The idea of acquiring the human emotional state from one's voice, we've collected the necessary data that helps us understand the principle behind the process. Human emotions can be predicted by his / her facial expressions or voice tone. One of the main tasks involved in image processing is the interpretation of the facial expressions. Computer is well-trained to identify and distinguish human faces, and is also trained to recognize emotions. It is programmed and trained using examples from the real world, since emotions often vary in facial expressions. Each emotion has a different expression to it. It can be trained using various faces in this way. Every emotion likewise holds different tone in one's speech. To measure and evaluate the emotions it needs a particular level of emotional tones. We need to retrieve estimated emotional frequencies. It's the daunting job, as each speaker has different pitch when speaking, and the same person's pitches vary in emotion. Another big problem is the noise in the audio when a person is speaking, because of low quality recordings or surroundings. The emotion list is Happy, angry, sad, bored, shocked. The prior important principle for this project is speech recognition. The computer must be able to interpret the input as a speech in shape, and must be able to recognize the words. The provided input is converted to text format, and the text is parsed in tokens. Computer must also be capable of fetching the frequencies at the same time.The calculation is carried out using several established methodologies, and we have put into practice another methodology which is capable of recognizing emotions.