An adaptive and secure audio watermarking algorithm robust to MP3 compression

MP3 is an important and popular audio compression standard. Watermarking directly in MP3 stream can result in better performance (robustness and fidelity) against MP3 compression. This paper presents an adaptive and secure watermarking algorithm for MP3 compressed audio signals, based on human auditory system. In the proposed algorithm, the watermark is embedded adaptively and inaudibly after the Modified Discrete Cosine Transformation (MDCT) and before the quantization. The Gaussian distribution statistic analysis is introduced to make this watermarking algorithm adaptive and correlation coefficient analysis is used to implement blind watermark detection. We tested our adaptive watermarking algorithm on various types of audio signals. The experimental results show that the new algorithm can survive most common signal manipulation including MP3 compression and other attacks.