A semi-fragile speech watermarking algorithm based on speech model is researched, which is realized by quantization of line spectrum pair (LSP) parameters.

• 在语音产生模型的基础上研究了一种基于该模型的半脆弱性数字水印算法，通过对语音信号的线谱对(LSP)参数嵌入水印信息来实现。