Multi-Pitch Detection with Gender Identification and Emotion Detection

Abstract

Pitch is a perceptive attribute of a sound. Pitch has a very important acoustical feature in speech analysis play a vital role in many applications like speech recognition, prosody analysis, speaker identification, emotion recognition and computational auditory scene analysis (CASA). Pitch perception is very complex process. Pitch determination of single source is easy as compare to multi-Sources or of polyphonic signals. The problem of estimating the fundamental frequency or pitch of periodic waveforms occurs in various form of application, and has received notable interest over the recent years, for example, several speech and audio problems notably depend on the initial forming of an estimate on the pitch or pitches including problems. Fundamental frequency plays a vital role in Gender identification also. There are many effective and accurate proposed algorithms on single source pitch determination and detection. But the multi-pitch real life scenario occurs regularly than single pitch case, and often also in speech processing. It is difficult to accurately estimate the multiple pitches of a mixed signal; multi-pitch estimation has potential application in speech sources separation, speech enhancement, speech recognition.