Human-human communication in social environment is only possible through speech, facial expressions and bodily changes. In this research work, speech and facial expressions is used in order to estimate basic emotions (angry, sad, happy, boredom, disgusting and surprise) using a computer. A new asynchronous hybrid technique through Relative Bin Frequency Coefficients (RBFC) and Relative Sub-Image Base (RSB) is used in order to deal with the above modalities which are different in time, scale and dimension. Support Vector Machine (SVM) is used for classification. From experimental results, the proposed system performs better than the conventional systems. © 2005 - 2015 JATIT & LLS. All rights reserved.