Bit Replacement Audio Watermarking Using Stereo Signals

In this paper, a blind audio watermarking scheme using stereo signals is proposed. In the proposed scheme, each bit of watermark bit stream is embedded into ldquonon-silentrdquo samples of stereo audio signals. The watermark bits can be embedded by replacing least significant bit or bits in higher bit layers of original signal samples. Robustness and performance of this audio watermarking scheme under different types of attacks, such as addition of noise, MP3 compression and re-sampling are investigated. Watermark extraction rates are calculated. Different audio signals including music, speech are used for the experiments. It is found that the proposed audio watermarking scheme, compared with the use of single channel signals, is much more robust against addition of white noise.