Study On Adaptive Audio Watermarking Algorithms In Domain Wavelet

While the development of Internet and digital multimedia technologies, people pay more attention to information security. To solve copyright protection problem, a new technology—digital watermarking has been found to be a promising solution. In order to protect or authenticate the digital audio, a mark with special meaning is embedded into the audio, without affecting the quality of the audio.This paper reviews the background and significance of the audio watermarking, describes the development of digital watermarking technology. Then the principle of digital audio watermarking, some traditional algorithms and evaluation method of audio watermarking scheme's performance are introduced. The article mainly focuses on designing an audio watermarking algorithm with good robustness and less audible. The main work of this paper is given as follows.1,An improved adaptive quantization algorithm in wavelet domain has been proposed. Based on the advantage of the wavelet transform, the audio is transformed into frequency domain using discrete wavelet transform firstly. Then, the 2-dimensional digital watermark is embedded into the selected location of the low frequency coefficients using quantization technique. The quantization step is determined adaptively by the coefficients'values in each frame. Experimental results show that the proposed watermarking scheme has good transparency to many different kinds of audios, and is robust to common digital signal processing such as re-sampling, re-quantifying, adding noise, low-pass filtering and lossy compression.2,In wavelet packet domain, an adaptive algorithm using masking threshold has been designed. The existing audio watermarking algorithms based on psychoacoustic model are non-blind extraction, or time-consuming. In this paper, the masking threshold is calculated in wavelet pocket domain, and the coefficients suitable to the embedding are selected. The algorithm has the following advantages:(1)the sub-bands are got more reasonably; (2)chaotic sequences are used to encrypt the watermark information; (3)because masking threshold is used as the quantization step, the watermarked audio has good inaudibility; (4)the original audio is not needed during the detection, so it is suitable for practical application. Simulation and experimental results show that the algorithm can preserve inaudibility and good robustness against signal processing.