文章分析了语音的谐波加噪声模型,给出了基于谐波加噪声模型的男/女声转换算法流程,利用高斯混合模型(GMM)分别实现了谐波分量和噪声分量特征参数的转换。文中利用Matlab进行了仿真实验,验证了算法的有效性,对转换后的语音质量进行了ABX测试,并计算其谱失真进行客观评价,测试结果表明文中所述算法,较好的实现了语音转换。 In this paper, the harmonic and noise model of speech is analyzed, and the flow of male / female conversion algorithm based on harmonic and noise model is given. The conversion of the characteristic parameters of harmonic components and noise components is achieved by Gaussian mixture model (GMM) respectively. In this paper, Matlab was used to simulate the experiment to verify the effectiveness of the algorithm, the speech quality of the converted ABX test, and calculate the spectral distortion of the objective evaluation, the test results show that the algorithm described in the article, the better realization of the voice conversion .