The development of such system that automatically recognizes the input speech and translates in another language like Sanskrit is a challenging task. Sanskrit language is much more conjured language. The purpose of this paper is to explain a system which convert the English Speech into English text and then translate that English text into Sanskrit text and again convert that into speech. This system falls into the category of Speech-to-Speech translation. It unifies the isolated words class under the Speech Recognition type, traditional dictionary rule based machine translation approach and text to speech synthesizer. So basically it is classifies into three areas: Speech Recognition, Machine Translation and Speech Synthesis. This system matches tokens [1] from database to differentiate Subject, Object, Verb, Preposition, Adjective, and Adverb. This paper presents approach for translating well-structured English sentences into Sanskrit sentences. Since Sanskrit is free ordering language (or syntax free language) or we can say its meaning won't be change even if the order of words changes.

Rule based dictionary approach Synthesizer Translator Speech Recognition Rule based machine translation