The frequency analysis of the psychoacoustic model can give the inaudible , redundan elements of audio . in this paper we employ it to guarantee the inaudibility of the embedded watermark 心理声学模型的时频域分析主要用于音频数据的压缩编码,这里主要考虑将其用于音频水印算法中去,以便从理论的角度来考虑嵌入水印的不可感知性。