Abstract: An approach of searching the audio for appropriate segments is propo sed in this paper. It computes the AMT(audio masking threshold) of every audio s egments, and compares it to the corresponding audio power spectrum, by which the appropriate segments are found out. This method is effective for short-time aud io frames and has been successfully applied into an audio watermarking project b ased on OFDM.

摘要： 针对音频信号信息隐藏，提出了一种基于人耳掩蔽效应寻找适当嵌入点的方法 。该方法在时域上对音频信号按一定大小进行分帧，计算每一帧若干个临界带宽的掩蔽阈值 ，然后通过与对应子带的能量进行比较来找到适当的嵌入点。当音频帧长度为数十毫秒数量 级时取得了良好的效果。这种技术已成功地应用于一种高度稳健的音频数据隐藏方案中。

关键词: 掩蔽效应, 音频信号处理, 信息隐藏

CLC Number: 

• TN912.3