An analysis on LPC, RASTA and MFCC techniques in Automatic Speech recognition system

Speech is an ancient field of study and research is being done on it till date. Automatic Speech recognition system deals with analysis and recognition of the input speech signal by the machine or computer in various environments. To enhance the accuracy and capability of the system various feature extraction techniques are implemented. This research paper provides a brief overview of Speech recognition system and its various phases like analysis, feature extraction, modeling and testing or matching. In addition it also includes detailed and comparative study on Linear Predictive Coding (LPC), Relative Spectral Filtering (RASTA) and Mel-Frequency Cepstral Coefficient (MFCC) feature extraction techniques used in Automatic Speech Recognition systems. The main objective of this research paper is to briefly summarize speech recognition system and three feature extraction methods that are an integral part of ASR.