as Spanish and German.
  • Keywords
    speech processing; statistical analysis; /v/ consonant; /v/V context; Romanian spoken language; pronunciation classification; statistical feature analysis; Arrays; Clustering algorithms; Couplings; Feature extraction; Indexes; Prototypes; Support vector machine classification; Euclidian distance; cluster validity; formants; hierarchical clustering; k-means; prototypes;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Speech Technology and Human-Computer Dialogue (SpeD), 2011 6th Conference on
  • Conference_Location
    Brasov
  • Print_ISBN
    978-1-4577-0440-6
  • Type

    conf

  • DOI
    10.1109/SPED.2011.5940730
  • Filename
    5940730