A Review Paper on MFCC based Hindi Speech Recognition System using HTK Toolkit

To implement the robot’s capabilities, it is important for us to communicate with them efficiently. Thus, Human Robot interchange is attracting the observance of the researchers these days. A speech recognition technique has been developed using various characteristic extraction method like MFCC (mel frequency cepestral coefficient), LPC (linear predictive coding) and HMM (hidden markov model) is used as the classifier. Less work has been finished for Hindi language in this domain with a vocabulary size not very large. So, work finished for Hindi database, with a vocabulary size a bit protracted. HMM has been enforced using HTK Toolkit. Afterwards the presentation of both of the method used have been compared. The work has been finished using audacity for sound recordings and Cygwin to implement the HTK commands in Linux kind environment in windows platform. As well as, the tech improvement has been checked in the speaker dependent and speaker independent both kind of environments, whose present results, as well as, the comparison graph of the technique shows that MFCC done well as compared to LPC in each and every condition.