含噪语音的基音周期提取算法

Pitch detection algorithm for noisy speech

背景噪声对语音信号的基音周期提取有很大的影响.结合能够对背景噪声进行有效降噪的EMD(Empirical Mode Decomposition)软阈值法,提出了一种噪声背景下语音基音周期提取的方法.首先将EMD软阈值降噪法和中心削波法相结合,对信号进行预处理,然后利用语音信号波形本身的信息,进行波形之间的匹配估计,从而获得准确的语音信号基音周期.仿真实验表明:与传统的基音周期提取方法相比较,所提方法在基音周期提取的准确度方面有很大的提高.

The background noise has a great influence on the accurate extraction of pitch period. For the EMD soft threshold denoising method can be effective against the background noise, one novel method is proposed for pitch period extraction of noisy speech. Combined with the EMD softthreshold denoising method and center clipping, this method preprocesses the speech signal, then employs the signal waveform in time donation for matching estimation, in order to acquire the accurate pitch period. The implementing simulation results demonstrate that, the proposed method has a higher accuracy compared with traditional methods.