Audio Watermarking Based on the Psychoacoustic Model and Modulated Complex Lapped Transform

The music industry suffers a multi-billion dollar annual revenue loss due to piracy. Audio watermarking technology is an important technique that is gaining widespread use. However, one of the challenges facing watermarkers is the high sensitivity of the Human Auditory System (HAS). In this paper, we present a novel audio spread spectrum watermarking scheme where the watermark is extracted from the original file.By using the psychoacoustic model this technique efficiently takes advantages of masking phenomena in HAS, in order to embed an important amount of watermark data below the masking threshold of the original input signal. The watermarking is performed in the Modified Complex Lapped Transform (MCLT) domain, which presents a good loss-less reconstruction for audio reconstruction. The computer simulation shows a good imperceptibility and a high robustness to a wide range of unintended and intended attacks.