International Journal on Recent and Innovation Trends in Computing and Communication ISSN: 2321-8169 Volume: 2 Issue: 11 3734 – 3736 _______________________________________________________________________________________________ 3734 IJRITCC | November 2014, Available @ http://www.ijritcc.org _______________________________________________________________________________________ A Review on Multilingual Text to Speech Synthesis by Syllabifying the Words of Devanagari and Roman Mr. S. B. Chaudhari Research Scholar Computer Engineering JJT University, Rajasthan. sbchaudhari@yahoo.com Abstract: Speech synthesis is process of spoken language as an input text and converted into speech waveforms. This paper describes the text to speech system for Devanagari scripted language and Roman Language. There are many earliest TTS systems are available but for Devanagari and Roman scripts are not available. Keywords: Unicode, Scripting Language, Optimization and Validation. __________________________________________________*****_________________________________________________ Introduction: General concept of Text to Speech synthesis technology is that it takes the text as input and it converts it into speech. In everyday life for communication we use speech. There are many people in India that they can’t read but they can hear the sound. For them text to speech system is very useful, Such TTS system is very useful for learning new languages. Now days there are many applications in which we require such type of tool. Presently there are many systems are available in the market, some of them are free to use and some of them are not free to use. These systems are having their pros and cons. Majority of system is available for single language like English or Hindi or other language. Basically the Multilingual Text to speech synthesis by syllabifying the words of Devanagari is a proposed system which will take input as typed text of Marathi, English, Hindi, Sanskrit and all languages that use Devanagari script and that text will be converted to speech as output. Text processing and generation of speech are the main steps involved in this system. Designing such type of system it is necessary to consider the input and the expected output. There are many TTS systems are available in the market that reads the text of a particular language. Here it is necessary to design a system that will read Indian languages and English. In globalization the English is the language of the world, Hindi is the official language of India and Marathi is the language of Maharashtra state. Now a day’s people use combination of different languages like Marathi, Hindi and English .To type the text of these languages in most of the available text to speech system different local fonts are used. In the proposed system we are going to use i18n concept for typing Devanagari and Roman script using UTF-8, Unicode. So input will be like, “What day is it today, आज Sunday आहे . आज हम या करंगे . आपण आज university ऱा visit देवू”. So such type of input will be converted into speech. To use combination such languages in written and spoken there is need of transliteration and syllabification. There are many applications in which speech enabled interface is required. But there no such system available that reads the multilingual text with Devanagari and Roman script. There are mainly three approaches are used in speech synthesis Concatenative speech synthesis, Formant synthesis and Articulatory synthesis. Each approach has its own strengths and weakness. Comparatively Concatenative speech synthesis is widely used. Presently unlimited words can be generated by concatenating the phonemes, but we do not get the natural quality speech using this method. When we use phoneme unit for concatenation the concatenation cost increases, also speed and we do not get quality of speech. Instead of selecting unit as phoneme if we select Syllabified units of words (all-phones) as segment we can generate unlimited, natural quality speech. When we use one writing system like we can use ASCII character, but this code table is not sufficient for multiple languages. The proposed system will work with high accuracy and has a great potential to be used for variety of applications. Need of TTS: In the application field of Text to Speech is expanding fast while the quality and use of TTS system is also increasing steadily. These synthesis systems are more affordable for common customer and also more suitable for everyday use. Identification of Problems: In the present Text to Speech System different journals, Conferences and Books have been considered from year 1990 to 2014 Problem Identification: At different level different problems are identified.