Scholarlink

<p>For realistic speech generation, variation in glottal waveform models has long been proposed. Due to simplicity and efficiency, the parametric models of the glottal flow are very popular in the field of speech generation. The proposed work presents a new approach to modeling the glottal flow. The current model is comprised of two piecewise differential equations that generate a glottal pulse. The first and second differential equations generate the opening and closing phases of the vocal folds, respectively while the closed phase is taken as zero. There are four parameters involved in the proposed model to bring variation in the shape of the glottal pulse. The current model is very flexible in designing a glottal pulse and is comparable with the famous Liljencrants-Fant model, Rosenberg model, and KLGLOTT88 model. This comparison supports its successful implementation as a voice source in speech synthesis which also leads to the validity of our differential equation-based glottal model.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Parametric Modeling of Glottal Flow in Form of Piecewise Differential Equations

  • 期刊论文

  • 发表日期：

    2025年1月23日

查看全文
  • Tahir Mushtaq,
  • Sadaf Murtaza,
  • Muhammad Saleem,
  • Khurram Farhad,
  • Muhammad Faisal Ejaz

摘要

For realistic speech generation, variation in glottal waveform models has long been proposed. Due to simplicity and efficiency, the parametric models of the glottal flow are very popular in the field of speech generation. The proposed work presents a new approach to modeling the glottal flow. The current model is comprised of two piecewise differential equations that generate a glottal pulse. The first and second differential equations generate the opening and closing phases of the vocal folds, respectively while the closed phase is taken as zero. There are four parameters involved in the proposed model to bring variation in the shape of the glottal pulse. The current model is very flexible in designing a glottal pulse and is comparable with the famous Liljencrants-Fant model, Rosenberg model, and KLGLOTT88 model. This comparison supports its successful implementation as a voice source in speech synthesis which also leads to the validity of our differential equation-based glottal model.