Non-blind audio-watermarking using compression-expansion of signals

One of the advantages of non-blind watermarking scheme is that the original signal must be available to extract the watermark. As such, the possession of the original signal becomes important in the proof of ownership. In this paper, a non-blind audio-watermarking method is proposed. The method makes use of compression-expansion of adjacent frames of samples of the audio signal. Some frames of the signal are thus 'modified'. The modified frames can be readily identified by comparing them with the original signal. The modified frames can be used to embed hidden information. By applying psychoacoustic model in the compression-expansion process, it is found that the modification does not affect the quality of the audio signal. The embedding of watermarks for single channel audio signal and two-channel stereo signals require different approaches which are described in detail in the paper. The performance of the method under several types of attacks is assessed. It is found that high percentage of recovery can be achieved with spectrally rich audio signals.