【摘要】 针对以往单通道噪声和浊语音分离算法的不足,改进了谐波组织算法。算法利用载波包络能量比将时频单元分为确定和非确定。提取基频作为组织线索。组织阶段分别使用谐波原理和最小幅度原理对确定时频单元组织;使用改进包络自相关函数度量幅度调制率对非确定时频单元组织。对比以往算法的处理结果,改进算法平均信噪比(SNR)提高0.96 dB。通过对谐波组织规则的改进,提高了分离性能。更多还原

【Abstract】 An enhanced algorithm based on elaborate harmonic grouping rules for monaural voiced speech segregation was proposed.The algorithm classified the time-frequency(T-F) units into resolved and unresolved by carrier- to-envelope energy ratio.Then pitch is extracted as grouping cue.During the grouping stage,the resolved T-F units were grouped according to harmonicity principle and minimum amplitude principle.And for the unresolved T-F units grouping,"enhanced" envelope autocorrelation function was employed to detect the amplitude modulation rate.The signal to noise ratio was improved by 0.96 dB compared with that of previous method.The proposed algorithm improved the performance of separation by elaborated harmonic rules.更多还原