International Journal of Computer Applications |
Foundation of Computer Science (FCS), NY, USA |
Volume 49 - Number 13 |
Year of Publication: 2012 |
Authors: Simmi Luthra, Parminder Singh |
10.5120/7690-1006 |
Simmi Luthra, Parminder Singh . Punjabi Speech Generation System based on Phonemes. International Journal of Computer Applications. 49, 13 ( July 2012), 40-44. DOI=10.5120/7690-1006
This paper discusses the development of a concatenated speech synthesis system for the Punjabi text written in Gurmukhi script based on Phoneme as basic unit. The research work was carried out with the aim that the developed system must be able to produce synthetic speech, corresponding to two qualities of Speech Synthesis: intelligibility and naturalness. In order to get the naturalness in the synthetic speech, Concatenative speech synthesis techniques have been used, with Punjabi Phonemes as a basic unit. In this work, phonemes have been selected as the speech unit for the development of the Punjabi speech database. After analyzing of Phonemes, a carefully selection of unbiased Punjabi Corpus was made, having nearly,2,33,009 unique and more than four million total words. The two types of Phonemes (V and CV) gave rise to 380 phonemes with non nasalized vowels and 380 phonemes with nasalized vowels, resulting total 722 phonemes. The system is based on a Punjabi speech database that contains the starting and ending positions of phoneme-sounds labeled carefully in a wave file of recorded words. The input text is first processed then these phonemes are searched in the database for corresponding phoneme-sound positions in recorded wave file.