A method of automatic pitch segmentation for pitch-synchronous speech processing

Bibliographic Information

Other Title
  • ピッチ同期音声処理のためのピッチ区間自動区分化の一手法

Search this article

Abstract

This paper presents a new method for automatic pitch extraction and segmentation of voiced speech.Primary pitch extraction is performed at suitable portion within a voiced segment by autocorrelation analysis with multi window lengths.Pitch segmentation is done based on peak picking of sinusoidal-like waveform obtained by low-pass filtering with cut-off frequency which is little higher than the average pitch frequency estimated frame by frame referring to the result of primary pitch extraction. Experimental results of the automatic pitch extraction for many speech samples show that it is useful for speech rate or pitch period conversion by pitch synchronous speech processing for any kind of speech.

Journal

Citations (7)*help

See more

Details 詳細情報について

  • CRID
    1571698602294106624
  • NII Article ID
    110003296668
  • NII Book ID
    AN10013221
  • Text Lang
    ja
  • Data Source
    • CiNii Articles

Report a problem

Back to top