 The analysis by synthesis theory claims that during listening we permanently compare the input with our own auditory experience. This activity is controlled by a recognition device in the brain and a set of auditory patterns with which the input is matched. The whole process looks like this. First, the ear amplifies the incoming signal and transmits it to the auditory nerve. Then the signal is filtered and non-speech components are removed and auditory patterns, such as formant frequencies, are extracted. At the end of the recognition process, phonological units, that is phonemes or features, are extracted from the signal and are combined to form larger units.