An experimental technique on text normalization and its role in speech synthesis

Abstract

Text Normalization is an important step in Natural Language Processing. It plays an important role in various applications such as text-to-speech systems, speech recognition, email voice activators, for visually impaired persons and also for educational purpose. Text normalization means converting the non-standard words into their standard formats (i.e. in the string format or plain text format). The normalization of text is required for words such as dates, currency symbols, abbreviations, acronyms, and numbers etc. to be pronounced or read by the system. The paper presents an overview of work done so far in the field of text normalization and speech synthesis. The work on text normalization is being done from last few decades. The paper discusses in detail the numerous techniques used by different authors for achieving the accuracy in the text normalization for several different languages all over the world.