An Approach towards Robustness of Digital Audio Watermarking

Abstract

Digital Watermarking has been proposed as a mean to identify the owner or distributor of digital data. It involves the concealment of data within a discrete audio file. Audio Watermarking has been proposed as a possible solution, embeds copyright information into audio files as a proof of their ownership. In this paper, we propose an effective, robust and inaudible Audio Watermarking algorithm. Audio Watermarking scheme has been carried out where a visually recognizable binary image is used as a watermark. The effectiveness of the algorithm is brought by applying a cascade of two powerful mathematical transforms: the discrete wavelet transform (DWT) and the singular value decomposition (SVD). In the proposed algorithm, watermark bits are not embedded directly into DWT’s coefficient, but rather on the elements of singular values of the DWT sub-bands of the audio frames. This system is implemented using MATLAB programming language and validated through a set of music records. Experimental results are presented in this paper to demonstrate the effectiveness of the proposed algorithm.