ABSTRACT
This chapter argues that the reader is sufficiently familiar with audio watermarking to allow him to appreciate the analysis on the pros and cons of this technique with respect to the audio fingerprinting technique discussed next. It introduces the reader to the basic technical aspects of audio fingerprinting. The chapter provides an analysis on the commonalities and the differences between watermarking and fingerprinting and the reader with insights on which identification technology to use in which application. It discusses some issues for future research and presents the essential ingredients of an audio fingerprinting technology developed at Philips Research. The chapter argues that the choice for a particular technology is highly application dependent. It focuses on the different parameters of an audio fingerprint system, and presents the main parameters: robustness; reliability; fingerprint size; granularity; and search speed and scalability. The subfingerprints are obtained by comparing and thresholding energy differences. The proposed fingerprint extraction scheme is based on this general streaming approach.
