基于幅度压缩滤波的清浊音分类及基音估计

Voiced/Unvoiced Classification and Pitch Estimation Based on Amplitude Compression Filter

该文针对传统算法在实环境(不同噪声类型和信噪比)下容易发生清浊误判和基音估计错误问题,提出一种基于幅度压缩基音估计滤波(PEFAC)的清浊音分类及基音估计方法。首先,通过PEFAC削弱语音的低频噪声,提取出基音谐波;然后,采用基于对称平均幅度和函数的脉冲序列加权算法(SIM)确定谐波数目;最后,利用动态规划估计出基音,用基于3元素特征矢量的高斯混合模型对清浊音进行分类。仿真结果表明,在实环境下,所提方法能有效抑制清浊误判及基音估计错误现象的发生,性能优于传统方法。

A method of voiced/unvoiced classification and pitch estimation based on Pitch Estimation Filter with Amplitude Compression（PEFAC） is proposed in this paper. The method first attenuates strong noise components at the low frequencies based on PEFAC and extracts pitch harmonic from noisy speech in the log-frequency domain. Then, the harmonic number associated with the pitch harmonic is determined by Symmetric average magnitude sum function weighted Impulse-train Matching（SIM） scheme in time domain. A pitch tracking scheme using dynamic programming is applied to select the pitch candidates and a voiced speech probability is computed from the likelihood ratio of Gaussian Mixture Models（GMMs） classifiers based on 3-element feature vector. The simulated results show that the proposed method efficiently reduces voiced/unvoiced and pitch estimation error, and it is superior to some of the state-of-the–art method in the real environment.