The Embeding Strength Analysis And The Algorithm Research Of Audio Information Hiding

Because of the rapid development of multimedia technology and computer network, along with more and more widely usage of voice and image in communication, information security in the environment of network communication system is becoming more and more important. Malicious individuals’or groups’activities of stealing communication information are increasingly rampant, this forces the users not only asking for communication of high quality and functional requirements, but also paying more attention to secure transmission of information, which can be solved by information hiding technology.Audio information is the usual content in network communication, so the research of audio information hiding technology has very high practical value. The study of image information hiding technology is more common than audio now, this is because the human auditory system is more sensitive than the human visual system, the practice of imperceptibility in audio information hiding is more difficult than the practice in image. Therefore, in the design of audio information hiding, we should be more careful when we are selecting embedding position and controlling embedding strength.Information hiding method in time domain is presented earlier, and some classic algorithm have been applied to various fields of digital watermarking. Time domain algorithm is easy to be implemented and has great practical value. But the time domain algorithm is not robust, so the application of it is faced with some difficulties. The current study of audio information hiding algorithm usually combined the analysis of audio signal in time domain with analysis in transform domain to enhance the robustness of the algorithm, but this will sacrifice some advantages of time domain algorithm.To solve the problems above, two part of researches have been done in this paper:(1) Audio information hiding algorithm is based on the analysis of audio signal’s transform domain or time domain. By modifying the location specific data, information is embedded into the source carrier. Based on the DCT and DWT principle, this paper analyzes the effect to the original audio signal when the transform domain coefficient is changed; combined with the characteristics of human auditory system, this paper analyzes the embedding strength meeting imperceptibility when DCT or DWT coefficient is modified.(2) In this paper, combining with the neural network and chaos theory, a time domain audio information hiding algorithm is introduced. The algorithm modifies the quantized audio data, makes the audio signal to meet specific time domain waveform, so as to achieve the purpose of embedding hidden information. The experimental results show that the algorithm runs fast, has large embedding capacity and good robustness.