Vol. 1 No. 1 (2021): African Journal of Artificial Intelligence and Sustainable Development
Articles

Text-to-Speech Synthesis - Techniques and Evaluation: Analyzing techniques and evaluation metrics for text-to-speech (TTS) synthesis systems for converting text input into spoken audio output

Sofia Kim
Assistant Professor, AI Research Center, Pacific University, Los Angeles, USA
Cover

Published 30-05-2021

Keywords

  • Text-to-Speech Synthesis,
  • TTS Techniques,
  • TTS Evaluation

How to Cite

[1]
Sofia Kim, “Text-to-Speech Synthesis - Techniques and Evaluation: Analyzing techniques and evaluation metrics for text-to-speech (TTS) synthesis systems for converting text input into spoken audio output”, African J. of Artificial Int. and Sust. Dev., vol. 1, no. 1, pp. 31–36, May 2021, Accessed: Jul. 01, 2024. [Online]. Available: https://africansciencegroup.com/index.php/AJAISD/article/view/29

Abstract

Text-to-Speech (TTS) synthesis plays a crucial role in various applications, including accessibility tools, virtual assistants, and entertainment. This paper provides a comprehensive overview of the techniques and evaluation metrics used in TTS synthesis systems. We discuss various methods such as concatenative synthesis, formant synthesis, and statistical parametric synthesis, highlighting their strengths and weaknesses. Additionally, we delve into the evaluation metrics used to assess the quality of synthesized speech, including subjective evaluations, objective metrics, and listener preference tests. By analyzing these techniques and metrics, this paper aims to provide insights into the advancements and challenges in TTS synthesis, paving the way for future research and development in this field.

Downloads

Download data is not yet available.

References

  1. Tatineni, Sumanth. "Blockchain and Data Science Integration for Secure and Transparent Data Sharing." International Journal of Advanced Research in Engineering and Technology (IJARET) 10.3 (2019): 470-480.