A Redundant Audio Watermarking Technique Using Discrete Wavelet Transformation

Abstract

A frame based audio watermarking scheme using Discrete Wavelet Transformation (DWT) is presented in this paper. The algorithm embeds the copyright information into the consecutive levels of approximate coefficients of the DWT transformed audio signal. To increase the robustness of the proposed scheme, the information bits are embedded multiple times at the different levels. Embedding of redundant bits results in degradation of SNR and PSNR values; however, experimental results show that the degradation are within acceptable range. Extraction of the watermarked audio signal is carried out by considering the majority of the watermark bits extracted from different levels. The proposed scheme demonstrates results which tend to resist infiltration against many signal processing manipulations, like MP3 compression, re-sampling, quantization and noise addition.