International Journal of Inventive Engineering and Sciences (IJIES) ISSN: 2319–9598, Volume-2, Issue-5, April 2014 4 Published By: Blue Eyes Intelligence Engineering & Sciences Publication Pvt. Ltd. Abstract— This paper aims at providing design and development solution of an Android application whose objective is to provide a solution to overcome the barrier of languages by implementing text to speech conversion in different languages. The Android application developed text to speech conversion to facilitate the translation of English language text into speech output in different languages. We have also proposed few improvements which can further advance this system to include more target audiences so as to make it more beneficial and useful. The proposed English Text to Multilingual Speech Translator using Android (T2MSTA) aims at providing assistance to the people lacking the power of speech or non-native speakers like people who do not share a common language. Index Terms— Android, T2MSTA, Text to Speech conversion, Intent, MATLAB. I. INTRODUCTION Android is probably the most popular operating system being used by millions of smart phones and tablets today and is growing by leaps and bounds. This is one of the most sophisticated and user friendly platform. Thus we have selected android as a platform to develop a system that provides a solution for the people who are unable to read or speak a particular language either because of illiteracy or not sharing a common language, or for other reasons. There are many procedures available in the market for text to speech conversion [1][2][3]. Each language uses different sets of algorithms and these cannot be employed to translate languages. Some multilingual translators are also available but they are limited for text to text [4][5][6][7]. Speech to speech translation is also available [8][9]. The proposed method is a simple and handy solution to overcome the barrier of languages by developing an android application for multilingual translation implementing text to speech conversion. It provides the methodology to implement text to speech conversion to facilitate the translation of English language text into speech output in different languages as desired. Design and development of this T2MSTA will thus provide assistance to the people those lack the power of speech or non-native speakers. Manuscript received on April, 2014. Vivek Hanumante, Department of Electronics & Communication Engineering, NIT Arunachal Pradesh, Yupia, India. Rubi Debnath, Department of Electronics & Communication Engineering, NIT Arunachal Pradesh, Yupia, India. Disha Bhattacharjee, Department of Electronics & Communication Engineering, NIT Arunachal Pradesh, Yupia, India. Deepti Tripathi, Department of Electronics & Communication Engineering, NIT Arunachal Pradesh, Yupia, India. Sahadev Roy, Department of Electronics & Communication Engineering, NIT Arunachal Pradesh, Yupia, India. With the design and development of this smart multilingual translator we strive for developing a system which will in turn embellish the devices which assists aforementioned people. Thus a novel system T2MSTA has been proposed which can take input from user and can translate in various languages. T2MSTA so far can take an input of a single word or a sentence in English language and can translate as output voice in three different languages namely Hindi, English and Bengali as per users’ choice. The system can be further upgraded to generate syllable based sounds for the words which are not present in library stored [10] [11]. II. OVERVIEW OF T2MSTA The proposed T2MSTA is centralized on an objective which is a twofold: i) Text to Speech conversion and ii) Language Translation. Sound libraries of different languages for T2MSTA has been created using MATLAB, which includes most frequently used words in daily life conversations. In the android application user is asked to enter the text and click on the button named with language to listen the translated speech output. Figure 1 presents the system architecture of T2MSTA: Fig. 1. System Architecture III. DEVELOPMENT AND OPTIMIZATION OF SOUND LIBRARY Libraries for T2MSTA have been developed using MATLAB. Firstly the sound units are stored in .wav format, which then, is converted into .ogg format after certain English Text to Multilingual Speech Translator Using Android Vivek Hanumante, Rubi Debnath, Disha Bhattacharjee, Deepti Tripathi, Sahadev Roy Android application Text Input from User Selection of language from options Text Pre-processing Word Comparison Plays the sound unit that matches the input word Sound unit Library Translated text in form of speech at output