A framework for Bangla text to speech synthesis
K. M. Azharul Hasan, Muhammad Hozaifa, Sanjoy Dutta, Rafsan Zani Rabbi · 2014
We describe a basic framework and methodology to convert Bangla Text to Speech. Articulated words are automatically produced from Bangla input text by the methodology from the basic pronunciation of the Bangla words. The single tone syllables are considered as the fundamental units for analysis. The methodology selects phonetic units from uttered vocabulary and then combined the appropriate diphones to get the final output. The uttered syllables are selected on the basis of synthesized input text. The input text is analyzed according to the very basic grammar rules of Bangla pronunciation. Using the methodology, we developed a prototype system and the output generated from the system has been analyzed with the natural human voice.