Abstract: Speech is known to be the most convenient form of communication between individuals. This study deals with the modus operandi for estimation of basics acoustic features as also of the function of the vocal tract area, direct from the acoustic of speech wave form. Linear Predictive Coding (LPC) is used for estimation of the function of vocal tract area. LPC is based on autoregressive modelling of the speech signals. The method is an effective technique for speech analysis. It is predominant in speech research work for locating and estimating basic acoustic features and vocal tract area function and also to represent the speech for low bit rate transmission and application in storage area.The principal objective of emotional speech analysis/recognition is to identify the different basic emotional states (primary emotions) and to categorize them under positive (non-negative) and negative emotions. Emotion analysis finds use as a tool for improvement of the quality of service found in many speech supported application like call centres. It also helps in interfaces for human computer application and others.