Recognizing and Concatenating of the Tamil Syllables Using FPGA

This project describes the designing of FPGA based Tamil syllable recognition and developing creative applications to enable users to hear Tamil in the language. Around four Thousand Syllables have been identified. This process transforms the text to speech by applying various layers of processing over the text. The layers are Character recognizing, Syllable Formation & identification and creating library for Tamil syllables. The Character recognizing module used for normalization, transliteration of the given Text. The syllable Formation & identification module is used for syllable splitting. In the Library module the Word will be split in to its corresponding syllables, will take a word as input and give out an array of syllables. In this paper a text-to-speech synthesizer which concatenates pre-recorded acoustic elements of the Tamil syllables and forms a continuous speech element has been proposed. The text input is analysed and compared with the stored data and it is synthesized for final output. Hardware implementation of the proposed system is achieved by FPGA (Cyclone II), VHDL and synthesizer tools.