A Digital Audio Watermarking Using Two Masking Effects

Abstract

In this paper, a new digital audio watermarking algorithm is presented. The proposed algorithm embeds watermark by eliminating some frequency information of audio signal based on human auditory system (HAS). This algorithm is a blind audio watermarking method, which does not require any prior information during watermark extraction process. In this paper, two masking effects are used for audio watermarking, frequency-domain masking effect and time-domain masking effect. This algorithm finds watermarking position using time-domain masking effect and embeds/detects the watermark using frequency-domain masking effect. Detection of embedded watermark is obtained by finding the eliminated frequency using band-pass filter. By using the two masking effects, the degradation of the audio quality can be minimized. It is confirmed that the proposed algorithm has robustness against various attacks such as cropping, down sampling, time stretch, MPEG1-layer3, and MPEG2-AAC compression with good audio quality.