A new average quantized watermark for audio signal

Embedding watermark to audio signal has two affections, if the embedded signal is too light, it can't be extracted, if the embedded signal is to heavy, it's easy to feel the distortion. A new average quantized watermark is realized, firstly, watermark signal is scrambled by Arnold transform, secondly, audio signal blocked N blocks and analyzed by discrete Fourier transform (DFT), average operations is processed on the energy signal, thirdly, it quantizes the average signal and embeds scrambled watermark. Experimental results show that this algorithm has strong robustness, and can counteract attacking.