A Spread Spectrum Audio Watermarking System with High Perceptual Quality

We propose a system to embed watermark message into audio signal, which can be used for copyright protection. It uses spread spectrum theory to generate a watermark that resistants to different removal attempts. We exploit the psychoacoustic auditory model to guarantee the audio signal's perceptual quality after the watermark embedding procedure. Recovery is performed without knowledge of the original audio signal. We use different pseudo-random sequence in synchronization marks and watermark message to avoid the false synchronization in watermark message. Experimental results show that our watermarking scheme is robust against common signal processing attacks while keeps the watermarked audio a high perceptual quality.