Voice Feature Analysis For Speaker Recognition To Improve Content Management

We are introducing feature extraction approach which helps in content control on broadband television. Speaker recognition-based content control are more effective as compare to traditional permission-based method. Television content can be controlled based on the program requester’s age and program requester’s age can be identified based on his/her voice. Speaker is adult or not that can be identified based on his/her voice. Every person’s voice is unique based on voice features like pitch, frequency, jitter, amplitude, and spectral power. This article explains how we can effectively use these features to classify child and adult voice. In general, Mel-Frequency Cepstral Coefficient (MFCC) is used for effective feature extraction and it provides features’ numerical values. Artificial Intelligence (AI) based system uses these features for accurate speaker recognition. Child and adult voice can be differentiate at time, frequency, mel scale level, which can help for clear classification of the speaker. Power spectrum and mel-spectrum also represent difference in child and adult voice.