Speech signal carries rich emotional information except semantic information. Speech and emotion recognition improve the quality of human computer interaction and allow more easy to use interfaces for every level of user in software applications. Common  emotion namely exclamatory, neutral and question mark were discussed and recognize through a propose frame work which combines of  Mel Frequency Cepstrum Coefficients (MFCC) and Power spectrum are used for feature extraction and back propagation neural network are used for recognition  of the emotional speech signals. This further will be used for transplanting that emotion in synthetic speech so that output quality of synthesis is improved.