USING THE SDIF SOUND DESCRIPTION INTERCHANGE FORMAT FOR AUDIO FEATURES

摘要: We present a set of extensions to the Sound Description Interchange Format (SDIF) for purpose storage and/or transmission general audio descriptors. The aim is allow portability and interoperability between feature extraction module an information retrieval application remaining modules, such as training, classification or clustering. A techniques addressing needs short-time features temporal modeling over longer windows are proposed, together with mechanisms that further adaptations by user. paper completed overview aspects SDIF its practical use means existing programming interfaces for, among others, C, C++ Matlab.