Robust Audio Watermarking Based on Secure Spread Spectrum and Auditory Perception Model

The impetus for the work presented in this paper arose from the need to provide copyright protection for digital audio, including CD-quality music and DVD, under heavy compression rates, conventional audio processing manipulations, and transmission through noisy media. Our approach is established on modelling audio watermark as information transmitted through a probabilistic communication channel represented by digital music. Watermark information is embedded within significant portions of the audio streams making it tightly coupled with audio content and able to sustain even the most severe attacks intended to remove the hidden data. This paper presents how highly transparent, robust and secure watermark information can be embedded into digital audio streams using the model for acquiring significant audio components in conjunction with conventional spread spectrum data hiding techniques.