EXTRACTION OF PHONEMES FROM THE MERGED SPEECH AND THEIR IDENTIFICATION