隠れマルコフモデルに基づいた歌声合成システム  [in Japanese] A Singing Voice Synthesis System Based on Hidden Markov Model  [in Japanese]

Access this Article

Search this Article

Author(s)

Abstract

隠れマルコフモデルに基づく音声合成方式を歌声合成に拡張することにより構築した歌声合成システムについて述べる.本システムでは,歌い手の声の質と基本周波数パターンに関する特徴をモデル化するため,スペクトルと基本周波数パターンをHMMにより同時にモデル化している.特に,自然な歌声を合成するうえで重要な要素となる音符の音階や音長の基本周波数パターンへの影響を精度良くモデル化するため,楽譜から得られる音階と音長を考慮したコンテキスト依存モデルを構築している.これらのモデルに対して決定木によるコンテキストクラスタリングを行うことで,未知の楽曲からの歌声合成が可能となっている.実験から,歌い手の特徴を再現し歌声の合成が可能であることを示す.We describe a singing voice synthesis system by applying HMM-basedspeech synthesis technique.In this system, a sequence of spectrum and F0 are modeledsimultaneously in a unified framework of HMM, and context dependentHMMs are constructed by taking account of contextual factors thataffects singing voice.In addition, the distributions for spectral and F0 parameter areclustered independently by using a decision-tree based contextclustering technique.Synthetic singing voice is generated from HMMs themselves by usingparameter generation algorithm. In the experiments, we confirmed that smooth and natural-soundingsinging voice is synthesised. It is also maintains the characteristicsand personality of the donor of the singing voice data for HMMtraining.

We describe a singing voice synthesis system by applying HMM-based speech synthesis technique. In this system, a sequence of spectrum and FO are modeled simultaneously in a unified framework of HMM, and context dependent HMMs are constructed by taking account of contextual factors that affects singing voice. In addition, the distributions for spectral and FO parameter are clustered independently by using a decision-tree based context clustering technique. Synthetic singing voice is generated from HMMs themselves by using parameter generation algorithm. In the experiments, we confirmed that smooth and natural-sounding singing voice is synthesised. It is also maintains the characteristics and personality of the donor of the singing voice data for HMM training.

Journal

  • IPSJ journal

    IPSJ journal 45(3), 719-727, 2004-03-15

    Information Processing Society of Japan (IPSJ)

References:  16

Cited by:  19

Codes

  • NII Article ID (NAID)
    110002712123
  • NII NACSIS-CAT ID (NCID)
    AN00116647
  • Text Lang
    JPN
  • Article Type
    Journal Article
  • ISSN
    1882-7764
  • NDL Article ID
    6885035
  • NDL Source Classification
    ZM13(科学技術--科学技術一般--データ処理・計算機)
  • NDL Call No.
    Z14-741
  • Data Source
    CJP  CJPref  NDL  NII-ELS  IPSJ 
Page Top