Development of Grapheme-to-Phoneme Conversion System for Yorùbá Text-to-Speech Synthesis

Main Article Content

Abímbólá R. Ìyàndá
Odétúnjí A. Odéjobí
Festus A. Soyoye
Olúbénga O. Akinadé

Abstract

Grapheme-to-Phoneme (G2P) conversion is concerned with describing the process of automatic conversion of letters in a text into their phonemic transcription and it plays an important role in text to speech synthesis. The G2P component of Yorùbá TTS is yet to be addressed. To achieve the conversion of the grapheme to phoneme, knowledge of the process underlying standard Yorùbá text to sound is expedient. This paper presents a system for generating the phonemic description of the sound corresponding to a piece of Yorùbá text. Standard Yorùbá text data was collected and digitised and resulting digital text was edited for correction of orthographic items using Tákádá text editor and Àkotó Yorùbá software. The corresponding speech data for the text collected was recorded in a quiet environment with a noise cancelling microphone on a typical multimedia computer system using the Speech Filing System software and analysed and annotated using PRAAT speech processing software. The system was designed with Finite State Transducer and implemented using Python 2.7 programming language. The system was able to generate multiple correct representations for graphemes in Yorùbá language with 100\% accuracy on both grapheme level and word level. The results obtained in this study confirmed the hypothesis that the Yorùbá G2P system has a systematic procedure underlying it and this procedure can be computationally specified, analysed and represented using computational tools.

Article Details

How to Cite
Ìyàndá, A. R., Odéjobí, O. A., Soyoye, F. A., & Akinadé, O. O. (2014). Development of Grapheme-to-Phoneme Conversion System for Yorùbá Text-to-Speech Synthesis. INFOCOMP Journal of Computer Science, 13(2), 44-53. Retrieved from http://infocomp.dcc.ufla.br/index.php/infocomp/article/view/394
Section
Articles