Exploring Gender Classification from Speech Signal using Java-DSP Tool

Abstract

Abstract: More than a century, speech and audio processing
play a crucial role in the entertainment and telecommunication
systems. Due to the high demands, expecting more
sophisticated features, still audio and speech processing require
advanced techniques to serve better. This article focuses mainly
on the gender classification from speech signals based on pitch
period and pitch frequency analysis. To achieve the required
tasks, the online graphical simulation, such as Java-DSP, is
used here. J-DSP environment blocks can be dragged and
placed in the editor easily, which helps to reduce the time
required to complete the design issues. Also, the pitch period
estimation is done using autocorrelation and cepstrum
methods. The zero-crossing detection is used to focus on the
identification of the voiced/unvoiced portion of the speech. The
survey results from the participants who attended the training
program on speech and audio processing course were obtained
as feedback.

Index Terms:Autocorrelation, gender classification, J-DSP,
on-line DSP simulator, pitch estimation .