Main fundamental challenge for recent research work on speech based on science and technology is to understand and model the user variants in Spoken Languages. Users have their style of speaking, reliant on various factors, adding the dialect and accent of the speaker as well as the social and economic
background of the speaker and contextual attributes like degree of knowledge between the listener, speaker and the position or rank of the speaking condition, from very normal to formal. In the past few decades, an extensive progress has been seen in automatically verifying the language of a speaker offered a sample speech. The main purpose of dialect verification is the recognition of a speaker’s region dialect, within a pre-determined language, offered the acoustic signal alone. DR (Dialect Recognition) is a main issue in particular, since even within the similar dialect and accent or register user change may occur. For illustration, In Spontaneous speech, few speakers tend to exhibit more optimizing and alteration of function words than others. The main issue of dialect recognition system has been viewed as challenging than that of language classification or recognition due to the maximum similarity among dialects of the similar language. While, dialects may differ in any dimensions of the linguistic spectrum such as syntactic, lexical, morphological, phonological differences, these changes are likely to be more indirect across dialects than those across languages such as Hindi, Punjabi and English etc.