A data hiding for audio using band division based on QMF bank

In the watermarking techniques for audio are required quality, security and robustness. This paper presents method that improves security and robustness, besides restricting the noise. Special attention is given to band-division by using a filter bank based on the Quadrature Mirror Filter (QMF) that several audio compression standards adopt. First, we divide the time-domain data into several subbands and select the sub band to embed the watermarking. Then, we spread the signal included in the selected sub band using Direct Spectrum Spreading (DSS), one of the spectrum spreading methods, and embed the watermarking in the frequency-domain data. Finally, we reconstruct the signal by synthesizing whole subbands. By this method, we can reduce the noise and improve security and robustness against MPEG1 Audio Layer3 compression. Experimental results show good robustness of the approach against some signal processing manipulations. Quality and robustness are evaluated by computer simulation