D.C. Wyld et al. (Eds.): Advances in Computer Science, Eng. & Appl., AISC 167, pp. 843–852. springerlink.com © Springer-Verlag Berlin Heidelberg 2012 High Capacity Lossless Semi-fragile Audio Watermarking in the Time Domain Sunita V. Dhavale, R.S. Deodhar, and L.M. Patnaik Defence Institute of Advanced Technology, Girinagar, Pune-411025, India sunitadhavale75@rediffmail.com, {rsdeodhar,lalit}@diat.ac.in Abstract. A blind high capacity lossless semi-fragile audio watermarking algo- rithm based on the statistical quantity related to the correlation among the audio sample values is proposed. Time domain embedding is used to reduce the compu- tational time in searching the synchronization codes. The watermark is embedded into the non-silent high energy frames (HEF) to take advantage of the perceptual properties of the Human Auditory System (HAS) and to improve the transparency of the digital watermark. The Offset value used for embedding is made adaptive to the required SNR for the final watermarked audio signal. The watermark can be removed using a secret watermarking key with only minimal remaining distortion. The method proposed is media format independent and it can be used with lossy compression. Both subjective and objective tests reveal that the proposed water- marking scheme maintains high audio quality and is simultaneously highly robust to pirate attacks, including MP3 compression, cropping, time shifting, filtering, re- sampling, and re-quantization., and re-quantization. Keywords: Audio watermarking, digital rights management, Lossless, Reversi- ble, Self Synchronization, Time Domain, Semi-fragile, Blind. 1 Introduction Due to outstanding progress of digital audio technology, ease of reproducing and retransmitting digital audio has been greatly facilitated. Hence there is a need for the protection and enforcement of intellectual property rights for digital media. Digital wa- termarking is one of the promising ways to meet this requirement. The primary objec- tive of digital watermarking is to hide the copyright information (e.g. owners/company name, logo etc.) into a multimedia object, without disturbing the perceivable quality of the content [1]. Watermarking of audio signals is more challenging compared to the watermarking of images or video sequences due to wider dynamic range of the human auditory sys- tem (HAS) in comparison with the human visual system (HVS). Two properties of the HAS dominantly used in audio watermarking algorithms are frequency masking and temporal masking. According to the International Federation of the Phonographic Industry (IFPI), SNR of watermarked audio signal should be always greater than 20 dB. The embed- ded watermarks should not be removed or degraded using common audio processing techniques. The watermark embedding process should be faster, so that integrated