Digital Media Processing For Multimedia Interactive Services, 2003, pp.338-344
Abstract The paper presents a novel classification and segmentation scheme for MP3 and AAC audio in the compressed domain. The input audio is split into speech, music and silent segments using features such as total energy, band energy ratio, pause rate, subband centroid and fundamental frequency. Simulation results show the efficiency of the proposed algorithm.
Session 4: Audio Processing
World Scientific Books (World Scientific Publishing Co.)
View this record in World Scientific Books