一种基于快速傅立叶变换的音频数字水印

提出了一种新的数字音频水印算法,首先对信号进行分段离散快速傅立叶变换(FFT),再根据人类听觉系统(HAS)模型计算出各段的纯音、噪音,以及各段的总体掩蔽阈值,通过改变大于总体遮蔽阈值的纯音能量值来嵌入水印信息。仿真实验结果表明本算法不仅能保持较好的音质,而且对诸如叠加噪声、有损压缩、低通滤波、重新采样等攻击都具有较好的鲁棒性。

Audio Digital Watermark Algorithm Based on Fast Fourier Transform

A new digital audio watermarking algorithm is presented. First, signals will be converted into segments by the FFT. Then, according to the Human Auditory system model, every segment's tones, noises and final masking threshold will be gotten. Therefore, the aim of embedding into the watermarking signal will be achieved through changing the value of the tones' energy which is bigger than the final masking threshold. Experimental results show that the proposed algorithm not only can keep better timbre but also can get better robust against various signal processing operations, such as noise addition, lossy compression, low pass filtering, re-quantization.