Transform Domain-based Digital Audio Watermarking Algorithm

With the popularity of computer and the Internet, a large number of digital products are emerging and widely disseminated. We have entered the information and the digital age, Digital audio as an important part of digital products has a profound impact on our lives. With the development and wide application of digital technology, people begin to pay more attention to protect the safety of digital products. Digital audio watermarking technology is an effective way to protect the security and rights, which has becoming a hot topic of today's research.The algorithm of digital audio watermarking can roughly be divided into two broad categories, the time domain algorithm and transform-domain algorithm. The time-domain algorithm has less robust, so it is a good way to hide the watermarking by changing the frequency domain coefficients of the audio signal. Therefore, current researches focus on transform domain.Digital watermarking involves a large number of images, audio processing algorithms, mathematical calculations and tools, etc. it is necessary to use scientific and engineering calculation software which is of high-performance, Matlab provides toolboxes of image processing, wavelet analysis and digital signal processing, it is very good option for digital watermarking.The focus on this paper is the algorithm of transform domain on digital audio watermarking, designed two kinds of watermarking algorithm according to the different types of watermarking. the research results are as follows:(1) Introduced the significance and background of the subject, describes the relevant knowledge about digital audio watermarking technology, and analyzed the existing algorithms.(2) Designed a high-capacity digital audio watermarking algorithm based on DWT and DCT. Doing watermarking pre-processing by using Arnold transform and DCT, decomposing the original audio by DWT, combined with the auditory characteristics and masking effect, modifying the wavelet coefficients overall to achieve watermarking embedding, this method enhanced the stability of the watermarking. Because of the rounding errors, the values of watermarking fluctuated widely, provided a simple, efficient method to select the express value. This algorithm enriched the type of watermarking, it provides a solution for high-capacity watermarking.(3) Designed a double digital audio watermarking algorithm based on DWT, owing to the wavelet decomposition of audio matched to Auditory characteristics, the difficulty of perceiving watermarking are unequal when the watermarking are embedded in the different wavelet components. The robustness of watermarking is difference while audio are attacked. Using this method of watermarking re-embedding, the watermarking is embedded into the high-frequency and intermediate frequency components, using pairs of watermarking to resist different attacks and enhance the protection for audio. It is possible to extract watermarking without the original watermarking.(4) Designed the experiment platform, doing simulation experiments by matlab to test the performance of watermarking algorithm.It can be seen form the experimental results that the two kinds of digital audio watermarking algorithm which are mentioned above are to meet the requirements of security and transparency, and can resist the common attacks such as filtering, resampling, noise interfering, and compressing effectively, it has good robustness.