Abstract:
There are not many researches undertaken in the area of digital synthesizing of Sri Lankan personal names. The authors could only find a small number of comprehensive works, which includes Sri Lankan personal name translation as well as synthesis. Even the existing works are generalized and may not cater the specific needs of name synthesis. The objective of this project is to address the problem of synthesizing Sri Lankan personal names as much as close to their natural pronunciation. This paper presents an empirical study on the translation and synthesis of Sri Lankan personal names. The solution is based on natural language processing techniques with translation and transliteration methods. Sinhala names written in English are input to the system. System translates them to Sinhala Unicode with a support of a special rule base. This rule base is composed with mappings between English name segments with Sinhala Unicode name segments. Special data structure Trie is used to hold the mappings. Input follows steps such as text formatting, syntax parsing, stress assignment and smoothing in the translation process. System uses digital synthesizing techniques in Microsoft Speech API to generate synthesis voice. This solution can convert any English written Sri Lankan personal name into spoken words, close to their natural pronunciation
