PARAMETRIC SPEECH SYNTHESIS BY SOURCE-FILTER MODEL

  • P. SHERLY ARUNODHAYAMARY, S. SELVA NIDHYANANTHAN

Abstract

Speech synthesis is a notable research field in speech processing which provides smart phone applications and also supports blind people. Process of obtaining natural speech from gadgets are gaining lots of attention due to their popularity in the market. But the major drawback is the voice quality. In this work, levels of synthesis are performed and their performance at various levels is quoted to predict the better performance among the levels of synthesis. The digits and vowels are data considered for work. The first part of synthesis is the process of generating voice from text in case 1. Recorded database is used in case 2. A level is increased with a module of source-filter synthesis. The performance of the module is evaluated and compared between the present results from case 1, case 2. To evaluate performance of the levels, measure have been taken by Mean Square Error, Peak Signal-to-Noise Ratio  

Published
2020-04-18
How to Cite
P. SHERLY ARUNODHAYAMARY, S. SELVA NIDHYANANTHAN. (2020). PARAMETRIC SPEECH SYNTHESIS BY SOURCE-FILTER MODEL. International Journal of Advanced Science and Technology, 29(7s), 1077 - 1084. Retrieved from https://sersc.org/journals/index.php/IJAST/article/view/10624