Audio watermarking with high embedding capacity based on multiple access techniques

This paper deals with a new digital audio watermarking scheme based on multiple access techniques. In digital communication, multiple access techniques allow several users transmissions by taking the same communication channel. The present work proposes to embed multiple sub-watermarks in the same channel that is the audio signal. Our main objective is to investigate embedding capacity (the amount of information that can be hidden) limitations of our proposed audio watermarking system. Three multiple access techniques have been employed in the new multi-watermarking system: DS-CDMA (Direct-Sequence Code Division Multiple Access), FHMA (Frequency Hopped Multiple Access) and FDMA (Frequency Division Multiple Access). Experimental results allow us to make the choice on the best multiple access technique, in the multi-watermarking system, which significantly permits the highest embedding capacity (a data embedding rate up to 6 kb/s) with almost no loss of data imperceptibility and with acceptable extraction fidelity (BER≃10−2) even in presence of disturbances.