Thai Study Text Analysis Text To Speech System

Thailand is located in the south of China, in the heart of Southeast Asia, it is a friendly neighbor of China. Thailand is the key partnerships in economic and politic exsit between Thailand and China.Thai is the official language of Thailand, currently it has more than60million people speak Thai language. Thai is an analysis language and also an isolated language, its basic vocabulary is mostly made up of monosyllables. At the same time, Thai is a tonal language, tone can be used to distinguish between different vocabularies and grammars. It contains a huge potential market in Southeast Asia, languages in Southeast Asian will become the main research in the field.The main work includes as follows:1. Download the words from a professional Thai Online Dictionary, pick out the words such as the common words, compound words, place names, quantifiers, loanwords and so on, and add the standard pronunciation and POS information to the Thai Dictionary.2. Chooses the commonly used sentences from the Thai specialized books as well as the websites, and select the useful language materials from sentences, then removes the sentences which length and form is inappropriate, finally these sentences can be used for text analysis.3. Constructing Thai TTS system on the basis of the Thai dictionary, first of all should do the textual analysis. In view of Thai linguistic feature, this paper designs a Thai word segmentation algorithm combine forward maximum matching algorithm with backward maximum matching algorithm based on the dictionary, then replaces the words with Thai syllable information.4. For Thai dictionary based on forward and backward maximum matching algorithm did not match the words, we designed a spelling rules based on syllable processing method for processing.5. Designs an improvement Thai romanisation code scheme, in foundation that in the above process completes, coding the Thai text with this code scheme, then compared with the standard Thai pronunciation again.In this paper, segmenting Thai text through the forward and backward matching algorithm based on Thai dictionary, have been able to properly cut the words which contained in the dictionary. To deal with the unknown word, we try to show the text into syllables combination, the syllable accuracy rate of about78%; Speech synthesis system requires the Romanization of text to extract syllable pitch information, Improved Thai romanisation code scheme can express the Thai syllable and tone information more accurately.