High Imperceptible Copyright Protection: Adaptive Wavelet for Security

Abstract

This study is an extension of a recently conducted study in audio watermarking system. This study shows the imperceptibility of audio watermarking system based on Bose-Chaudhuri-Hocquenghem (BCH) code, Discrete Wavelet Transform (DWT) and Singular Value Decomposition (SVD) methods. Watermarked image is embedded in the audio host signal accompanied by several attacks. The audio watermarking performances show that perceptual quality of audio with attacks is varies. At quantization spacing of 0.1 perceptual quality, objective difference degree (ODG) can reach -1.55 for level 1 in the audio watermarking system with echo attack. While signal to noise ratio (SNR) can reach 10.5 dB in the audio watermarking sistem with echo attack and bit error rate (BER) performance can reach 0.03 in our audio watermarking system also with echo attack. However, the audio watermarking system with equalizer attack shows that the higher the DWT level, the better imperceptible and BER performance is. Different from that of the system with equalizer attack, the watermarking system with resampling attack shows that the higher the DWT level, the better imperceptible and SNR performance is.