Robust, transparent and high-capacity audio watermarking in DCT domain

Novel secured scheme for blind audio/speech norm-space watermarking by Arnold algorithm

A scheme jointly exploring the rational dither modulation (RDM) and auditory masking properties in the discrete wavelet transform (DWT) domain is proposed to achieve effective blind audio watermarking. The embedding of binary information is carried out by modulating coefficient vectors in the 5th-level approximation subband using the quantization steps estimated from past watermarked vectors. The robustness and payload capacity of the proposed scheme are maneuverable by varying vector dimensions, while the imperceptibility is ensured by constraining quantization noise below the auditory masking threshold. Furthermore, the periodic characteristic inherited in the RDM formulation can be used to re-establish synchronization for accurate watermark extraction. Experimental results show that the proposed DWT---RDM approach renders a near-zero objective difference grade in the perceptual evaluation of audio quality even when the signal-to-noise ratio maintains at a level near 20 dB. In most digital signal processing attacks, the bit error rates of retrieved watermarks are sufficiently low as compared to other recently developed methods with fewer payload capacities.

A DWT-Based Rational Dither Modulation Scheme for Effective Blind Audio Watermarking

Digital representation of multimedia is more advantageous than the analog one due to potentially improving the portability, efficiency, and accuracy of the information presented. Despite the challenge of having high hiding capacity like other media watermarking, still audios are likely candidates for data hiding due to their possible capabilities of achieving impressive robustness and protection against online music piracy and content identification. In this paper, we propose efficient audio watermarking embedding and extracting techniques, which mainly use Discrete Wavelet Transform (DWT) and Singular Value Decomposition (SVD), in which a new matrix formation of details sub-bands is proposed. Additionally, massive experimental work was conducted to investigate the contributions of operating different watermark intensities and multiple levels of DWT to our proposed techniques. Two performance objectives are employed in this work which involve imperceptibility and robustness. To further boost the imperceptibility, we incorporate the code assignment method to our techniques that do outperform what are closely connected in the literature.

/pdf/imperceptible-and-robust-dwt-svd-based-digital-audio-sml61971g9.pdf

Imperceptible and Robust DWT-SVD-Based Digital Audio Watermarking Algorithm

It is a challenging work to design a robust audio watermarking scheme against various attacks. Moments and moment invariants have become a powerful tool in robust audio watermarking owing to their description capability and invariance property. Exponent moments (EMs) is a new kind of orthogonal moment defined on the circular domain. EMs has a better reconstruction, lower noise sensitivity, lower computational complexity, and excellent invariant property. Based on invariant exponent moments and synchronization code technique, a new and robust audio watermarking method with good auditory quality and reasonable resistance against most attacks is proposed in this paper. Specifically, the watermark data is efficiently inserted in host audio by taking advantage of both invariant EMs and quantization index modulation (QIM). EMs magnitude is robust to common signal processing operations, and QIM can render watermarking scheme blind in nature. To resist against desynchronization attack, the synchronization code technique is also integrated into our invariant EMs based audio watermarking method. Compared with the typical and related audio watermarking schemes, the proposed method is more robust to most attacks and has higher imperceptibility. The effectiveness of our method is demonstrated by simulation results.

Ling-Yuan Hsu

Papers

Variable-dimensional vector modulation for perceptual-based DWT blind audio watermarking with adjustable payload capacity