Research On Adaptive Audio Watermarking Algorithm Based On QIM

With the development of the Internet, network audio document transmission is faster than before, Audio copyright problem also will be more and more serious. Therefore, the audio watermarking has been under more and more attention from researchers.This paper has been studied audio watermarking technology, and analyzes and summarized the current research status and existing problems of the audio watermarking, proposed research ideas based on quantitative, designed two audio watermarking algorithms based on QIM.And then studied non-perception and robustness of DM and DC-DM algorithm.Firstly, for traditional QIM algorithm quantization step can not be control, And the Sharp decline in detection performance receiver When the external noise intensity over a certain range, Introduce A Dither Modulation with Distortion Compensation Audio Watermarking Algorithm Based on DWT. Algorithm selected low frequency coefficients of wavelet transform to embed data and used the Fibonacci transformation scrambling the watermark information to improve security of watermark scrambled, The characteristic of the methods is able to control the expansion factor, so as to achieve the purpose of controlling the quantization step.Secondly, for previous QIM Embedding Method can not adapt to human's auditory system, introduce DM audio watermarking algorithm which can adapt to human's auditory system. This method used DM which can insure the relatively independence between quantization noise and carrier signal watermarking. Also, it was based on Frequency masking characteristics that allows the size of the quantization step size according to the frequency adaptive control of the human auditory system. The adaptive of quantization step improved the performance of non-perception and robustness. It can adaptively select the longer quantization step with larger audio frequency, so robustness is improved; It can adaptively select the shorter quantization step with smaller audio frequency, so non-perception is improved. Embedded DM and detection algorithm improved the anti-attack of watermark and the fidelity of the audio carrier.Through experimental analysis, the two proposed audio watermarking algorithm based on QIM can achieved quantization step controlled and adaptive selection.To control the quantization step and choice, which makes audio watermarking system is robust and imperceptible. Because the quantization step to control and select, which makes audio watermarking system have better robust and non-perception than the traditional QIM algorithm.