Analysis of Word-Based and Unit-Based Diphone Concatenation for Myanmar Text-To-Speech

Abstract

Speech synthesis system is a popular field in natural language processing of computer science for various languages. The process of speech synthesis is to produce the human-like speech from corresponding language and it can be divided into two key phases such as high-level and low-level text-to-speech synthesis. In high-level synthesis, the input text is converted into such form that the low-level synthesizer can produce the output speech. Speech synthesis can be used three types: Articulatory, Format and Concatenative synthesis. Concatenative speech synthesis is the most popular technique among three types and it is the easiest technique to synthesis the speech. Diphone-concatenative method is one of the popular methods in nowadays because of speech quality is better than other techniques and complexity is less than others. This paper gives the analysis of two type of Diphone-concatenation; word-based and unit-based synthesis for Myanmar language by applying Pitch Synchronous Overlap and Add (PSOLA) algorithm to smooth the joints of the speech signals. This paper describes the Time-Domain Pitch Synchronous Overlap and Add method in Diphoneconcatenation speech synthesis and it is to maintain the consistency and accuracy of the pitch marks of the speech signal and Diphones database with integrated vowels and consonants of Myanmar language. This paper also describes the building of Myanmar Diphone database for concatenation method. Diphone Database for Myanmar pronunciations is constructed in this paper to reduce the ambiguity in pronunciations.