Recognition of speech with the help of machine automatically is an important area of research for more than forty years. As Voice is a signal of limitless information, so processing of speech signal by converting into digital is very efficient tool for high and precise automatic signal or voice recognition technology. Speech recognition has found its application in various areas of our daily lives as automatic answering machine to convey text and issue voice signal for the machines. Feature Extraction and Classification are main processing part of ASR system. The main part for the speech processing system to improve capability is the selection of Feature Extraction method which plays an important role in the system precision. This paper gives brief overview on the survey of various methods in speech processing such as `Wavelet Transform' and `Soft Computing Techniques' as ANN, HMM and GMM for isolated digit recognition.