A speech signal filterbank, using melscale frequency and framebanking.- 1. For speech signal can be represented as a discrete sequence of frames (or feature vectors) that can be used as the input to a speech recogniser. Important ideas and technique