Research On Digital Audio Watermarking Using Frequency-Hopping Spread-Spectrum Scheme

In recent years, the rapid development of computer and network technology allows us to access and disseminate multimedia works quickly and easily. But it also results in a series of security and copyright protection issues of multimedia digital works. As digital Watermarking is an effective solution, it has obtained more and more attention. In this paper, we firstly review the current situation of the research of digital watermarking, and analysis the system models, basic principles and the evaluation criteria of digital audio watermarking. After understanding the features and common technical of audio signal, we propose two novel digital audio watermarking schemes based on frequency-hopping spread-spectrum(FHSS) method to improve the existing spread-spectrum(SS) watermarking scheme. The innovation and results are as follows:(1) A frequency-hopping watermarking scheme using the single tone(STFH) is presented, in which SS sequence is replaced by the single tone(ST) sequence. Watermark bit is embedded into the whole frequency domain of original audio by SS sequence, while STFH adds watermark symbol on a single frequency which has better auditory perception. The bit error rate(BER) and the discrete channel capacity of binary watermark is derived.(2) On this basis of STFH, we design the other frequency-hopping watermarking algorithm using orthogonal single tone sequences(OSTFH). Multiple orthogonal single tone sequences are used to greatly improve the capacity of discrete channel. The bit error rate(BER) and the discrete channel capacity of binary watermark is also derived. The theoretical expression of the optimum parameter for the maximal channel capacity is obtained, and it is proved to be true under the condition that the ratio of host audio signal to noise(SNR) is infinity. We introduce the theory of channel coding to reduce the bit error rate since the maximal channel capacity of our algorithm is larger than that of the existing scheme.(3) The audio attack is tested under the same payload. Experimental results show that the proposed algorithm OSTFH has better imperceptibility and achieves higher robustness against common audio attacks.