Betreuer:

The present work describes the design, implementation and evaluation of a system for automatic audio signal classification. The signals are classified according to audio type, such as speech, background noise and several musical genres. The classification process is organized hierarchically, that is, as a succession of type decisions. At the highest level of the hierarchy, signals are recognized as speech, music or background noise. Speech signals are further divided into male speech, female speech and speech with background music or noise. Special focus is given to the distinction between music genres, where a total of 13 classical and non-classical genres have been considered. Also, an effort was made in selecting the music genres, aiming at simplicity and generality. A large number of audio features are evaluated for their suitability in such a classification task, including well-known physical and perceptual features, audio descriptors defined in the MPEG-7 standard, as well as new features proposed in this work. They are selected with regard to their robustness to noise an