A Correcting Method for Pitch Extraction Using Neural Networks (Special Section of Papers Selected from JTC-CSCC'93)

概要

論文の詳細を見る
Pitch frequency is a basic characteristic of human voice, and pitch extraction is one of the most important studies for speech recognition. This paper describes a simple but effective technique to obtain correct pitch frequency from candidates (pitch candidates) extracted by the short-range auto-correlation function. The correction is performed by a neural network in consideration of the time continuation that is realized by referring to pitch candidates at previous frames. Since the neural network is trained by the back-propagation algorithm with training data, it adapts to any speaker and obtains good correction without sensitive adjustment and tuning. The pitch extraction was performed for 3 male and 3 female announcers, and the proposed method improves the percentage of correct pitch from 58.65% to 89.19%.
社団法人電子情報通信学会の論文
1994-06-25