A Study on Performance of Pitch Extraction by BPFP Method under Pitch Variation, Wave Disturbance and Noise

スポンサーリンク

概要

• 論文の詳細を見る

• Pitch extraction of speech signals is one of the most important subjects in speech analysis and synthesis. In this work, we study two pitch extraction methods using the bank of bandpass filter pairs (BPFP), and using neural networks combined with the bank of BPFP (BPFP-NN), in performance with respect, to variations of pitch period, disturbance of waveforms for each pitch period and addition of white noise, as compared with cepstrum and modified autocorrelation analysis. The results of experiments show that both BPFP and BPFP-NN methods are advantageous to other typical methods, and that our proposed rearrangement of center frequencies of BPFP at logarithmic intervals is effective for especially low pitch region.

• 社団法人電子情報通信学会の論文
• 2002-03-01

著者

• Funada Tetsuo

  Faculty Of Engineering Kanazawa University

• MIYABAYASHI Hideo

  Toyama National College of Maritime Technology

関連論文

• STUDY OF COMPUTER DISPLAY OF EPICARDIAL MAPPING : Electrocardiography : IInd Auditorium : Proceedings of the 43rd Annual Meeting of the Japanese Circulation Society, Tokyo, 1979
• Speech enhancement by ARUI model and a new type of adaptive filter with sequential processing
• A new method for extracting the glottal closure intervals of voiced speech
• A Study on Performance of Pitch Extraction by BPFP Method under Pitch Variation, Wave Disturbance and Noise

スポンサーリンク