中文摘要：

介绍了一种改进的抗噪性强的基音周期检测算法,该算法利用经典的自相关函数和平均幅度差函数,并将二者相结合。它继承了自相关函数的抗噪特性,并通过能量函数进一步加重基音峰值点,进而提高了在噪声环境下的提取准确度。与传统自相关方法相比,该算法很好地克服了基音半倍和双倍错误。实验证明：这种算法简单易行,又有很好的抗噪性能,在存在噪声的条件下,仍能比较精确地检测出基音周期。

英文摘要：

A modified noise-robust method for pitch period detection is presented. It combines autocorreetion energy function and magnitude difference energy function based on autoeorreetion function （ACF） and average magnitude difference function （AMDF）. The method inherits noiserobust performance of ACF and has enhanced the period peak by energy function, moreover improves the estimation accuracy in noise environment. It overcomes the halving and doubling errors compared with ACF. The experiment shows that the algorithm is simple and feasible, but has good performance in noise-robust. It can detect pitch period accurately without any procession with noise.