The Research On Audio Digital Watermarking

With the development of multimedia and network, the distribution of digital audio becomes much easier and faster. But it brings the problem of ownership protection, so it is important to protect digital audio. As a solution, digital audio watermarking is an effective method to solve the problem.The algorithms based on cepstrum transform are very few, and there are some problems such as some algorithm is not blind detection and robust character is insufficient and so on.Based on the analysis,comparison and test of existing digital audio watermarking some digital audio watermarking algorithms are proposed and improved as followed:The paper proposes a blind audio watermarking algorithm based on the cepstrum domain. The algorithm of the embedded watermark is very important in order to solving the conflict between imperceptibility and robust. With the characteristics of complex cepstrum, the coefficients of complex cepstrum are transformed and we embed a binary gray image in the original audio signal so as to meet he utmost need of imperceptibility and robust of audio digital. The embedded watermark is imperceptibility and robust to many common attacks, such as noise adding, resampling, low pass filtering, requantization and so on.The first algorithm can not be robust to cropping attack, so the paper also proposes an audio watermark algorithm with self-synchronization based on complex cepstrum domain. At first,it divides original audio signal into several segments, then synchronization code is embedded into audio with the water- mark. The watermarks can be extracted without the original digital audio signal. Compared with the SMM algorithm, not only the embedded watermark is perceptibility, and also the correlation is improved significantly after some common attacks, such as noise adding, resampling, low pass filtering, requantization and cropping and so on。...