Text To Speech Engine For Gujarati Language

Abstract

Text-to-Speech system is the process of converting a text input into a speech output. TTS newlinesystems can be used in various applications such as public announcing systems, automatic newlinemessaging systems, email readers, voice recognition systems, newspaper readers for the newlinevisually impaired and computer aided learning systems. This research work presents the newlinedevelopment of a text-to-speech (TTS) engine tailored for the Gujarati language, a widely newlinespoken language in the Indian state of Gujarat and among the Gujarati diaspora worldwide. newlineDespite the major developments in TTS technology for many of the world s languages, such newlineas English and Mandarin, many others, including Gujarati have remained underrepresented in newlinethe digital realm. This research s central goal is to lay the groundwork for a high-quality, newlinenatural-sounding TTS system that precisely converts Gujarat-written into spoken words. newlineThe research work encompasses several key components: Text Analysis, Text Pre-Processing, newlineText Normalization, and speech synthesis. The text Pre-Processing module deals with the newlinenormalization of Gujarati text, handling non-standard words and complex script features. The newlinefinal speech output is synthesized by a concatenative synthesis technique to generate the bestquality audio output, utilizing a database of pre-recorded Gujarati words. This method is newlinechosen due to the excellent balance between naturalness and computational efficiency newline