A Novel Audio Watermarking Algorithm Based on Reduced Singular Value Decomposition

This paper presents a transparent and accurate semi-blind audio watermarking algorithm, based on the Reduced Singular Value Decomposition (RSVD). The proposed scheme manipulates coefficients of one of the resulting unitary matrices for embedding watermark bits. To preserve audio fidelity a threshold-based distortion control technique is given. Test results on real music signals show that this watermarking scheme is transparent, accurate and robust against MP3 compression.