基于谐波提取的欠定语音盲分离方法

摘要

The existing underdetermined speech blind source separation (BSS)methods can hardly concurrently possess high efficiency and high performance.To solve this problem,a harmonics extraction based underdetermined speech BSS al-gorithm is proposed in this paper.Firstly,introduce the spectrum correction technique to extract the harmonic components from the mixtures’short time Fourier transform (STFT);Secondly,apply a phase-coherence criterion on these harmonic components to identify the single source components;Thirdly,employ the adaptive K-means clustering on these refined sin-gle source patterns to estimate the source number and the mixing matrix;Finally,combining the subspace projection algo-rithm with this estimated matrix yields the source recovery.Specifically,the combination of harmonics extraction and sin-gle-source component identification ensures that the mixing matrix can be accurately estimated in low complexity.Simula-tion demonstrates that,compared to the original subspace projection algorithm,the proposed BSS method can acquire a higher recovery quality,which presents a potential application in other harmonic related fields.%现有的欠定语音信号盲分离算法往往不能同时兼顾分离性能及效率。针对此问题，本文提出一种基于谐波提取的欠定盲分离方法。首先，利用频谱校正从混合信号的短时傅里叶变换中提取谐波参数，其次利用相位一致性准则甄别这些参数的单源属性，进而用自适应 K-均值方法对单源模式做聚类而获得源数估计和混合矩阵估计，最后再用子空间投影法恢复源信号。其中谐波提取和单源参数筛选可保证低复杂度地精确估计出混合矩阵。仿真实验表明，相比于原始子空间投影算法，本文方法可获得更高的信号恢复质量，且在谐波相关领域也具有潜在应用价值。