ENCODER AND METHOD THEREOF, PROGRAM, AND RECORDING MEDIUM

• Main Authors: MORIYA TAKEHIRO, HARADA NOBORU, KAMAMOTO MASARU
• Format: Patent
• Language: English; Japanese
• Published: 11.07.2019
• Subjects: ACOUSTICS; MUSICAL INSTRUMENTS; PHYSICS; SPEECH ANALYSIS OR SYNTHESIS; SPEECH OR AUDIO CODING OR DECODING; SPEECH OR VOICE PROCESSING; SPEECH RECOGNITION

To provide a technique for accurately encoding and decoding a coefficient that can be converted into a linear prediction coefficient even for a frame in which the spectrum varies widely while suppressing an increase in encoding amount as a whole.SOLUTION: An encoder includes: a first encoding unit that encodes a coefficient that can be converted into a plurality of orders of linear prediction coefficients to acquire a first code; and a second encoding unit that, in a case (A-1) where an index Q corresponding to the largeness of a peak and a valley of a spectrum envelope corresponding to the coefficient that can be converted into a plurality of orders of linear prediction coefficient is more than or equal to a predetermined threshold value Th1 and/or in a case (B-1) where an index Q' corresponding to the smallness of a peak and a valley of the spectrum envelope is less than or equal to a predetermined threshold value Th1', at least encodes a quantization error of the first encoding unit to acquire a second code.SELECTED DRAWING: Figure 3 【課題】全体として符号量の増大を抑えつつ、スペクトルの変動が大きいフレームについても線形予測係数に変換可能な係数を精度良く符号化及び復号する技術を提供する。【解決手段】符号化装置は、複数次の線形予測係数に変換可能な係数を符号化して第一符号を得る第一符号化部と、(A-1)複数次の線形予測係数に変換可能な係数に対応するスペクトル包絡の山谷の大きさに対応する指標Ｑが所定の閾値Th1以上である場合、および／または、(B-1)スペクトル包絡の山谷の小ささに対応する指標Ｑ'が所定の閾値Th1'以下である場合、少なくとも第一符号化部の量子化誤差を符号化して第二符号を得る第二符号化部とを含む。【選択図】図３