Speech pitch period estimation using circular AMDF
Gang Xu, Liangrui Tang · 2004
Pitch period is a key parameter in speech compression, synthesis and recognition. The AMDF is often used to determine this parameter. Due to the disadvantage of the AMDF, however, it is easy to make the estimated pitch doubled or halved, and increase the complexity of the algorithm. In this paper, we propose a new function, circular AMDF (CAMDF), and describe a new pitch detection algorithm based on the CAMDF. The CAMDF conquer the defect of the AMDF, and bring us some useful features, which efficiently decrease not only the error estimation in pitch detection, also simplify the process and improve the precision of estimated pitch. Lots of experiments show that this algorithm can avoid pitch smoothing, reduce the delay of speech coding, and satisfy the requirement of real-time pitch detection, when the restriction on the error estimation is not too high.