In this paper a system is being proposed to classify gender of a speaker depending on his/her voice. Earlier, pitch was used to differentiate between male and female voices. But this method basically not work in cases of similar pitch value for both male and female voices. Thus, MFCC (Mel Frequency Cepstral Coefficient) is used in this proposed method for feature extraction from the speech which best suits the current scenario. ANN (Artificial Neural Network) performs gender classification using extracted features obtained from the speech. The experimental analysis shows that using MFCC for feature extraction with ANN as a classifier can give better results particularly in case of voice based gender identification.
Volume 11 | 05-Special Issue
Pages: 2136-2142