Analysis of Channel Capacity of Spread Spectrum Audio Watermarking System

Spread spectrum watermarking (SSW) is one of the most widespread watermarking methods. In this paper, the characterization of the SSW from the information theoretic point of view is performed. Previous works on channel capacity analysis of watermarking systems are reviewed. Since SSW method is designed so that the host signal is unknown to the embedding encoder, Costa´s results for capacity analysis of watermarking systems are not directly applicable. In this paper, watermarking embedding and detection is modeled as a channel that transmits information in the presence of noise of host and channel capacity analysis of that model is derived and the most reliable transmission rate of hidden information obtained. In addition, channel capacity for the case of presence of additive noise in the watermark channel is derived and robustness of the SSW against additive noise examined