Research On Technology Of Digital Audio Watermarking

As a kind of effective digital copyright protection technological means, the digital watermark is a foreland subject in the field of information security presently in the world. And now the digital watermark both at home and abroad is mostly based on static picture imbedded, while less attention is focused on audio imbedded watermark. With the two-dimension binary image as watermark information, the audio data as imbedded target, on the basis of analyzing the digital technological characteristic and demand of audio frequency watermark in detail, in order to enhance the imperceptible and robust of the digital watermark, and with the improvement of the existing algorithm, this paper discusses two new digital blind watermark methods of audio for copyright protection.At one hand, there is an improvement of the existing algorithm — the digital blind watermark algorithm of audio based on discreet wavelet transform. In the pretreatment before the watermark is imbedded, besides using the nonrandom sort method, which also arranges in an order at random, to dispel the dependence of the element in the one- dimensional array, it introduces mistake control code (namely: Error correction code), which is quite matured technique in the field of telecommunication and so on, to raise the robust of the digital watermark. This article uses (7, 4) Hamming code to error- correcting code the watermark message. In order to enhance the robust of digital watermark algorithm, the error-correcting code is used for correcting the wrong message in the watermark if any, while distilling and inspecting in watermark information.On the other hand, it discusses the feasibility of substitute thediscreet wavelet transform of twain transform, which are combined by discreet wavelet transform and discreet cosine transform, to improve the robust of the digital audio bland watermark algorithm, on the basis of introducing the mistake control code of digital audio bland watermark algorithm based on discreet wavelet transform.The discreet wavelet transform is characteristic with multi-resolution, and the discreet cosine transform is excellent with the ability of de-correlation and cohesive energy. It combines the advantage of the two, and uses low frequency weight of discreet wavelet transform to discreet cosine transform, and then in return uses the result of the cosine transform - coefficient of the cosine transform - to embed in the digital watermark, thus raises the performance of the digital watermark algorithm.And there are a lot of imperceptible and robust experiments for both the two kinds of betterment algorithm. The experiment results indicate that the digital audio bland watermark algorithm, which uses the error control code on the basis of discreet wavelet transform, is good at imperceptible and better robust for familiar signal process than that of original algorithm; and that the digital audio bland watermark algorithm, which uses twain transform and error - control code, as a whole boosts up the robust of the algorithm, with huge enhances the robust when the environment has noise assault and a little descend when cut and low pass filter attacks.