JP2004302259A - Hierarchical encoding method and hierarchical decoding method for audio signal - Google Patents

Abstract

To perform high-quality encoding at a low bit rate.
An n-th layer residual signal is input from an input terminal, and supplied to a time domain coding unit and a target signal generation unit. The time domain encoding unit 203 encodes the n-th layer residual signal in the time domain using the n-th layer residual signal and an input signal input from the input terminal 202 to generate an encoded code. I do. The target signal generation unit 204 generates an input signal of the frequency domain coding unit 205 using the input signal input from the input terminal 201 and the coded code obtained by the time domain coding unit 203. The frequency domain coding unit 205 generates a coded code by performing coding in the frequency domain using the signal generated by the target signal generation unit 204 and the input signal input from the input terminal 202. Output to
[Selection] Fig. 2

Description

ここで、Ｆｓ（ｎ）は第ｎレイヤの信号のサンプリング周波数を表す。本実施の形態によれば、複数のサンプリング周波数に対応した符号化を行うことが可能となる。
【００７２】
入力端子１０１から、サンプリング周波数Ｆｓ（３）の音響信号が入力されＤＳ１部１０２に与えられる。
【００７３】
ＤＳ１部１０２は、入力音響信号をダウンサンプリングし、この入力音響信号のサンプリング周波数をＦｓ（３）からＦｓ（１）に下げる。そして、ＤＳ１部１０２は、サンプリング周波数Ｆｓ（１）の入力信号を第１レイヤ符号化部１０３に出力する。
【００７４】
第１レイヤ符号化部１０３は、過去に生成した駆動音源信号を内部状態として保持している適応符号帳を有し、適応符号帳を用いることで周期性の強い信号を効率的に符号化することができる。第１レイヤ符号化部１０３は、入力音響信号と符号化後に生成される復号信号との間の聴感的な歪が最小となるように第１符号化コードを決定する。第１レイヤ符号化部１０３に適用される代表的な方法として符号励信線形予測法（ＣＥＬＰ）がある。
【００７５】
そして、第１レイヤ符号化部１０３は、得られた第１符号化コードを第１レイヤ復号化部１０４及び多重化部１１８に出力する。第１レイヤ復号化部１０４は、第１符号化コードを用いて第１レイヤ復号信号を生成し、この第１レイヤ復号信号をＵＳ１部１０５に出力する。
【００７６】
ＵＳ１部１０５は、第１レイヤ復号信号をアップサンプリングし、サンプリング周波数をＦｓ（１）からＦｓ（２）に上げる。そして、ＵＳ１部１０５は、サンプリング周波数Ｆｓ（２）の第１レイヤ復号信号を減算器１０８と加算器１１１に出力する。
【００７７】
次に、入力端子１０１から入力される音響信号がＤＳ２部１０６に与えられる。ＤＳ２部１０６は、入力音響信号をダウンサンプリングし、この入力音響信号のサンプリング周波数をＦｓ（３）からＦｓ（２）に下げる。そして、ＤＳ２部１０６は、サンプリング周波数Ｆｓ（２）の入力信号を遅延器１０７に出力する。
【００７８】
遅延器１０７は、入力端子１０１から入力される音響信号を所定の時間長だけ遅延して減算器１０８に出力する。すなわち、ＤＳ１部１０２、第１レイヤ符号化部１０３、第１レイヤ復号化部１０４、ＵＳ１部１０５およびＤＳ２部１０６にて生じる遅延を補正する役割を持つ。
【００７９】
減算器１０８は、遅延器１０７の出力信号と前述の第１レイヤ復号信号との差をとり第２レイヤ残差信号を生成する。そして、減算器１０８は、第２レイヤ残差信号を第２レイヤ符号化部１０９に出力する。
【００８０】
第２レイヤ符号化部１０９は、第２レイヤ残差信号を聴感的に品質改善が成されるように符号化を行い、第２符号化コードを決定する。そして、第２レイヤ符号化部１０９は、第２レイヤ復号化部１１０と第２符号化コードを多重化部１１８に出力する。第２レイヤ復号化部１１０は、第２符号化コードを用いて復号処理を行い、第２レイヤ復号残差信号を生成し、この第２レイヤ復号残差信号を加算器１１１に出力する。
【００８１】
加算器１１１は、第１レイヤ復号信号と第２レイヤ復号残差信号の和をとり、第２レイヤ復号信号を生成する。そして、加算器１１１は、この第２レイヤ復号信号をＵＳ２部１１２に出力する。
【００８２】
ＵＳ２部１１２は、第２レイヤ復号信号をアップサンプリングし、サンプリング周波数をＦｓ（２）からＦｓ（３）に上げる。そして、ＵＳ２部１１２は、サンプリング周波数Ｆｓ（３）の第１レイヤ復号信号を減算器１１４に出力する。
【００８３】
次に、遅延器１１３は、入力端子１０１から入力される音響信号を所定の時間長だけ遅延した後、この音響信号を減算器１１４に出力する。すなわち、遅延器１１３は、前段までの符号化部と復号化部で生じる遅延、具体的にはＤＳ１部１０２からＵＳ２部１１２までの信号処理で生じる遅延を補正する役割を持つ。
【００８４】
減算器１１４は、遅延器１１３の出力信号と前述の第２レイヤ復号信号との差をとり第３レイヤ残差信号を生成する。そして、減算器１１４は、この第３レイヤ残差信号を第３レイヤ符号化部１１５に出力する。
【００８５】
第３レイヤ符号化部１１５は、第３レイヤ残差信号を聴感的に品質改善が成されるように符号化して第３符号化コードを決定し、この第３符号化コードを多重化部１１８に出力する。
【００８６】
多重化部１１８は、第１符号化コード、第２符号化コードおよび第３符号化コードを所定の手段によって多重化を行い、符号化ビット列を生成する。そして、多重化部１１８は、この符号化ビット列を出力端子１１９より出力する。
【００８７】
次に、第２レイヤ以降の符号化の詳細について説明する。本実施の形態の階層符号化装置は、第２レイヤ以降の符号化において、入力音響信号と前段の符号化信号を復号化した信号との差分から残差信号を生成し、この残差信号を時間領域符号化部と周波数領域符号化部により符号化を行う点に特徴がある。
【００８８】
次に、第ｎレイヤ（２≦ｎ≦Ｎ）符号化部について説明を行う。図２は、本実施の形態の階層符号化装置の第ｎレイヤ（２≦ｎ≦Ｎ）符号化部の構成を示すブロック図である。
【００８９】
入力音響信号と第ｎレイヤ（２≦ｎ≦Ｎ）符号化信号を復号化した信号との差分である第ｎレイヤ残差信号には、上位層までの符号化ノイズとサンプリング周波数が高くなったことによる高周波成分とが含まれる。
【００９０】
第ｎレイヤ残差信号には時間領域で処理した場合に効率的に符号化できる成分と周波数領域で処理した場合に効率的に符号化できる成分とが混在している。そのため時間領域および周波数領域の２つの領域で符号化を行うことにより効率的な符号化が実現できるという効果が得られる。また、時間領域符号化部と周波数領域符号化部の両者には、入力信号がそれぞれ与えられている。この入力信号は、聴覚的に高品質な符号化を実現するために聴覚マスキングの算出などに利用される。以下、図２を用いて詳細な説明を行う。
【００９１】
入力端子２０１から第ｎレイヤ残差信号が入力され、時間領域符号化部２０３と目標信号生成部２０４に与えられる。時間領域符号化部２０３は、前記第ｎレイヤ残差信号と入力端子２０２より入力される入力信号を用いて、第ｎレイヤ残差信号を時間領域にて符号化を行い、符号化コードを生成する。そして、時間領域符号化部２０３は、符号化コードを目標信号生成部２０４と多重化部２０６に出力する。時間領域符号化部２０３の詳細については図３を用いて後述する。
【００９２】
次に、目標信号生成部２０４は、入力端子２０１より入力される入力信号と時間領域符号化部２０３で求められた符号化コードを使い、周波数領域符号化部２０５の入力信号を生成する。目標信号生成部２０４の詳細については図４を用いて後述する。
【００９３】
次に、周波数領域符号化部２０５は、目標信号生成部２０４で生成される信号と入力端子２０２より入力される入力信号を用いて、周波数領域にて符号化して符号化コードを生成し、多重化部２０６に出力する。周波数領域符号化部２０５の詳細については図５を用いて後述する。
【００９４】
以下、各ブロックの詳細について説明する。図３は、本実施の形態の階層符号化装置の時間領域符号化部の構成を示すブロック図である。図３の時間領域符号化部２０３は、入力端子３０１と、ＬＰＣ分析器３０２と、ＬＰＣ量子化器３０３と、ＬＰＣ復号器３０４と、聴感重みフィルタ３０５と、合成フィルタ３０６と、適応符号帳３０７と、雑音符号帳３０８と、乗算器３０９と、乗算器３１０と、ゲイン符号帳３１１と、加算器３１２と、減算器３１３と、探索器３１４と、多重化部３１５と、出力端子３１６とから主に構成される。
【００９５】
ＬＰＣ分析器３０２は、入力端子３０１から入力されたサンプリングレートＦｓ（ｎ）の音響信号からＬＰＣ係数を求める。このＬＰＣ係数は、聴感的な品質向上のために利用される係数である。ＬＰＣ分析器３０２は、このＬＰＣ係数を聴感重みフィルタ３０５とＬＰＣ量子化器３０３に出力する。
【００９６】
ＬＰＣ量子化器３０３は、ＬＰＣ係数をＬＳＦ係数などの量子化に適したパラメータに変換し、量子化を行う。そして、ＬＰＣ量子化器３０３は、この量子化で得られる符号化コードを多重化部３１５とＬＰＣ復号器３０４に出力する。
【００９７】
ＬＰＣ復号器３０４は、符号化コードから量子化後のＬＳＦ係数を算出し、このＬＳＦ係数をＬＰＣ係数に変換する。この処理により、量子化後のＬＰＣ係数が求められる。そして、ＬＰＣ復号器３０４は、この量子化後のＬＰＣ係数を合成フィルタ３０６に出力する。
【００９８】
合成フィルタ３０６は、この量子化後のＬＰＣ係数を用いて適応ベクトル、適応ゲイン、雑音ベクトルおよび雑音ゲインの探索を行う。次に、適応ベクトル、適応ベクトルゲイン、雑音ベクトル、雑音ベクトルゲインの探索方法について説明する。
【００９９】
適応符号帳３０７は、過去に生成した駆動音源信号を内部状態として保持しており、この内部状態を所望のピッチ周期で繰り返すことにより適応ベクトルを生成する。ピッチ周期の取る範囲は６０Ｈｚ〜４００Ｈｚの間が適当である。また、雑音符号帳３０８は、あらかじめ記憶領域に格納されている雑音ベクトル、もしくは代数（ａｌｇｅｂｒａｉｃ）構造のように記憶領域を持たずにルールに従い生成される雑音ベクトルを出力する。
【０１００】
ゲイン符号帳３１１は、適応ベクトルに乗じられる適応ベクトルゲインを乗算器３０９に出力し、雑音ベクトルに乗じられる雑音ベクトルゲインを乗算器３１０に出力する。
【０１０１】
乗算器３０９は、適応ベクトルに適応ベクトルゲインを乗算して加算器３１２に出力する。乗算器３１０は、雑音ベクトルに雑音ベクトルゲインを乗算して加算器３１２に出力する。
【０１０２】
加算器３１２は、適応ベクトルゲインが乗じられた適応ベクトルと雑音ベクトルゲインが乗じられた雑音ベクトルとを加算して駆動音源信号を生成する。そして、加算器３１２は、この駆動音源信号を合成フィルタ３０６に出力する。
【０１０３】
合成フィルタ３０６は、駆動音源信号を合成フィルタに通して合成信号を生成し、この合成信号を減算器３１３に出力する。
【０１０４】
減算器３１３は、入力端子３１７から入力される第ｎレイヤ予測残差信号から合成信号を減算し、聴感重みフィルタ３０５に減算後の信号を出力する。
【０１０５】
聴感重みフィルタ３０５は、ＬＰＣ分析器３０２で求められたＬＰＣ係数を基に減算器３１３で求められる信号に重み付けを行う。これは、量子化歪のスペクトルを入力信号のスペクトル包絡にマスクされるようスペクトル整形を行うことを目的として行われる。
【０１０６】
探索器３１４では、減算後の信号から定義される歪が最小となる適応ベクトル、適応ベクトルゲイン、雑音ベクトル、雑音ベクトルゲインの組み合わせを効率よく探索し、それら符号化コードを多重化部３１５に送る。
【０１０７】
探索器３１４では、以下の式（２）または式（３）で定義される歪を最小とする符号化コードｉ、ｊ、ｍまたはｉ、ｊ、ｍ、ｎを決定してそれらを多重化部３１５に送ることになる。
【数２】

【数３】

ここで、ｔ（ｋ）は第ｎレイヤ残差信号、ｑ_ｉ（ｋ）は第ｉ番目の適応ベクトル、ｃ_ｊ（ｋ）は第ｊ番目の雑音ベクトル、βとγはそれぞれ適応ベクトルゲインと雑音ベクトルゲインを表す。
【０１０８】
式（２）と式（３）とではゲイン符号帳の構成が異なる。式（２）の場合、ゲイン符号帳は適応ベクトルゲインβ_ｍと雑音ベクトルゲインγ_ｍを要素として持つベクトルとして表されており、ベクトルを特定するための符号化コードｍが決定されることになる。式（３）の場合、ゲイン符号帳は適応ベクトルゲインβ_ｍと雑音ベクトルゲインγ_ｎをそれぞれ独立に有しており、それぞれの符号化コードｍ、ｎが独立に決定されることになる。また、ｈ（ｌ）は聴感重みフィルタのインパルス応答を表す。

は畳み込みを表す演算子である。
【０１０９】
全ての符号化コードが決定された後に、多重化部３１５は、符号化コードを一つにまとめて出力端子３１６より出力する。そして、次のフレーム（もしくはサブフレーム）での復号化処理に備えて、選択された適応ベクトル、適応ベクトルゲイン、雑音ベクトル、雑音ベクトルゲインを用いて表される駆動音源信号を用いて適応符号帳の内部状態を更新する。
【０１１０】
次に、目標信号生成部２０４の詳細につい説明する。図４は、本実施の形態の階層符号化装置の目標信号生成部の構成を示すブロック図である。図４の目標信号生成部２０４は、入力端子４０１と、入力端子４０２と、復号部４０３と、遅延器４０４と、減算器４０５と、出力端子４０６とから主に構成される。
【０１１１】
入力端子４０１から時間領域符号化部２０３で得られる符号化コードが入力される。復号部４０３は、この符号化コードの情報に従い復号信号を生成する。
【０１１２】
遅延器４０４は、入力端子４０２から入力される第２レイヤ残差信号に時間領域符号化部２０３および復号部４０３で生じる遅延を補正するように遅延を与えた後、減算器４０５に出力する。
【０１１３】
減算器４０５は、遅延器４０４の出力信号から復号部４０３で得られる復号信号を減算して周波数領域符号化部２０５の目標信号を生成し、減算した信号を出力端子４０６から出力する。
【０１１４】
次に、周波数領域符号化部２０５の詳細について説明する。図５は、本実施の形態の階層符号化装置の周波数領域符号化部の構成を示すブロック図である。図５の周波数領域符号化部２０５は、入力端子５０１と、入力端子５０２と、周波数領域変換部５０３と、聴覚マスキング算出部５０４と、量子化部５０５と、出力端子５０６とから主に構成される。
【０１１５】
入力端子５０１から周波数領域符号化部２０５に入力される信号は、目標信号生成部２０４で求められた目標信号である。
【０１１６】
周波数領域変換部５０３は、目標信号に分析窓を乗じた後に周波数変換が行われ、この周波数変換で得られる変換係数が量子化部５０５に出力される。ここでの周波数変換の方法としては、変形離散コサイン変換（ＭＤＣＴ）や離散フーリエ変換（ＤＦＴ）などを用いることができる。
【０１１７】
入力端子５０２からはサンプリング周波数Ｆｓ（ｎ）の音響信号が与えられ、聴覚マスキング算出部５０４に入力される。聴覚マスキング算出部５０４は、人間には知覚されないノイズパワーの閾値を表す聴覚マスキングを算出し、量子化部５０５に聴覚マスキングを出力する。
【０１１８】
量子化部５０５は、聴覚マスキングを利用して周波数領域変換部５０３で求められた変換係数を量子化し、そのとき得られる符号化コードを出力端子５０６より出力する。
【０１１９】
次に聴覚マスキングの算出法を、図６を用いて詳細に説明する。図６は、本実施の形態の階層符号化装置の聴覚マスキング算出部の構成を示すブロック図である。人間の聴覚特性には、ある信号が与えられたとき、その信号の周波数の近傍に位置する信号が聞こえ難くなるというマスキング効果がある。この特性を利用して、入力信号のスペクトルから聴覚マスキングを算出し、量子化歪をこのマスキング値以下になるように変換係数の量子化を行うことにより、少ないビットレートで効率よく変換係数を量子化することができる。
【０１２０】
入力端子６０１から入力信号が与えられ、周波数変換部６０２にて周波数領域への変換が行われ変換係数が算出される。周波数領域への変換の方法として、前述のように変形離散コサイン変換（ＭＤＣＴ）や離散フーリエ変換（ＤＦＴ）などを用いることが可能である。ここでは、ＤＦＴを用いる場合について説明することとし、ＤＦＴにより求められたフーリエ係数を｛Ｒｅ（ｍ），Ｉｍ（ｍ）｝と表すものとする。
【０１２１】
図６において、周波数変換部６０２は、遅延器１０７から出力された入力信号をフーリエ変換し、フーリエ係数｛Ｒｅ（ｍ），Ｉｍ（ｍ）｝を算出する。ここでｍは周波数を表す。
【０１２２】
バークスペクトル算出部６０３は、以下の式（４）を用いてバークスペクトルＢ（ｋ）を算出する。
【数４】

ここで、Ｐ（ｍ）はパワースペクトルを表し、以下の式（５）より求められる。
【数５】

また、ｋはバークスペクトルの番号に対応し、ＦＬ（ｋ）、ＦＨ（ｋ）はそれぞれ第ｋバークスペクトルの最低周波数（Ｈｚ）、最高周波数（Ｈｚ）を表す。バークスペクトルＢ（ｋ）はバークスケール上で等間隔に帯域分割されたときのスペクトル強度を表す。ヘルツスケールをｆ、バークスケールをＢと表したとき、ヘルツスケールとバークスケールの関係は以下の式（６）で表される。
【数６】

【０１２３】
スプレッド関数畳み込み部６０４は、以下に示す式（７）を用いてバークスペクトルＢ（ｋ）にスプレッド関数ＳＦ（ｋ）を畳み込み、Ｃ（ｋ）を算出する。
【数７】

【０１２４】
トーナリティ算出部６０５は、以下の式（８）を用い、パワースペクトルＰ（ｍ）から各バークスペクトルのスペクトル平坦度ＳＦＭ（ｋ）を求める。
【数８】

ここで、μｇ（ｋ）は第ｋバークスペクトルの幾何平均、μａ（ｋ）は第ｋバークスペクトルの算術平均を表す。そして、トーナリティ算出部６０５は、以下の式（９）を用いてスペクトル平坦度ＳＦＭ（ｋ）のデシベル値ＳＦＭｄＢ（ｋ）からトーナリティ係数α（ｋ）を算出する。
【数９】

【０１２５】
聴覚マスキング算出部６０６は、以下の式（１０）を用いてトーナリティ算出部６０５で算出したトーナリティ係数α（ｋ）から各バークスケールのオフセットＯ（ｋ）を求める。
【数１０】

【０１２６】
そして、聴覚マスキング算出部６０６は、以下の式（１１）を用いてスプレッド関数畳み込み部６０４で求めたＣ（ｋ）からオフセットＯ（ｋ）を減算して聴覚マスキングＴ（ｋ）を算出する。
【数１１】

ここで、Ｔ_ｑ（ｋ）は絶対閾値を表す。絶対閾値は、人間の聴覚特性として観測される聴覚マスキングの最小値を表す。そして、聴覚マスキング算出部６０６は、バークスケールで表される聴覚マスキングＴ（ｋ）をヘルツスケールＭ（ｍ）に変換して出力する。
【０１２７】
このように、本実施の形態の階層符号化装置によれば、上位レイヤの符号化結果を復号化した信号と入力音響信号との差分を時間領域と周波数領域で符号化することにより、周期性のある信号を時間領域の符号化し、周期性のない信号を周波数領域の符号化することができ、符号化して低ビットレートで高品質な符号化を行うことができる。
【０１２８】
特に、本実施の形態の階層符号化装置によれば、第２レイヤ以下の符号化において、上位レイヤの符号化結果を復号化した信号と入力音響信号との差分を時間領域で符号化し、前記差分信号と時間領域での符号化した信号を復号した復号信号との差分を周波数領域で符号化することにより、周期性のある信号を時間領域の符号化し、周期性のない信号を周波数領域の符号化することができ、符号化して低ビットレートで高品質な符号化を行うことができる。
【０１２９】
また、本実施の形態の階層符号化装置によれば、下位レイヤで符号化する信号のサンプリング周波数を上位レイヤで符号化する信号のサンプリング周波数より高くすることにより、様々なサンプリング周波数に対応させて入力信号を符号化することができる。
【０１３０】
（実施の形態２）
本実施の形態では、実施の形態１の階層符号化装置で符号化された信号を復号する例について説明する。本実施の形態の特徴は、実施の形態１で説明された階層符号化法の符号化コードを復号することができ、その結果高品質な音響信号を復号することが可能になる点にある。
【０１３１】
図７は、本発明の実施の形態２に係る階層復号化装置の構成を示すブロック図である。図７の階層復号化装置７００は、入力端子７０１と、分離部７０２と、第１レイヤ復号化部７０３と、ＵＳ１部７０４と、加算器７０５と、第２レイヤ復号化部７０６と、ＵＳ２部７０７と、第３レイヤ復号化部７０８と、加算器７０９と出力端子７１０とから主に構成される。
【０１３２】
入力端子７０１から図１の階層符号化装置にて符号化された符号化ビット列が入力される。
【０１３３】
分離部７０２は、符号化ビット列を分離し、第１レイヤ符号化で得られる第１符号化コード、第２レイヤ符号化で得られる第２符号化コードおよび第３レイヤ符号化で得られる第３符号化コードを生成する。そして、分離部７０２は、第１符号化コードを第１レイヤ復号化部７０３に出力し、第２符号化コードを第２レイヤ復号化部７０６に出力し、第３符号化コードを第３レイヤ復号化部７０８に出力する。
【０１３４】
第１レイヤ復号化部７０３は、分離部７０２で得られた第１符号化コードを用いて復号処理を行い、第１レイヤ復号信号を生成する。
【０１３５】
ＵＳ１部７０４は、第１レイヤ復号信号をアップサンプリングし、サンプリング周波数をＦｓ（１）からＦｓ（２）に上げる。そして、ＵＳ１部７０４は、サンプリング周波数Ｆｓ（２）の第１レイヤ復号信号を加算器７０５に出力する。
【０１３６】
次に、第２レイヤ復号化部７０６は、分離部７０２で得られた第２符号化コードを用いて復号処理を行い、第２レイヤ復号残差信号を生成する。加算器７０５では、前述の第１レイヤ復号信号と第２レイヤ復号残差信号とを加算し、第２レイヤ復号信号を生成する。そして、加算器７０５は、第２レイヤ復号信号をＵＳ２部７０７に出力する。
【０１３７】
ＵＳ２部７０７は、第２レイヤ復号信号をアップサンプリングし、サンプリング周波数をＦｓ（２）からＦｓ（３）に上げる。そして、ＵＳ２部７０７は、サンプリング周波数Ｆｓ（３）の第１レイヤ復号信号を加算器７０９に出力する。
【０１３８】
次に、第３レイヤ復号化部７０８は、分離部７０２で得られた第３符号化コードを用いて復号処理を行い、第３レイヤ復号残差信号を生成する。加算器７０９は、前述の第２レイヤ復号信号と第３レイヤ復号残差信号とを加算し、第３レイヤ復号信号を生成する。加算器７０９は、第３レイヤ復号信号を出力端子７１０に出力する。
【０１３９】
次に、第ｎレイヤ（２≦ｎ≦Ｎ）復号化部について説明を行う。図８は、本実施の形態の階層復号化装置の第２レイヤ以降の復号化部の構成を示すブロック図である。
【０１４０】
入力端子８０１より第ｎレイヤ（２≦ｎ≦Ｎ）符号化コードが入力される。分離部８０２は、第ｎレイヤ（２≦ｎ≦Ｎ）符号化コードを時間領域符号化コードと周波数領域符号化コードに分離する。そして、分離部８０２は、時間領域符号化コードを時間領域復号化部８０３に出力し、周波数領域符号化コードを周波数領域復号化部８０４に出力する。
【０１４１】
時間領域復号化部８０３は、時間領域符号化コードを用いて時間領域復号信号を生成し、時間領域復号信号を加算器８０５に出力する。時間領域復号化部８０３の詳細については図９を用いて後述する。
【０１４２】
同様に、周波数領域復号化部８０４は、周波数領域符号化コードを用いて周波数領域復号信号を生成し、加算器８０５に出力する。周波数領域復号化部８０４の詳細については図１０を用いて後述する。加算器８０５は、時間領域復号信号と周波数領域復号信号との加算を行い、出力端子８０６より出力する。
【０１４３】
次に、図９を用いて時間領域復号化部８０３の説明を行う。図９は、本実施の形態の階層復号化装置の時間領域復号化部の構成を示すブロック図である。
【０１４４】
図９において、分離部９０２は、入力端子９０１より入力される時間領域符号化コードから符号化コードを分離し、適応符号帳９０３、雑音符号帳９０４、ゲイン符号帳９０５、及びＬＰＣ復号器９０９にそれぞれ出力する。ＬＰＣ復号器９０９は、与えられる符号化コードを用いてＬＰＣ係数を復号し、合成フィルタ９１０に出力する。
【０１４５】
次に、適応符号帳９０３、雑音符号帳９０４およびゲイン符号帳９０５は、符号化コードを利用してそれぞれ適応ベクトルｑ（ｋ）、雑音ベクトルｃ（ｋ）、適応ベクトルゲインβ_ｑおよび雑音ベクトルゲインγ_ｑをそれぞれ復号する。
【０１４６】
乗算器９０６は、適応ベクトルに適応ベクトルゲインを乗じて加算器９０８に出力する。同様に、乗算器９０７は、雑音ベクトルに雑音ベクトルゲインを乗じて加算器９０８に出力する。加算器９０８は、乗算後の適応ベクトルと雑音ベクトルとを加算して駆動音源信号を生成する。駆動音源信号をｅｘ（ｋ）と表すと、駆動音源信号ｅｘ（ｋ）は次の式（１２）のように求められる。
【数１２】

【０１４７】
次に、復号されたＬＰＣ係数と駆動音源信号ｅｘ（ｋ）を用いて合成フィルタ９１０にて合成信号ｓｙｎ（ｋ）を次の式（１３）に従い生成する。
【数１３】

ここで、α_ｑは復号されたＬＰＣ係数、ＮＰはＬＰＣ係数の次数を表す。このように復号された復号信号ｓｙｎ（ｎ）は出力端子９１１より出力される。上記復号化処理が終了した後に、次のフレーム（もしくはサブフレーム）での復号化処理に備えて、適応符号帳の内部状態を最新の駆動音源信号を用いて更新する。
【０１４８】
次に図１０を用いて周波数領域復号化部８０４の説明を行う。図１０は、本実施の形態の階層復号化装置の周波数領域復号化部の構成を示すブロック図である。変換係数復号化部１００２は、入力端子１００１から入力される周波数領域符号化コードから量子化された変換係数を復号する。次に時間領域変換部１００３は、変換係数復号化部１００２から得られる変換係数に時間領域変換処理を施し、時間領域の信号を生成する。時間領域の信号にはフレーム（またはサブフレーム）間の不連続が生じないように重ね合わせ加算などの処理が施される。そして、時間領域変換部１００３は、この出力信号を出力端子１００４より出力する。
【０１４９】
このように、本実施の形態の階層復号化装置によれば、上位レイヤの符号化結果を復号化した信号と入力音響信号との差分を時間領域と周波数領域で符号化した符号化信号を、周期性のある信号を時間領域と、周期性のない信号を周波数領域とで復号することにより、低ビットレートで高品質な符号化及び復号化を行うことができる。
【０１５０】
また、本実施の形態の階層復号化装置によれば、下位レイヤで復号する信号のサンプリング周波数を上位レイヤで復号する信号のサンプリング周波数より高くすることにより、様々なサンプリング周波数に対応させて信号を符号化した信号を復号することができる。
【０１５１】
（実施の形態３）
図１１は、本発明の実施の形態３に係る階層符号化装置の構成を示すブロック図である。本実施の形態では、第ｎレイヤ（２≦ｎ≦Ｎ）符号化部が時間領域符号化部と周波数領域符号化部で構成される音響信号符号化方式において、上位レイヤにて求められたピッチ周期を利用して符号化を行う時間領域符号化部を有する点に特徴がある。
【０１５２】
本実施の形態によれば、上位レイヤで求めたピッチ周期を利用することにより、時間領域符号化部のピッチ周期の符号化をより効率的に行うことが可能となり、その結果として低ビットレートで高品質に符号化を行うことができる。図１１において、図２と同じ名称を持つ構成要素は同一の機能を有するため、そのような構成要素についての詳細な説明は省略する。
【０１５３】
入力端子１１０８から、上位レイヤにて求められたピッチ周期Ｔが入力される。時間領域符号化部１１０３は、入力された上位レイヤのピッチ周期を利用して符号化を行う。この場合の時間領域符号化部１１０３の構成を図１２に示す。図１２は、本実施の形態の階層符号化装置の時間領域符号化部の構成を示すブロック図である。図１２において、図３と同じ名称をもつ構成要素は同一の機能を有するため、そのような構成要素についての詳細な説明は省略する。
【０１５４】
入力端子１２１８から入力される下位レイヤのピッチ周期Ｔは探索候補決定部１２１９に与えられる。探索候補決定部１２１９は、上位レイヤのピッチ周期Ｔを基に適応符号帳１２０７に含まれる探索の対象となる適応ベクトルの候補を限定する。
【０１５５】
上記限定により、適応符号帳１２０７に含まれる全ての候補を探索の対象とする場合に比べ、この方法によれば探索の対象となる適応ベクトルの候補が少なくなるために当該レイヤのピッチ周期を表すための符号量を少なくできる。さらに適応符号帳の探索に必要な演算量が削減できるなどの効果が得られる。
【０１５６】
探索候補決定部１２１９は、上位レイヤのピッチ周期Ｔを使用して次の式（１４）で示される範囲に含まれるピッチ周期に対応する適応ベクトルを探索の対象とすることができる。ただし、上位レイヤのサンプリング周波数と当該レイヤのサンプリング周波数が異なる場合、当該レイヤのサンプリング周波数に適合するように上位レイヤのピッチ周期Ｔを修正して使用するものとする。
【数１４】

ここで、Ｔ（ｎ）は当該レイヤ（第ｎレイヤ）のピッチ周期を表す。Ｔ（ｍ）は上位レイヤのピッチ周期を表し、ｍの範囲は、１≦ｍ＜ｎと表される。また、ΔＴ１とΔＴ２はピッチ周期の範囲を決定する定数を表す。適応ベクトルの探索は式（１４）に含まれるピッチ周期Ｔ（ｎ）に対応する適応ベクトルについてのみ行われることになり、探索の結果、相対ピッチ周期ΔＴが決定され、この情報が符号化コードとして多重化部１２１５に与えられる。
【０１５７】
また、上位レイヤのピッチ周期が倍ピッチもしくは半ピッチになっている場合を考慮して、次に示す式（１５）に従い適応符号帳１２０７に含まれる適応ベクトルの探索候補を決定しても良い。
【数１５】

ここで、ｋはｋ＝｛…，１／４，１／３，１／２，１，２，３，４，…｝のように整数倍もしくは整数分の１を表す変数である。またΔＴ１（ｋ）およびΔＴ２（ｋ）と表記しているのは、ｋの値に依存してピッチ周期の探索範囲が異なることがある場合を示している。
【０１５８】
このように、本実施の形態の階層符号化装置によれば、第２レイヤより下位の時間領域符号化において、上位レイヤにて求められたピッチ周期を利用して適応符号帳の適応ベクトルから探索の対象となる適応ベクトルの候補を限定し、限定した適応ベクトルを用いて符号化を行うことにより、時間領域符号化のピッチ周期の符号化をより効率的に行うことができ、低ビットレートで高品質に符号化できる。
【０１５９】
（実施の形態４）
図１３は、本発明の実施の形態４に係る階層復号化装置の構成を示すブロック図である。本実施の形態では、第ｎレイヤ（２≦ｎ≦Ｎ）符号化部が時間領域符号化部と周波数領域符号化部で構成される階層符号化方式において、上位レイヤにて求められたピッチ周期を利用して符号化を行う時間領域符号化部により生成された符号化コードを復号できる点に特徴がある。
【０１６０】
本実施の形態によれば、上位レイヤで求めたピッチ周期を利用することにより、時間領域符号化部のピッチ周期の符号化がより効率的に行うことが可能となり、その結果として低ビットレートで高品質に符号化を行うことができる階層符号化方式の符号化コードを復号することにより、高品質な復号信号を得ることができるという効果が得られる。
【０１６１】
図１３において、図８と同じ名称を持つ構成要素は同一の機能を有するため、そのような構成要素についての詳細な説明は省略する。入力端子１３０７から、上位レイヤにて復号されたピッチ周期Ｔが入力され、時間領域復号化部１３０３に与えられる。
【０１６２】
時間領域符号化部１３０３は、入力された上位レイヤのピッチ周期を利用して復号化を行う。この時間領域復号化部１３０３の構成を図１４に示す。図１４は、本実施の形態の階層復号化装置の時間領域復号化部の構成を示すブロック図である。図１４において、図９と同じ名称をもつ構成要素は同一の機能を有するため、そのような構成要素についての詳細な説明は省略する。
【０１６３】
入力端子１４１２から入力される上位レイヤにて復号されたピッチ周期Ｔは適応ベクトル決定部１４１３に与えられる。さらに、分離部１４０２にて相対ピッチ周期ΔＴが復号され適応ベクトル決定部１４１３に与えられる。
【０１６４】
適応ベクトル決定部１４１３は、下位レイヤのピッチ周期Ｔおよび相対ピッチ周期ΔＴを用いて、次の式（１６）に従い当該レイヤのピッチ周期Ｔを算出する。
【数１６】

ここで、Ｔ（ｎ）は当該レイヤ（第ｎレイヤ）のピッチ周期を表し、Ｔ（ｍ）は上位レイヤ（１≦ｍ＜ｎ）のピッチ周期を表す。式（１５）に従い適応ベクトルの探索候補が決定されている場合には、当該レイヤのピッチ周期は次の式（１７）に従い算出される。
【数１７】

ここで、ｋはｋ＝｛…，１／４，１／３，１／２，１，２，３，４，…｝のように整数倍もしくは整数分の１を表す変数である。このようにして復号した当該レイヤのピッチ周期を適応符号帳１４０３に与える。適応符号帳１４０３では、復号したピッチ周期に対応した適応ベクトルを出力することになる。
【０１６５】
このように、本実施の形態の音声復号化装置によれば、第２レイヤより下位の時間領域復号化において、符号化側の上位レイヤにて求められたピッチ周期を利用して適応符号帳の適応ベクトルから復号化に用いる適応ベクトルの候補を限定して復号化を行うことにより、時間領域符号化のピッチ周期の符号化及び復号化をより効率的に行うことができ、低ビットレートで高品質に符号化できる。
【０１６６】
（実施の形態５）
実施の形態５では、実施の形態３の入力端子１１０８から入力されるパラメータが異なる例について説明する。実施の形態３では上位レイヤで求められたピッチ周期が入力されていたが、本実施の形態では上位レイヤにて求められたＬＰＣ係数が入力される。
【０１６７】
本実施の形態では、第ｎレイヤ（２≦ｎ≦Ｎ）符号化部が時間領域符号化部と周波数領域符号化部で構成される階層符号化方式において、上位レイヤにて求められたＬＰＣ係数を利用して符号化を行う時間領域符号化部を有する点に特徴がある。本実施の形態によれば、上位レイヤで求めたＬＰＣ係数を利用することにより、時間領域符号化部のＬＰＣ係数の符号化がより効率的に行うことが可能となり、その結果として低ビットレートで高品質に符号化を行うことができる。図１１において、図２と同じ名称を持つ構成要素は同一の機能を有するため、そのような構成要素についての詳細な説明は省略する。
【０１６８】
図１１において、入力端子１１０８から、上位レイヤにて求められたＬＰＣ係数が入力され、時間領域符号化部１１０３に与えられる。時間領域符号化部１１０３は、入力された下位レイヤのＬＰＣ係数を利用して符号化を行う。この場合の時間領域符号化部１１０３の構成を図１５に示す。図１５は、本実施の形態の階層符号化装置の時間領域符号化部の構成を示すブロック図である。図１５において、図３と同じ名称をもつ構成要素は同一の機能を有するため、そのような構成要素についての詳細な説明は省略する。
【０１６９】
入力端子１５１８から入力される上位レイヤのＬＰＣ係数は、ＬＰＣ量子化器１５０３に与えられる。ＬＰＣ量子化器１５０３は、ＬＰＣ分析器１５０２から与えられる当該レイヤのＬＰＣ係数を上位レイヤのＬＰＣ係数を利用して効率的に符号化を行う。ＬＰＣ量子化器１５０３の構成を、図１６を用いて説明する。図１６は、本実施の形態の階層符号化装置のＬＰＣ量子化器の構成を示すブロック図である。
【０１７０】
入力端子１６０９からここでは図示されないＬＰＣ分析器１５０２で求められた当該レイヤのＬＰＣ係数が入力される。当該レイヤのＬＰＣ係数を｛αｐ； ｐ＝１〜ＮＰ（ｎ）｝と表す。ここでＮＰ（ｎ）は当該レイヤ（第ｎレイヤ）のＬＰＣ係数の次数を表す。
【０１７１】
次に、ＬＳＦ変換部１６０６は、当該フレームのＬＰＣ係数をＬＳＦ係数に変換する。ＬＳＦ係数は、ＬＰＣ係数と相互に変換可能なパラメータで、フィルタの安定条件判定が容易、パラメータの補間特性が良い、スペクトル歪に対するパラメータの感度がほぼ一定などの利点があり、音声符号化の分野では広く利用されている。
【０１７２】
ここでＬＳＦ係数を｛Ｆｐ； ｐ＝１〜ＮＰ（ｎ）｝と表すと、ＬＳＦ係数は０〜１の間の値を取り、かつＦｐ＜Ｆｐ＋１の関係がある。同様に入力端子１６０１から入力される上位レイヤのＬＰＣ係数を｛βｐ； ｐ＝１〜ＮＰ（ｍ）｝と表す。ここでＮＰ（ｍ）は上位レイヤ（第ｍレイヤ、ｍ＜ｎ）のＬＰＣ係数の次数を表す。
【０１７３】
次に、ＬＳＦ変換部１６０２は、上位レイヤのＬＰＣ係数｛βｐ； ｐ＝１〜ＮＰ（ｍ）｝をＬＳＦ係数｛Ｇｐ； ｐ＝１〜ＮＰ（ｍ）｝に変換する。次に、修正部１６０３は、当該レイヤのサンプリング周波数に適合するように下位レイヤのＬＳＦ係数に定数を乗じる。この定数は、Ｆｓ（ｍ）／Ｆｓ（ｎ）で表される。
【０１７４】
加算器１６０５は、修正部１６０３から与えられる変換後の下位レイヤのＬＳＦ係数とデルタＬＳＦ符号帳１６０４に格納されているデルタＬＳＦベクトルとを加算する。減算器１６０７は、当該レイヤのＬＳＦ係数から加算器１６０５の出力ベクトルを減じ、その誤差信号を探索器１６０８に出力する。
【０１７５】
探索器１６０８は、前記誤差信号のエネルギーまたは聴感的に重み付けされたエネルギーを最小にするデルタＬＳＦ符号帳１６０４に格納されているデルタＬＳＦベクトルを効率的に探索し、そのインデックスを符号化コードとして出力端子１６１０より出力する。
【０１７６】
このように、本実施の形態の階層符号化装置によれば、第２レイヤより下位の時間領域符号化において、上位レイヤにて求められたＬＰＣ係数（またはＬＳＦ係数）と当該レイヤのＬＰＣ係数（またはＬＳＦ係数）を用いて最適なデルタＬＳＦベクトルを探索することにより、上位レイヤにて求められたＬＰＣ係数を考慮して最適なデルタＬＳＦベクトルを探索することができ、時間領域符号化をより効率的に行うことができ、低ビットレートで高品質に符号化できる。
【０１７７】
（実施の形態６）
実施の形態６では、実施の形態４の入力端子１３０７から入力されるパラメータが異なる例について説明する。実施の形態４では上位レイヤで求められたピッチ周期が入力されていたが、実施の形態６では上位レイヤにて求められたＬＰＣ係数が入力される。
【０１７８】
本実施の形態では、第ｎレイヤ（２≦ｎ≦Ｎ）復号化部が時間領域復号化部と周波数領域復号化部で構成される音響信号復号化方式において、上位レイヤにて復号されたＬＰＣ係数を利用して当該レイヤのＬＰＣ係数の復号化を行う時間領域復号化部を有する点に特徴がある。本実施の形態によれば、下位レイヤで復号化されたＬＰＣ係数を利用することにより、ＬＰＣ係数の符号化を効率的に行う時間領域符号化部の符号化コードを復号することが可能となり、その結果として低ビットレートで高品質な復号信号を生成することができる。図１３において、図８と同じ名称を持つ構成要素は同一の機能を有するため、そのような構成要素についての詳細な説明は省略する。
【０１７９】
入力端子１３０７から、上位レイヤにて復号されたＬＰＣ係数が入力され、時間領域復号化部１３０３に与えられる。時間領域復号化部１３０３は、入力された上位レイヤのＬＰＣ係数を利用して復号化を行う。この場合の時間領域復号化部１３０３の構成を図１７に示す。図１７は、本実施の形態の階層復号化装置の時間領域復号化部の構成を示すブロック図である。図１７において、図９と同じ名称をもつ構成要素は同一の機能を有するため、そのような構成要素についての詳細な説明は省略する。
【０１８０】
入力端子１７１２から入力される上位レイヤのＬＰＣ係数はＬＰＣ復号器１７０９に与えられる。ＬＰＣ復号器１７０９は、当該レイヤのＬＰＣ係数を上位レイヤのＬＰＣ係数を利用して復号する。ＬＰＣ復号器１７０９の構成を、図１８を用いて説明する。図１８は、本実施の形態の階層復号化装置のＬＰＣ復号器の構成を示すブロック図である。
【０１８１】
入力端子１８０１から上位レイヤのＬＰＣ係数｛βｐ； ｐ＝１〜ＮＰ（ｍ）｝が入力される。ＬＳＦ変換部１８０７は、上位レイヤのＬＳＦ係数｛Ｇｐ； ｐ＝１〜ＮＰ（ｍ）｝に変換する。修正部１８０３は、上位レイヤのサンプリング周波数Ｆｓ（ｍ）と当該レイヤのサンプリング周波数Ｆｓ（ｎ）で規定される定数Ｆｓ（ｍ）／Ｆｓ（ｎ）を上位レイヤのＬＳＦ係数｛Ｇｐ； ｐ＝１〜ＮＰ（ｍ）｝に乗じ、加算器１８０５に与える。
【０１８２】
入力端子１８０２からはデルタＬＳＦベクトルを表す符号化コードが入力される。デルタＬＳＦ符号帳１８０４は、この符号化コードを用いてデルタＬＳＦベクトルを復号し、加算器１８０５に与える。加算器１８０５は、修正後の上位レイヤＬＳＦ係数と復号されたデルタＬＳＦベクトルとを加算し、加算後のＬＳＦベクトルをＬＰＣ変換部１８０８に与える。ＬＰＣ変換部１８０８は、ＬＳＦベクトルからＬＰＣ係数に変換し、出力端子１８０６から出力する。
【０１８３】
このように、本実施の形態の音声復号化装置によれば、第２レイヤより下位の時間領域復号化において、符号化側の上位レイヤにて求められたＬＰＣ係数を考慮して探索した最適なデルタＬＳＦベクトルを利用して復号化することにより、時間領域符号化のＬＰＣ係数の符号化及び復号化をより効率的に行うことができ、低ビットレートで高品質に符号化できる。
【０１８４】
（実施の形態７）
次に、本発明の実施の形態７について、図面を参照して説明する。図１９は、本発明の実施の形態７に係る通信装置の構成を示すブロック図である。図１９における信号処理装置１９０３は前述した実施の形態１から実施の形態６に示した階層符号化装置の中の１つによって構成されている点に本実施の形態の特徴がある。
【０１８５】
図１９に示すように、本発明の実施の形態７に係る通信装置１９００は、入力装置１９０１、Ａ／Ｄ変換装置１９０２及びネットワーク１９０４に接続されている信号処理装置１９０３を具備している。
【０１８６】
Ａ／Ｄ変換装置１９０２は、入力装置１９０１の出力端子に接続されている。信号処理装置１９０３の入力端子は、Ａ／Ｄ変換装置１９０２の出力端子に接続されている。信号処理装置１９０３の出力端子はネットワーク１９０４に接続されている。
【０１８７】
入力装置１９０１は、人間の耳に聞こえる音波を電気的信号であるアナログ信号に変換してＡ／Ｄ変換装置１９０２に与える。Ａ／Ｄ変換装置１９０２はアナログ信号をディジタル信号に変換して信号処理装置１９０３に与える。信号処理装置１９０３は入力されてくるディジタル信号を符号化してコードを生成し、ネットワーク１９０４に出力する。
【０１８８】
このように、本発明の実施の形態の通信装置によれば、通信において前述した実施の形態１〜６に示したような効果を享受でき、少ないビット数で効率よく音響信号を符号化する階層符号化装置を提供することができる。
【０１８９】
（実施の形態８）
次に、本発明の実施の形態８について、図面を参照して説明する。図２０は、本発明の実施の形態８に係る通信装置の構成を示すブロック図である。図２０における信号処理装置２００３は前述した実施の形態１から実施の形態６に示した階層復号化装置の中の１つによって構成されている点に本実施の形態の特徴がある。
【０１９０】
図２０に示すように、本発明の実施の形態８に係る通信装置２０００は、ネットワーク２００１に接続されている受信装置２００２、信号処理装置２００３、及びＤ／Ａ変換装置２００４及び出力装置２００５を具備している。
【０１９１】
受信装置２００２の入力端子は、ネットワーク２００１に接続されている。信号処理装置２００３の入力端子は、受信装置２００２の出力端子に接続されている。Ｄ／Ａ変換装置２００４の入力端子は、信号処理装置２００３の出力端子に接続されている。出力装置２００５の入力端子は、Ｄ／Ａ変換装置２００４の出力端子に接続されている。
【０１９２】
受信装置２００２は、ネットワーク２００１からのディジタルの符号化音響信号を受けてディジタルの受信音響信号を生成して信号処理装置２００３に与える。信号処理装置２００３は、受信装置２００２からの受信音響信号を受けてこの受信音響信号に復号化処理を行ってディジタルの復号化音響信号を生成してＤ／Ａ変換装置２００４に与える。Ｄ／Ａ変換装置２００４は、信号処理装置２００３からのディジタルの復号化音声信号を変換してアナログの復号化音声信号を生成して出力装置２００５に与える。出力装置２００５は、電気的信号であるアナログの復号化音響信号を空気の振動に変換して音波として人間の耳に聴こえるように出力する。
【０１９３】
このように、本実施の形態の通信装置によれば、通信において前述した実施の形態１〜６に示したような効果を享受でき、少ないビット数で効率よく符号化された音響信号を復号することができるので、良好な音響信号を出力することができる。
【０１９４】
（実施の形態９）
次に、本発明の実施の形態９について、図面を参照して説明する。図２１は、本発明の実施の形態９に係る通信装置の構成を示すブロック図である。本発明の実施の形態９において、図２１における信号処理装置２１０３は、前述した実施の形態１から実施の形態６に示した音響符号化手段の中の１つによって構成されている点に本実施の形態の特徴がある。
【０１９５】
図２１に示すように、本発明の実施の形態９に係る通信装置２１００は、入力装置２１０１、Ａ／Ｄ変換装置２１０２、信号処理装置２１０３、ＲＦ変調装置２１０４及びアンテナ２１０５を具備している。
【０１９６】
入力装置２１０１は人間の耳に聞こえる音波を電気的信号であるアナログ信号に変換してＡ／Ｄ変換装置２１０２に与える。Ａ／Ｄ変換装置２１０２はアナログ信号をディジタル信号に変換して信号処理装置２１０３に与える。信号処理装置２１０３は入力されてくるディジタル信号を符号化して符号化音響信号を生成し、ＲＦ変調装置２１０４に与える。ＲＦ変調装置２１０４は、符号化音響信号を変調して変調符号化音響信号を生成し、アンテナ２１０５に与える。アンテナ２１０５は、変調符号化音響信号を電波として送信する。
【０１９７】
このように、本実施の形態の通信装置によれば、無線通信において前述した実施の形態１〜６に示したような効果を享受でき、少ないビット数で効率よく音響信号を符号化することができる。
【０１９８】
なお、本発明は、オーディオ信号を用いる送信装置、送信符号化装置又は音響信号符号化装置に適用することができる。また、本発明は、移動局装置又は基地局装置にも適用することができる。
【０１９９】
（実施の形態１０）
次に、本発明の実施の形態１０について、図面を参照して説明する。図２２は、本発明の実施の形態１０に係る通信装置の構成を示すブロック図である。本発明の実施の形態１０において、図２２における信号処理装置２２０３は、前述した実施の形態１から実施の形態６に示した音響復号化手段の中の１つによって構成されている点に本実施の形態の特徴がある。
【０２００】
図２２に示すように、本発明の実施の形態１０に係る通信装置２２００は、アンテナ２２０１、ＲＦ復調装置２２０２、信号処理装置２２０３、Ｄ／Ａ変換装置２２０４及び出力装置２２０５を具備している。
【０２０１】
アンテナ２２０１は、電波としてのディジタルの符号化音響信号を受けて電気信号のディジタルの受信符号化音響信号を生成してＲＦ復調装置２２０２に与える。ＲＦ復調装置２２０２は、アンテナ２２０１からの受信符号化音響信号を復調して復調符号化音響信号を生成して信号処理装置２２０３に与える。
【０２０２】
信号処理装置２２０３は、ＲＦ復調装置２２０２からのディジタルの復調符号化音響信号を受けて復号化処理を行ってディジタルの復号化音響信号を生成してＤ／Ａ変換装置２２０４に与える。Ｄ／Ａ変換装置２２０４は、信号処理装置２２０３からのディジタルの復号化音声信号を変換してアナログの復号化音声信号を生成して出力装置２２０５に与える。出力装置２２０５は、電気的信号であるアナログの復号化音声信号を空気の振動に変換して音波として人間の耳に聴こえるように出力する。
【０２０３】
このように、本実施の形態の通信装置によれば、無線通信において前述した実施の形態１〜６に示したような効果を享受でき、少ないビット数で効率よく符号化された音響信号を復号することができるので、良好な音響信号を出力することができる。
【０２０４】
なお、本発明は、オーディオ信号を用いる受信装置、受信復号化装置又は音声信号復号化装置に適用することができる。また、本発明は、移動局装置又は基地局装置にも適用することができる。
【０２０５】
また、本発明は上記実施の形態に限定されず、種々変更して実施することが可能である。例えば、上記実施の形態では、信号処理装置として行う場合について説明しているが、これに限られるものではなく、この信号処理方法をソフトウェアとして行うことも可能である。
【０２０６】
例えば、上記信号処理方法を実行するプログラムを予めＲＯＭ（Ｒｅａｄ Ｏｎｌｙ Ｍｅｍｏｒｙ）に格納しておき、そのプログラムをＣＰＵ（Ｃｅｎｔｒａｌ Ｐｒｏｃｅｓｓｏｒ Ｕｎｉｔ）によって動作させるようにしても良い。
【０２０７】
また、上記信号処理方法を実行するプログラムをコンピュータで読み取り可能な記憶媒体に格納し、記憶媒体に格納されたプログラムをコンピュータのＲＡＭ（Ｒａｎｄｏｍ Ａｃｃｅｓｓ ｍｅｍｏｒｙ）に記録して、コンピュータをそのプログラムにしたがって動作させるようにしても良い。
【０２０８】
なお、上記説明では、時間領域から周波数領域への変換法に離散フーリエ変換を用いる場合について説明を行っているがこれに限定されず直交変換であればいずれも適用できる。例えば、離散コサイン変換またはＭＤＣＴ（変形離散コサイン変換）等を適用することもできる。
【０２０９】
なお、本発明は、オーディオ信号を用いる受信装置、受信復号化装置又は音声信号復号化装置に適用することができる。また、本発明は、移動局装置又は基地局装置にも適用することができる。
【０２１０】
【発明の効果】
以上説明したように、本発明の音響信号の階層符号化方法および階層復号化方法によれば、第２レイヤ以下の符号化において、上位レイヤの符号化結果を復号化した信号と入力音響信号との差分を時間領域で符号化し、時間領域の符号化で符号化できない残差信号、すなわち差分信号と時間領域での符号化した信号を復号した復号信号との差分を周波数領域で符号化することにより、周期性のある信号を時間領域の符号化し、周期性のない信号を周波数領域の符号化することができ、低ビットレートで高品質な符号化を行うことができる。
【図面の簡単な説明】
【図１】本発明の実施の形態１に係る階層符号化装置の構成を示すブロック図
【図２】上記実施の形態の階層符号化装置の第ｎレイヤ（２≦ｎ≦Ｎ）符号化部の構成を示すブロック図
【図３】上記実施の形態の階層符号化装置の時間領域符号化部の構成を示すブロック図
【図４】上記本実施の形態の階層符号化装置の目標信号生成部の構成を示すブロック図
【図５】上記実施の形態の階層符号化装置の周波数領域符号化部の構成を示すブロック図
【図６】上記実施の形態の階層符号化装置の聴覚マスキング算出部の構成を示すブロック図
【図７】本発明の実施の形態２に係る階層復号化装置の構成を示すブロック図
【図８】上記実施の形態の階層復号化装置の第２レイヤ以降の復号化部の構成を示すブロック図
【図９】上記実施の形態の階層復号化装置の時間領域復号化部の構成を示すブロック図
【図１０】上記実施の形態の階層復号化装置の周波数領域復号化部の構成を示すブロック図
【図１１】本発明の実施の形態３に係る階層符号化装置の構成を示すブロック図
【図１２】上記実施の形態の階層符号化装置の時間領域符号化部の構成を示すブロック図
【図１３】本発明の実施の形態４に係る階層復号化装置の構成を示すブロック図
【図１４】上記実施の形態の階層復号化装置の時間領域復号化部の構成を示すブロック図
【図１５】本発明の実施の形態５に係る階層符号化装置の時間領域符号化部の構成を示すブロック図
【図１６】上記実施の形態の階層符号化装置のＬＰＣ量子化器の構成を示すブロック図
【図１７】本発明の実施の形態６に係る階層復号化装置の時間領域復号化部の構成を示すブロック図
【図１８】上記実施の形態の階層復号化装置のＬＰＣ復号器の構成を示すブロック図
【図１９】本発明の実施の形態７に係る通信装置の構成を示すブロック図
【図２０】本発明の実施の形態８に係る通信装置の構成を示すブロック図
【図２１】本発明の実施の形態９に係る通信装置の構成を示すブロック図
【図２２】本発明の実施の形態１０に係る通信装置の構成を示すブロック図
【符号の説明】
１０２ ＤＳ１部
１０３ 第１レイヤ符号化部
１０４、７０３ 第１レイヤ復号化部
１０７、１１３、４０４ 遅延器
１０５、７０４ ＵＳ１部
１０６ ＤＳ２部
１０８、１１４、４０５ 減算器
１０９ 第２レイヤ符号化部
１１０、７０４ 第２レイヤ復号化部
１１１、７０５、７０９、８０５、１３０５ 加算器
１１２、７０７ ＵＳ２部
１１５ 第３レイヤ符号化部
２０３ 時間領域符号化部
２０４ 目標信号生成部
２０５ 周波数領域符号化部
４０３ 復号部
５０３ 周波数領域変換部
５０４ 聴覚マスキング算出部
５０５ 量子化部
７０８ 第３レイヤ復号化部
８０３、１３０３ 時間領域復号化部
８０４、１３０４ 周波数領域復号化部
１００２ 変換係数復号化部
１００３ 時間領域変換部[0001]
TECHNICAL FIELD OF THE INVENTION
The present invention relates to a hierarchical encoding method and a hierarchical decoding method for an audio signal, and more particularly, to a hierarchical encoding method and a hierarchical decoding method suitable for efficiently using an audio signal such as a musical sound signal or an audio signal for compression encoding. About the method of conversion.
[0002]
[Prior art]
An acoustic coding technique for compressing a tone signal or a voice signal at a low bit rate is important for effective use of a transmission path capacity of radio waves and the like and a recording medium in mobile communication. There are G726 and G729 standardized by ITU (International Telecommunication Union) for audio coding for encoding an audio signal. These systems target narrowband signals (300 Hz to 3.4 kHz) and can perform high-quality encoding at 8 kbit / s to 32 kbit / s. In addition, there are ITU G722, G722.1, 3GPP (The 3rd Generation Partnership Project) AMR-WB, and the like as standard systems for wideband signals (50 Hz to 7 kHz). These methods can code a wideband audio signal with high quality at a bit rate of 6.6 kbit / s to 64 kbit / s.
[0003]
An effective method of encoding a speech signal at a low bit rate with high efficiency is CELP (Code Excited Linear Prediction). CELP is based on a model that simulates a human voice generation model by engineering, and passes an excitation signal represented by a random number or a pulse train through a pitch filter corresponding to the strength of the periodicity and a synthesis filter corresponding to the vocal tract characteristics, This is a method of determining an encoding code such that the square error between the output signal and the input signal is minimized under the weighting of auditory characteristics (for example, see Non-Patent Document 1). Many of the recent standard audio coding systems are based on CELP. For example, G729 can perform narrowband signal coding at 8 kbit / s, and AMR-WB can perform wideband signal coding at 6.6 kbit / s to 23.85 kbit / s. Can be encoded.
[0004]
On the other hand, in the case of musical sound encoding for encoding a musical sound signal, a musical sound signal is converted into a frequency domain like a layer III system or an AAC system standardized by MPEG (Moving Picture Expert Group), and the psychoacoustic is used. Transform coding in which coding is performed using a model is general. In these systems, it is known that a signal having a sampling frequency of 44.1 kHz has 64 kbit / s to 96 kbit / s per channel and hardly causes audible deterioration.
[0005]
However, when encoding a signal mainly composed of audio signals and having music or environmental sound superimposed on the background, if the audio encoding method is applied, the effect of the music or environmental sound in the background will cause the signal to be encoded only in the background. In addition, there is a problem that the audio signal is deteriorated and the overall quality is reduced. This is a problem that occurs because the speech coding system is based on a CELP-based system specialized for a speech model. In addition, the signal band that can be supported by the audio coding system is up to 7 kHz at most, and there is a problem that a signal having a higher band than that can not be sufficiently supported due to its configuration.
[0006]
On the other hand, music encoding can perform high-quality encoding on music, so that sufficient quality can be obtained even for audio signals having music and environmental sounds in the background as described above. The band of the target signal can be handled up to the CD quality of about 22 kHz. On the other hand, in order to realize high-quality encoding, it is necessary to use a high bit rate, and if the bit rate is suppressed to about 32 kbit / s, there is a problem that the quality of a decoded signal is reduced. For this reason, there is a problem that it cannot be used in a communication network having a low transmission rate.
[0007]
Combining these techniques to avoid the problems described above, the input signal is first coded by CELP in the first layer, and then the decoded signal is subtracted from the input signal to obtain a residual signal. A method of transform-encoding a signal in the second and subsequent layers can be considered. In this method, since the first layer uses CELP, the audio signal can be encoded with high quality, and the second layer and the subsequent layers cover the background music and environmental sound that cannot be expressed by the first layer, and the first layer. A signal having a frequency component higher than the frequency band can be efficiently encoded.
[0008]
However, in order to secure sufficient quality when music is input instead of voice, it is necessary to increase the bit allocation to the second and subsequent layers, resulting in a problem that the bit rate increases. This is a problem that arises because a speech-specific coding scheme such as CELP is applied to the first layer. That is, when a music signal is input, the CELP used in the first layer does not have high coding efficiency for music, so an error signal between the input signal and the decoded signal of the first layer (that is, the input signal of the second layer) Power is increased. As a result, it is necessary to allocate many bits to the second and subsequent layers to improve the quality of the final decoded signal.
[0009]
[Non-patent document 1]
"Code-Excited Linear Prediction (CELP): high quality speech at very low bit rates", Proc. ICASSP 85, pp. 937-940, 1985.
[0010]
[Problems to be solved by the invention]
As described above, the conventional apparatus has a problem that it is difficult to perform high-quality encoding at a low bit rate.
[0011]
The present invention has been made in view of such a point, and an object of the present invention is to provide a hierarchical encoding method and a hierarchical decoding method of an audio signal capable of performing high-quality encoding at a low bit rate.
[0012]
[Means for Solving the Problems]
A hierarchical encoding method according to the present invention is a hierarchical encoding method for encoding an input audio signal, decoding a signal encoded in an upper layer, and encoding a difference between the decoded signal and the input signal. A first encoding step of encoding an input audio signal in frame units of a length, and a second encoding step of encoding a difference between a signal obtained by decoding an encoding result of an upper layer and an input audio signal in one or more stages. Encoding step, wherein in the second encoding step, the difference between the signal obtained by decoding the encoding result of the upper layer and the input audio signal is encoded in both the time domain and the frequency domain. .
[0013]
According to this method, a signal having a periodicity is encoded in the time domain by encoding the difference between the signal obtained by decoding the encoding result of the upper layer and the input audio signal in the time domain and the frequency domain, and A signal having no characteristic can be encoded in the frequency domain, and the encoded signal can be encoded at low bit rate and high quality.
[0014]
In the hierarchical encoding method according to the present invention, the second encoding step includes a hierarchical decoding step of decoding a signal encoded in an upper layer to generate a decoded signal, and an upsampling step of increasing a sampling frequency of the decoded signal. And a hierarchical subtraction step of subtracting the decoded signal from the input audio signal to generate a difference signal, and a hierarchical encoding step of encoding the difference signal.
[0015]
According to this method, the input signal can be encoded corresponding to various sampling frequencies by setting the sampling frequency of the signal to be encoded in the lower layer higher than the sampling frequency of the signal to be encoded in the upper layer. .
[0016]
In the hierarchical encoding method of the present invention, the hierarchical encoding step includes a time domain encoding step of encoding a difference between a signal obtained by decoding an encoding result of an upper layer and an input audio signal in a time domain; A time domain signal decoding step of decoding the signal encoded in the domain encoding step to generate a decoded signal, and a time domain signal subtracting step of subtracting the decoded signal from the difference signal to generate a second difference signal; A frequency domain encoding step of encoding the second difference signal in a frequency domain.
[0017]
According to this method, in the encoding of the second layer or lower, the difference between the signal obtained by decoding the encoding result of the upper layer and the input audio signal is encoded in the time domain, and the difference signal is encoded in the time domain. By encoding the difference between the decoded signal and the decoded signal in the frequency domain, a periodic signal can be encoded in the time domain, and a non-periodic signal can be encoded in the frequency domain. High quality encoding can be performed at a low bit rate.
[0018]
In the hierarchical encoding method according to the present invention, the time domain encoding step includes a search candidate determining step of limiting an adaptive vector to be used in encoding from a pitch period obtained in an upper layer, and an input speech signal from the limited adaptive vector. And a search step of searching for an adaptive vector having the smallest difference.
[0019]
According to this method, in the time domain coding lower than the second layer, the candidate of the adaptive vector to be searched is limited from the adaptive vector of the adaptive codebook using the pitch period obtained in the upper layer. By performing coding using a limited adaptive vector, coding of a pitch period of time domain coding can be performed more efficiently, and high quality coding can be performed at a low bit rate.
[0020]
In the hierarchical encoding method according to the present invention, the time domain encoding step includes a quantization step of quantizing a pitch period, and the search candidate determining step includes a step of sampling a pitch period obtained in an upper layer by the layer. A modification is made to match the frequency, and the quantization step quantizes the pitch period of the layer using the modified pitch period.
[0021]
According to this method, the input signal can be encoded corresponding to various sampling frequencies by setting the sampling frequency of the signal to be encoded in the lower layer higher than the sampling frequency of the signal to be encoded in the upper layer. .
[0022]
In the hierarchical encoding method according to the present invention, in the time domain encoding step, an addition step of adding a parameter of an encoding obtained in an upper layer and a parameter that is a search candidate of the layer, and a result of the addition step A search step of searching for a parameter having the smallest difference from a parameter obtained from the input voice signal.
[0023]
In the hierarchical encoding method according to the present invention, the time domain encoding step includes a conversion step of converting an LPC coefficient obtained in an upper layer into an LSF coefficient, and the adding step is performed in the conversion step. The LSF coefficient added to the LSF coefficient held by the LSF codebook is added, and the searching step searches for an LSF coefficient that minimizes the difference between the added LSF coefficient and the LSF coefficient obtained from the input audio signal. did.
[0024]
According to these methods, in time domain decoding lower than the second layer, decoding is performed using the optimal adaptive vector searched in consideration of the LPC coefficient obtained in the upper layer on the encoding side. Accordingly, encoding and decoding of the pitch period of time domain encoding can be performed more efficiently, and high-quality encoding can be performed at a low bit rate.
[0025]
The hierarchical encoding method of the present invention includes an auditory masking step of calculating auditory masking from an input audio signal, and the frequency domain encoding step performs encoding using a signal after masking using the auditory masking. I made it.
[0026]
According to this method, the auditory masking is calculated from the spectrum of the input signal, and the transform coefficients are quantized so that the quantization distortion is equal to or less than the masking value, thereby efficiently quantizing the transform coefficients at a small bit rate. can do.
[0027]
A hierarchical decoding method according to the present invention is a hierarchical decoding method for encoding an input audio signal, decoding a signal encoded in an upper layer, and decoding a signal obtained by encoding a difference between the decoded signal and the input signal. A first decoding step of decoding an encoded code of a first layer; a second decoding step of decoding an encoded code of a layer lower than the second layer in both a time domain and a frequency domain; A decoding step and an adding step of adding the decoding result of the second decoding step are provided.
[0028]
According to this method, an encoded signal obtained by encoding a difference between a signal obtained by decoding an encoding result of an upper layer and an input audio signal in a time domain and a frequency domain is converted into a signal having periodicity in a time domain and a periodic signal. By decoding a signal having no characteristics in the frequency domain, high-quality encoding and decoding can be performed at a low bit rate.
[0029]
The hierarchical decoding method of the present invention includes an upsampling step of upsampling a sampling frequency of a decoding result of the first decoding step to a sampling frequency of a decoding result of the second decoding step, and the adding step includes an upsampling step. The decoding results of the later first decoding step and the second decoding step are added.
[0030]
According to this method, by making the sampling frequency of the signal to be decoded in the lower layer higher than the sampling frequency of the signal to be decoded in the upper layer, it is possible to decode the signal obtained by encoding the signal corresponding to various sampling frequencies. it can.
[0031]
In the hierarchical decoding method according to the present invention, the second decoding step includes a time domain decoding step of decoding the layer using a pitch cycle or an LPC coefficient of an upper layer in the time domain decoding.
[0032]
According to this method, in time domain decoding lower than the second layer, a candidate adaptive vector to be used for decoding is obtained from an adaptive vector of an adaptive codebook using a pitch period obtained in an upper layer on the encoding side. , The encoding and decoding of the pitch period of the time domain encoding can be performed more efficiently, and high-quality encoding can be performed at a low bit rate.
[0033]
According to this method, in the time domain decoding lower than the second layer, the adaptive vector used for decoding is obtained from the adaptive vector of the adaptive codebook using the LPC coefficient obtained in the upper layer on the encoding side. By limiting the candidates, the encoding and decoding of the pitch period of the time domain encoding can be performed more efficiently, and high-quality encoding can be performed at a low bit rate.
[0034]
In the hierarchical decoding method according to the present invention, the second decoding step includes a frequency domain decoding step of decoding an encoded code obtained by encoding a frequency domain coefficient using auditory masking.
[0035]
According to this method, the auditory masking is calculated from the spectrum of the input signal, and the transform coefficients are quantized so that the quantization distortion is equal to or less than the masking value, thereby efficiently quantizing the transform coefficients at a small bit rate. can do.
[0036]
A hierarchical encoding device of the present invention is a hierarchical encoding device that encodes an input audio signal, decodes a signal encoded in an upper layer, and encodes a difference between the decoded signal and the input signal. A first encoding unit that encodes an input audio signal in frame units of a length, and a second encoding unit that encodes a difference between a signal obtained by decoding an encoding result of an upper layer and an input audio signal in one or more stages. Encoding means, wherein the second encoding means encodes the difference between the signal obtained by decoding the encoding result of the upper layer and the input audio signal in both the time domain and the frequency domain. .
[0037]
According to this configuration, by encoding the difference between the signal obtained by decoding the encoding result of the upper layer and the input audio signal in the time domain and the frequency domain, a periodic signal is encoded in the time domain, A signal having no characteristic can be encoded in the frequency domain, and the encoded signal can be encoded at low bit rate and high quality.
[0038]
In the hierarchical encoding device according to the present invention, the second encoding means decodes a signal encoded in an upper layer to generate a decoded signal, and an upsampling means for increasing a sampling frequency of the decoded signal. And a hierarchical subtraction means for subtracting the decoded signal from the input audio signal to generate a difference signal, and a hierarchical encoding means for encoding the difference signal.
[0039]
According to this configuration, the input signal can be encoded corresponding to various sampling frequencies by setting the sampling frequency of the signal to be encoded in the lower layer higher than the sampling frequency of the signal to be encoded in the upper layer.
[0040]
The hierarchical encoding device according to the present invention, wherein the hierarchical encoding means includes a time domain encoding means for encoding a difference between a signal obtained by decoding an encoding result of an upper layer and an input audio signal in a time domain; A time-domain signal decoding unit that decodes the signal encoded by the region encoding unit to generate a decoded signal, a time-domain signal subtraction unit that subtracts the decoded signal from the difference signal to generate a second difference signal, Frequency domain encoding means for encoding the second differential signal in the frequency domain.
[0041]
According to this configuration, in the encoding of the second layer or lower, the difference between the signal obtained by decoding the encoding result of the upper layer and the input audio signal is encoded in the time domain, and the difference signal is encoded in the time domain. By encoding the difference between the decoded signal and the decoded signal in the frequency domain, a periodic signal can be encoded in the time domain, and a non-periodic signal can be encoded in the frequency domain. High quality encoding can be performed at a low bit rate.
[0042]
In the hierarchical coding apparatus according to the present invention, the time domain coding means may include a codebook holding an adaptive vector generated in the past, and a search candidate for limiting an adaptive vector used in coding from a pitch period obtained in an upper layer. A configuration including a determination unit and a search unit that searches for an adaptation vector having the smallest audible difference from the input speech signal from the limited adaptation vector is adopted.
[0043]
According to this configuration, in the time domain coding lower than the second layer, the candidate of the adaptive vector to be searched is limited from the adaptive vector of the adaptive codebook using the pitch period obtained in the upper layer. By performing coding using a limited adaptive vector, coding of a pitch period of time domain coding can be performed more efficiently, and high quality coding can be performed at a low bit rate.
[0044]
In the hierarchical coding apparatus according to the present invention, the time domain coding means includes quantization means for quantizing a pitch cycle, and the search candidate determining means determines a pitch cycle obtained in an upper layer by a sampling frequency of the layer. The quantization means adopts a configuration in which the pitch period of the layer is quantized using the corrected pitch period.
[0045]
According to this configuration, the input signal can be encoded corresponding to various sampling frequencies by setting the sampling frequency of the signal to be encoded in the lower layer higher than the sampling frequency of the signal to be encoded in the upper layer. .
[0046]
In the hierarchical coding apparatus according to the present invention, the time domain coding unit may include an addition unit that adds a coding parameter obtained in an upper layer and a parameter that is a search candidate of the layer, and a result of the addition unit. A configuration including a search unit that searches for a parameter having the smallest difference from the input voice signal is adopted.
[0047]
In the hierarchical coding apparatus according to the present invention, the time domain coding means includes an LSF codebook for holding LSF coefficients, and a conversion means for converting LPC coefficients obtained in an upper layer into LSF coefficients, The means adds the LSF coefficient converted by the conversion means and the LSF coefficient held by the LSF codebook, and the search means generates an audible difference between the added LSF coefficient and the LSF coefficient obtained from the input audio signal. A configuration for searching for an LSF coefficient that minimizes the difference is employed.
[0048]
According to these configurations, in time domain decoding lower than the second layer, decoding is performed using the optimal adaptive vector searched in consideration of the LPC coefficient obtained in the upper layer on the encoding side. Accordingly, encoding and decoding of the pitch period of time domain encoding can be performed more efficiently, and high-quality encoding can be performed at a low bit rate.
[0049]
The hierarchical encoding device of the present invention includes an auditory masking unit that calculates auditory masking from an input audio signal, and the frequency domain encoding unit encodes using the signal after the masking using the auditory masking. Take.
[0050]
According to this configuration, the auditory masking is calculated from the spectrum of the input signal, and the transform coefficients are quantized so that the quantization distortion is equal to or less than the masking value, thereby efficiently quantizing the transform coefficients at a small bit rate. can do.
[0051]
A hierarchical decoding device of the present invention encodes an input audio signal, decodes a signal encoded in an upper layer, and decodes a signal obtained by encoding a difference between the decoded signal and the input signal. A first decoding unit that decodes an encoded code of a first layer; a second decoding unit that decodes an encoded code of a layer lower than a second layer in both a time domain and a frequency domain; A configuration including a decoding unit and an adding unit for adding the decoding result of the second decoding unit is adopted.
[0052]
According to this configuration, an encoded signal obtained by encoding the difference between the signal obtained by decoding the encoding result of the upper layer and the input audio signal in the time domain and the frequency domain is converted into a signal having periodicity in the time domain, By decoding a signal having no characteristics in the frequency domain, high-quality encoding and decoding can be performed at a low bit rate.
[0053]
The hierarchical decoding apparatus according to the present invention includes up-sampling means for up-sampling a sampling frequency of a decoding result of the first decoding means to a sampling frequency of a decoding result of the second decoding means. A configuration is adopted in which the decoding results of the later first decoding means and the second decoding means are added.
[0054]
According to this configuration, by setting the sampling frequency of the signal to be decoded in the lower layer higher than the sampling frequency of the signal to be decoded in the upper layer, it is possible to decode a signal obtained by encoding a signal corresponding to various sampling frequencies. it can.
[0055]
The hierarchical decoding device of the present invention employs a configuration in which the second decoding means includes a time-domain decoding means for decoding a layer using a pitch cycle or an LPC coefficient of an upper layer in time-domain decoding.
[0056]
According to this configuration, in time domain decoding lower than the second layer, a candidate adaptive vector to be used for decoding is obtained from the adaptive vector of the adaptive codebook using the pitch period obtained in the upper layer on the encoding side. , The encoding and decoding of the pitch period of the time domain encoding can be performed more efficiently, and high-quality encoding can be performed at a low bit rate.
[0057]
According to this configuration, in the time domain decoding lower than the second layer, the adaptive vector used for decoding is obtained from the adaptive vector of the adaptive codebook using the LPC coefficient obtained in the upper layer on the encoding side. By limiting the candidates, the encoding and decoding of the pitch period of the time domain encoding can be performed more efficiently, and high-quality encoding can be performed at a low bit rate.
[0058]
The hierarchical decoding apparatus according to the present invention employs a configuration in which the second decoding unit includes a frequency domain decoding unit that decodes an encoded code obtained by encoding a frequency domain coefficient using auditory masking.
[0059]
According to this configuration, the auditory masking is calculated from the spectrum of the input signal, and the transform coefficients are quantized so that the quantization distortion is equal to or less than the masking value, thereby efficiently quantizing the transform coefficients at a small bit rate. can do.
[0060]
An audio signal transmitting apparatus according to the present invention includes an audio input unit for converting an audio signal into an electric signal, an A / D conversion unit for converting a signal output from the audio input unit into a digital signal, and an A / D converter. Means for encoding the digital signal output from the means, RF modulation means for modulating the coded code output from the coding apparatus into a radio frequency signal, and output from the RF modulation means. And a transmission antenna that converts a signal into a radio wave and transmits the radio wave.
[0061]
According to this configuration, by encoding the difference between the signal obtained by decoding the encoding result of the upper layer and the input audio signal in the time domain and the frequency domain, a periodic signal is encoded in the time domain, A signal having no characteristic can be encoded in the frequency domain, and the encoded signal can be encoded at low bit rate and high quality.
[0062]
The acoustic signal receiving apparatus according to the present invention includes a receiving antenna for receiving a radio wave, an RF demodulating means for demodulating a signal received by the receiving antenna, and the hierarchical decoding for decoding information obtained by the RF demodulating means. A D / A converter for converting a signal output from the decoding device into an analog signal, and an audio output unit for converting an electric signal output from the D / A converter into an audio signal. The configuration provided is adopted.
[0063]
According to this configuration, an encoded signal obtained by encoding the difference between the signal obtained by decoding the encoding result of the upper layer and the input audio signal in the time domain and the frequency domain is converted into a signal having periodicity in the time domain, By decoding a signal having no characteristics in the frequency domain, high-quality encoding and decoding can be performed at a low bit rate.
[0064]
The communication terminal device of the present invention employs a configuration including at least one of the above-described acoustic signal transmitting device and the above-described acoustic signal receiving device. The base station apparatus of the present invention employs a configuration including at least one of the above-described acoustic signal transmitting apparatus and the above-described acoustic signal receiving apparatus.
[0065]
According to these configurations, a signal having a periodicity is encoded in the time domain by encoding the difference between the signal obtained by decoding the encoding result of the upper layer and the input audio signal in the time domain and the frequency domain, Non-periodic signals can be coded in the frequency domain, and can be coded to perform high-quality coding at a low bit rate.
[0066]
Further, according to this configuration, an encoded signal obtained by encoding the difference between the signal obtained by decoding the encoding result of the upper layer and the input audio signal in the time domain and the frequency domain is referred to as a signal having periodicity in the time domain. By decoding a signal having no periodicity in the frequency domain, high-quality encoding and decoding can be performed at a low bit rate.
[0067]
BEST MODE FOR CARRYING OUT THE INVENTION
The present inventor has focused on the fact that the difference between the signal obtained by decoding the encoding result of the upper layer and the input audio signal includes both a periodic signal and a non-periodic signal. I came to.
[0068]
That is, the gist of the present invention is that, in the encoding of the second layer or lower, the difference between the signal obtained by decoding the encoding result of the upper layer and the input audio signal is encoded in the time domain, and the encoding is performed in the time domain. By encoding in the frequency domain the difference between the impossible residual signal, that is, the difference signal and the decoded signal obtained by decoding the encoded signal in the time domain, a periodic signal is encoded in the time domain, A non-existent signal can be coded in the frequency domain, and is coded to perform high-quality coding at a low bit rate.
[0069]
Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings. In the following description of the embodiment, the case where the number of layers N is set to 3 will be described. However, the present invention is not limited to this numerical value, and N is a natural number and a configuration satisfying the condition of N ≧ 2 It is possible to apply.
[0070]
(Embodiment 1)
FIG. 1 is a block diagram showing a configuration of a hierarchical encoding device according to Embodiment 1 of the present invention. 1 includes an input terminal 101, a DS1 unit 102, a first layer encoding unit 103, a first layer decoding unit 104, a US1 unit 105, a DS2 unit 106, a delay unit 107, a subtractor 108, a second layer encoding unit 109, a second layer decoding unit 110, an adder 111, a US2 unit 112, a delay unit 113, a subtractor 114, and a third layer code. Multiplexing section 115, multiplexing section 118, and output terminal 119.
[0071]
The present embodiment is characterized in that the sampling frequency of a signal input to each layer has a relationship represented by the following equation (1).
(Equation 1)

Here, Fs (n) represents the sampling frequency of the signal of the n-th layer. According to the present embodiment, it is possible to perform encoding corresponding to a plurality of sampling frequencies.
[0072]
From the input terminal 101, an acoustic signal of the sampling frequency Fs (3) is input and given to the DS1 unit 102.
[0073]
The DS1 unit 102 downsamples the input audio signal and reduces the sampling frequency of the input audio signal from Fs (3) to Fs (1). Then, DS1 section 102 outputs an input signal of sampling frequency Fs (1) to first layer encoding section 103.
[0074]
First layer coding section 103 has an adaptive codebook that holds a previously generated driving excitation signal as an internal state, and efficiently codes a signal having a high periodicity by using the adaptive codebook. be able to. First layer encoding section 103 determines the first encoded code such that the perceptual distortion between the input audio signal and the decoded signal generated after encoding is minimized. A typical method applied to the first layer encoding section 103 is a code excitation linear prediction method (CELP).
[0075]
Then, first layer encoding section 103 outputs the obtained first encoded code to first layer decoding section 104 and multiplexing section 118. First layer decoding section 104 generates a first layer decoded signal using the first encoded code, and outputs the first layer decoded signal to US1 section 105.
[0076]
US1 section 105 upsamples the first layer decoded signal and increases the sampling frequency from Fs (1) to Fs (2). Then, US1 section 105 outputs the first layer decoded signal of sampling frequency Fs (2) to subtractor 108 and adder 111.
[0077]
Next, an audio signal input from the input terminal 101 is provided to the DS2 unit 106. The DS2 unit 106 downsamples the input audio signal and reduces the sampling frequency of the input audio signal from Fs (3) to Fs (2). Then, DS2 section 106 outputs an input signal of sampling frequency Fs (2) to delay device 107.
[0078]
The delay unit 107 delays the acoustic signal input from the input terminal 101 by a predetermined time length and outputs the delayed audio signal to the subtracter 108. That is, it has a role of correcting a delay generated in the DS1 unit 102, the first layer encoding unit 103, the first layer decoding unit 104, the US1 unit 105, and the DS2 unit 106.
[0079]
The subtractor 108 calculates a difference between the output signal of the delay unit 107 and the above-described first layer decoded signal to generate a second layer residual signal. Then, subtracter 108 outputs the second layer residual signal to second layer encoding section 109.
[0080]
Second layer encoding section 109 encodes the second layer residual signal so that quality is perceptually improved, and determines a second encoded code. Then, second layer encoding section 109 outputs second layer decoding section 110 and the second encoded code to multiplexing section 118. Second layer decoding section 110 performs a decoding process using the second encoded code, generates a second layer decoded residual signal, and outputs the second layer decoded residual signal to adder 111.
[0081]
The adder 111 takes the sum of the first layer decoded signal and the second layer decoded residual signal to generate a second layer decoded signal. Then, adder 111 outputs the second layer decoded signal to US2 section 112.
[0082]
US2 section 112 up-samples the second layer decoded signal and increases the sampling frequency from Fs (2) to Fs (3). Then, US2 section 112 outputs the first layer decoded signal of sampling frequency Fs (3) to subtractor 114.
[0083]
Next, the delay unit 113 outputs the acoustic signal to the subtractor 114 after delaying the acoustic signal input from the input terminal 101 by a predetermined time length. That is, the delay unit 113 has a role of correcting a delay generated in the encoding unit and the decoding unit up to the previous stage, specifically, a delay generated in the signal processing from the DS1 unit 102 to the US2 unit 112.
[0084]
The subtractor 114 calculates a difference between the output signal of the delay unit 113 and the above-described second layer decoded signal to generate a third layer residual signal. Then, subtracter 114 outputs the third layer residual signal to third layer encoding section 115.
[0085]
Third layer encoding section 115 determines a third encoded code by encoding the third layer residual signal such that the quality is improved audibly, and multiplexes this third encoded code into multiplexing section 118. Output to
[0086]
The multiplexing unit 118 multiplexes the first coded code, the second coded code, and the third coded code by predetermined means, and generates a coded bit sequence. Then, the multiplexing unit 118 outputs the encoded bit sequence from the output terminal 119.
[0087]
Next, details of the encoding of the second and subsequent layers will be described. The hierarchical coding apparatus according to the present embodiment generates a residual signal from a difference between an input audio signal and a signal obtained by decoding a previous-stage encoded signal in encoding of a second layer or later, and generates the residual signal. A feature is that encoding is performed by a time domain encoding unit and a frequency domain encoding unit.
[0088]
Next, the n-th layer (2 ≦ n ≦ N) encoder will be described. FIG. 2 is a block diagram illustrating a configuration of an n-th layer (2 ≦ n ≦ N) encoding unit of the hierarchical encoding device according to the present embodiment.
[0089]
In the n-th layer residual signal, which is the difference between the input audio signal and the signal obtained by decoding the n-th layer (2 ≦ n ≦ N) encoded signal, the encoding noise up to the upper layer and the sampling frequency are higher. And high-frequency components due to this.
[0090]
The n-th layer residual signal contains both components that can be efficiently coded when processed in the time domain and components that can be coded efficiently when processed in the frequency domain. Therefore, there is an effect that efficient encoding can be realized by performing encoding in two regions, a time domain and a frequency domain. An input signal is provided to both the time domain coding unit and the frequency domain coding unit. This input signal is used for calculation of auditory masking and the like in order to realize audio quality of high quality. Hereinafter, a detailed description will be given with reference to FIG.
[0091]
An n-th layer residual signal is input from an input terminal 201 and provided to a time domain coding unit 203 and a target signal generation unit 204. The time domain encoding unit 203 encodes the n-th layer residual signal in the time domain using the n-th layer residual signal and an input signal input from the input terminal 202 to generate an encoded code. I do. Then, time domain encoding section 203 outputs the encoded code to target signal generating section 204 and multiplexing section 206. The details of the time domain coding unit 203 will be described later with reference to FIG.
[0092]
Next, the target signal generation unit 204 generates an input signal of the frequency domain encoding unit 205 using the input signal input from the input terminal 201 and the encoded code obtained by the time domain encoding unit 203. Details of the target signal generation unit 204 will be described later with reference to FIG.
[0093]
Next, the frequency domain encoding unit 205 performs encoding in the frequency domain using the signal generated by the target signal generation unit 204 and the input signal input from the input terminal 202 to generate an encoded code, Output to the conversion unit 206. The details of the frequency domain coding unit 205 will be described later with reference to FIG.
[0094]
Hereinafter, details of each block will be described. FIG. 3 is a block diagram illustrating a configuration of the time domain encoding unit of the hierarchical encoding device according to the present embodiment. 3 includes an input terminal 301, an LPC analyzer 302, an LPC quantizer 303, an LPC decoder 304, an auditory weighting filter 305, a synthesis filter 306, and an adaptive codebook 307. , Noise codebook 308, multiplier 309, multiplier 310, gain codebook 311, adder 312, subtractor 313, searcher 314, multiplexing section 315, and output terminal 316. Mainly composed.
[0095]
The LPC analyzer 302 calculates an LPC coefficient from the acoustic signal of the sampling rate Fs (n) input from the input terminal 301. The LPC coefficient is a coefficient used for improving the perceptual quality. LPC analyzer 302 outputs the LPC coefficient to perceptual weight filter 305 and LPC quantizer 303.
[0096]
The LPC quantizer 303 converts the LPC coefficients into parameters suitable for quantization such as LSF coefficients and performs quantization. Then, LPC quantizer 303 outputs the encoded code obtained by the quantization to multiplexing section 315 and LPC decoder 304.
[0097]
LPC decoder 304 calculates an LSF coefficient after quantization from the encoded code, and converts the LSF coefficient into an LPC coefficient. With this processing, the quantized LPC coefficient is obtained. Then, LPC decoder 304 outputs the quantized LPC coefficients to synthesis filter 306.
[0098]
The synthesis filter 306 searches for an adaptive vector, an adaptive gain, a noise vector, and a noise gain by using the quantized LPC coefficients. Next, an adaptive vector, an adaptive vector gain, a noise vector, and a method of searching for a noise vector gain will be described.
[0099]
Adaptive codebook 307 holds a previously generated drive excitation signal as an internal state, and generates an adaptive vector by repeating this internal state at a desired pitch cycle. An appropriate range of the pitch period is between 60 Hz and 400 Hz. The noise codebook 308 outputs a noise vector stored in a storage area in advance or a noise vector generated according to a rule without having a storage area like an algebraic structure.
[0100]
Gain codebook 311 outputs an adaptive vector gain multiplied by the adaptive vector to multiplier 309, and outputs a noise vector gain multiplied by the noise vector to multiplier 310.
[0101]
The multiplier 309 multiplies the adaptive vector by the adaptive vector gain and outputs the result to the adder 312. The multiplier 310 multiplies the noise vector by the noise vector gain and outputs the result to the adder 312.
[0102]
The adder 312 adds the adaptive vector multiplied by the adaptive vector gain and the noise vector multiplied by the noise vector gain to generate a drive excitation signal. Then, the adder 312 outputs the driving sound source signal to the synthesis filter 306.
[0103]
The synthesis filter 306 passes the driving sound source signal through the synthesis filter to generate a synthesized signal, and outputs the synthesized signal to the subtractor 313.
[0104]
The subtractor 313 subtracts the synthesized signal from the n-th layer prediction residual signal input from the input terminal 317, and outputs the signal after the subtraction to the audibility weighting filter 305.
[0105]
The auditory weight filter 305 weights the signal obtained by the subtractor 313 based on the LPC coefficient obtained by the LPC analyzer 302. This is performed for the purpose of performing spectrum shaping so that the spectrum of the quantization distortion is masked by the spectrum envelope of the input signal.
[0106]
The searcher 314 efficiently searches for a combination of an adaptive vector, an adaptive vector gain, a noise vector, and a noise vector gain that minimizes the distortion defined from the signal after the subtraction, and sends the encoded codes to the multiplexing unit 315. .
[0107]
The searcher 314 determines an encoded code i, j, m or i, j, m, n that minimizes the distortion defined by the following equation (2) or (3), and multiplexes them. 315.
(Equation 2)

[Equation 3]

Here, t (k) is the n-th layer residual signal, q _i (K) is the ith adaptation vector, c _j (K) indicates the j-th noise vector, and β and γ indicate the adaptive vector gain and the noise vector gain, respectively.
[0108]
Equations (2) and (3) differ in the configuration of the gain codebook. In the case of equation (2), the gain codebook is the adaptive vector gain β _m And noise vector gain γ _m As an element, and an encoded code m for specifying the vector is determined. In the case of equation (3), the gain codebook is the adaptive vector gain β _m And noise vector gain γ _n Respectively, and the respective encoded codes m and n are determined independently. H (l) represents an impulse response of the audibility weighting filter.

Is an operator representing convolution.
[0109]
After all the encoded codes are determined, the multiplexing unit 315 combines the encoded codes into one and outputs it from the output terminal 316. Then, in preparation for the decoding process in the next frame (or subframe), the adaptive codebook is selected by using the selected adaptive vector, adaptive vector gain, noise vector, and driving excitation signal represented by using the noise vector gain. Update the internal state of.
[0110]
Next, details of the target signal generation unit 204 will be described. FIG. 4 is a block diagram illustrating a configuration of the target signal generation unit of the hierarchical encoding device according to the present embodiment. 4 mainly includes an input terminal 401, an input terminal 402, a decoding unit 403, a delay unit 404, a subtractor 405, and an output terminal 406.
[0111]
An encoded code obtained by the time domain encoding unit 203 is input from an input terminal 401. The decoding unit 403 generates a decoded signal according to the information of the encoded code.
[0112]
Delay device 404 gives a delay to the second layer residual signal input from input terminal 402 so as to correct the delay generated in time domain encoding section 203 and decoding section 403, and outputs the result to subtractor 405.
[0113]
The subtracter 405 subtracts the decoded signal obtained by the decoding unit 403 from the output signal of the delay unit 404 to generate a target signal of the frequency domain encoding unit 205, and outputs the subtracted signal from an output terminal 406.
[0114]
Next, the details of the frequency domain coding unit 205 will be described. FIG. 5 is a block diagram illustrating a configuration of the frequency domain encoding unit of the hierarchical encoding device according to the present embodiment. 5 mainly includes an input terminal 501, an input terminal 502, a frequency domain transformation unit 503, an auditory masking calculation unit 504, a quantization unit 505, and an output terminal 506. You.
[0115]
The signal input from the input terminal 501 to the frequency domain coding unit 205 is the target signal obtained by the target signal generation unit 204.
[0116]
The frequency domain transform unit 503 performs frequency transform after multiplying the target signal by the analysis window, and outputs a transform coefficient obtained by the frequency transform to the quantization unit 505. As a method of the frequency conversion here, a modified discrete cosine transform (MDCT), a discrete Fourier transform (DFT), or the like can be used.
[0117]
An audio signal of the sampling frequency Fs (n) is provided from the input terminal 502 and input to the auditory masking calculation unit 504. The auditory masking calculation unit 504 calculates an auditory masking representing a threshold value of noise power that is not perceived by a human, and outputs the auditory masking to the quantization unit 505.
[0118]
The quantization unit 505 quantizes the transform coefficient obtained by the frequency domain transform unit 503 using auditory masking, and outputs an encoded code obtained at that time from an output terminal 506.
[0119]
Next, a method of calculating auditory masking will be described in detail with reference to FIG. FIG. 6 is a block diagram illustrating a configuration of the auditory masking calculation unit of the hierarchical encoding device according to the present embodiment. The human auditory characteristic has a masking effect that, when a certain signal is given, a signal located near the frequency of the signal becomes difficult to hear. By utilizing this characteristic, the auditory masking is calculated from the spectrum of the input signal, and the transform coefficients are quantized so that the quantization distortion is equal to or less than the masking value, so that the transform coefficients can be efficiently quantized at a small bit rate. Can be
[0120]
An input signal is provided from an input terminal 601, and the signal is converted into a frequency domain by a frequency conversion unit 602 to calculate a conversion coefficient. As a method of converting to the frequency domain, it is possible to use a modified discrete cosine transform (MDCT), a discrete Fourier transform (DFT), or the like as described above. Here, the case where DFT is used will be described, and the Fourier coefficient obtained by DFT is represented as {Re (m), Im (m)}.
[0121]
In FIG. 6, frequency conversion section 602 performs a Fourier transform on the input signal output from delay unit 107, and calculates a Fourier coefficient {Re (m), Im (m)}. Here, m represents a frequency.
[0122]
The bark spectrum calculator 603 calculates the bark spectrum B (k) using the following equation (4).
(Equation 4)

Here, P (m) represents a power spectrum and is obtained from the following equation (5).
(Equation 5)

K corresponds to the number of the bark spectrum, and FL (k) and FH (k) represent the lowest frequency (Hz) and the highest frequency (Hz) of the k-th bark spectrum, respectively. The bark spectrum B (k) represents the spectrum intensity when band-divided at equal intervals on the bark scale. When the Hertz scale is represented by f and the Bark scale is represented by B, the relationship between the Hertz scale and the Bark scale is represented by the following equation (6).
(Equation 6)

[0123]
The spread function convolution unit 604 convolves the spread function SF (k) with the bark spectrum B (k) using the following equation (7) to calculate C (k).
(Equation 7)

[0124]
The tonality calculation unit 605 calculates the spectrum flatness SFM (k) of each bark spectrum from the power spectrum P (m) using the following equation (8).
(Equation 8)

Here, μg (k) represents the geometric mean of the k-th bark spectrum, and μa (k) represents the arithmetic mean of the k-th bark spectrum. Then, the tonality calculation unit 605 calculates the tonality coefficient α (k) from the decibel value SFMdB (k) of the spectrum flatness SFM (k) using the following equation (9).
(Equation 9)

[0125]
The auditory masking calculation unit 606 calculates the offset O (k) of each bark scale from the tonality coefficient α (k) calculated by the tonality calculation unit 605 using the following equation (10).
(Equation 10)

[0126]
Then, the auditory masking calculation unit 606 calculates the auditory masking T (k) by subtracting the offset O (k) from C (k) obtained by the spread function convolution unit 604 using the following equation (11).
[Equation 11]

Where T _q (K) represents an absolute threshold. The absolute threshold represents the minimum value of auditory masking observed as a human auditory characteristic. Then, the auditory masking calculation unit 606 converts the auditory masking T (k) represented by the Bark scale into the Hertz scale M (m) and outputs the result.
[0127]
As described above, according to the hierarchical coding device of the present embodiment, the difference between the signal obtained by decoding the coding result of the upper layer and the input audio signal is coded in the time domain and the frequency domain, and thereby the periodicity is reduced. It is possible to encode a signal having a frequency in the time domain, and to encode a signal having no periodicity in the frequency domain, and to perform high-quality encoding at a low bit rate.
[0128]
In particular, according to the hierarchical encoding device of the present embodiment, in the encoding of the second layer or lower, the difference between the signal obtained by decoding the encoding result of the upper layer and the input audio signal is encoded in the time domain, By encoding in the frequency domain the difference between the difference signal and the decoded signal obtained by decoding the encoded signal in the time domain, a periodic signal is encoded in the time domain, and a non-periodic signal is encoded in the frequency domain. It can be coded and can be coded to perform high quality coding at a low bit rate.
[0129]
Further, according to the hierarchical encoding device of the present embodiment, by setting the sampling frequency of the signal to be encoded in the lower layer higher than the sampling frequency of the signal to be encoded in the upper layer, it is possible to cope with various sampling frequencies. The input signal can be encoded.
[0130]
(Embodiment 2)
In the present embodiment, an example in which a signal encoded by the hierarchical encoding device of Embodiment 1 is decoded will be described. A feature of the present embodiment is that the encoded code of the hierarchical encoding method described in the first embodiment can be decoded, and as a result, a high-quality audio signal can be decoded.
[0131]
FIG. 7 is a block diagram showing a configuration of a hierarchical decoding device according to Embodiment 2 of the present invention. 7 includes an input terminal 701, a separating unit 702, a first layer decoding unit 703, a US1 unit 704, an adder 705, a second layer decoding unit 706, and a US2 unit. 707, a third layer decoding section 708, an adder 709 and an output terminal 710.
[0132]
A coded bit sequence coded by the hierarchical coding device of FIG. 1 is input from an input terminal 701.
[0133]
Separating section 702 separates the coded bit string, and outputs a first coded code obtained by first layer coding, a second coded code obtained by second layer coding, and a third coded code obtained by third layer coding. Generate an encoded code. Then, separating section 702 outputs the first encoded code to first layer decoding section 703, outputs the second encoded code to second layer decoding section 706, and outputs the third encoded code to the third layer decoding section. Output to the decoding unit 708.
[0134]
First layer decoding section 703 performs a decoding process using the first encoded code obtained in separation section 702, and generates a first layer decoded signal.
[0135]
US1 section 704 up-samples the first layer decoded signal and increases the sampling frequency from Fs (1) to Fs (2). Then, US1 section 704 outputs the first layer decoded signal of sampling frequency Fs (2) to adder 705.
[0136]
Next, second layer decoding section 706 performs a decoding process using the second encoded code obtained in separation section 702, and generates a second layer decoded residual signal. The adder 705 adds the first layer decoded signal and the second layer decoded residual signal to generate a second layer decoded signal. Then, adder 705 outputs the second layer decoded signal to US2 section 707.
[0137]
US2 section 707 upsamples the second layer decoded signal and increases the sampling frequency from Fs (2) to Fs (3). Then, US2 section 707 outputs the first layer decoded signal of sampling frequency Fs (3) to adder 709.
[0138]
Next, third layer decoding section 708 performs a decoding process using the third encoded code obtained in separation section 702, and generates a third layer decoded residual signal. Adder 709 adds the above-described second layer decoded signal and third layer decoded residual signal to generate a third layer decoded signal. Adder 709 outputs the third layer decoded signal to output terminal 710.
[0139]
Next, the n-th layer (2 ≦ n ≦ N) decoding unit will be described. FIG. 8 is a block diagram illustrating a configuration of a decoding unit of the second layer and subsequent layers of the hierarchical decoding device according to the present embodiment.
[0140]
An n-th layer (2 ≦ n ≦ N) encoded code is input from an input terminal 801. Separating section 802 separates the n-th layer (2 ≦ n ≦ N) encoded code into a time-domain encoded code and a frequency-domain encoded code. Then, demultiplexing section 802 outputs the time domain encoded code to time domain decoding section 803, and outputs the frequency domain encoded code to frequency domain decoding section 804.
[0141]
Time domain decoding section 803 generates a time domain decoded signal using the time domain encoded code, and outputs the time domain decoded signal to adder 805. The details of the time domain decoding unit 803 will be described later with reference to FIG.
[0142]
Similarly, frequency domain decoding section 804 generates a frequency domain decoded signal using the frequency domain encoded code, and outputs it to adder 805. The details of the frequency domain decoding unit 804 will be described later with reference to FIG. Adder 805 performs addition of the time domain decoded signal and the frequency domain decoded signal, and outputs the result from output terminal 806.
[0143]
Next, the time domain decoding unit 803 will be described with reference to FIG. FIG. 9 is a block diagram showing a configuration of the time domain decoding unit of the hierarchical decoding device according to the present embodiment.
[0144]
In FIG. 9, separation section 902 separates the coded code from the time-domain coded code input from input terminal 901 and outputs the coded code to adaptive codebook 903, noise codebook 904, gain codebook 905, and LPC decoder 909. Output each. LPC decoder 909 decodes the LPC coefficient using the given encoded code, and outputs the result to synthesis filter 910.
[0145]
Next, the adaptive codebook 903, the noise codebook 904, and the gain codebook 905 use an encoded code to generate an adaptive vector q (k), a noise vector c (k), and an adaptive vector gain β, respectively. _q And noise vector gain γ _q Are respectively decoded.
[0146]
The multiplier 906 multiplies the adaptive vector by the adaptive vector gain and outputs the result to the adder 908. Similarly, multiplier 907 multiplies the noise vector by the noise vector gain and outputs the result to adder 908. The adder 908 adds the multiplied adaptive vector and the noise vector to generate a driving excitation signal. When the driving sound source signal is expressed as ex (k), the driving sound source signal ex (k) is obtained as in the following equation (12).
(Equation 12)

[0147]
Next, a synthesized signal syn (k) is generated by the synthesis filter 910 using the decoded LPC coefficients and the drive excitation signal ex (k) according to the following equation (13).
(Equation 13)

Where α _q Represents the decoded LPC coefficient, and NP represents the order of the LPC coefficient. The decoded signal syn (n) thus decoded is output from the output terminal 911. After the above-described decoding processing is completed, the internal state of the adaptive codebook is updated using the latest driving excitation signal in preparation for the decoding processing in the next frame (or subframe).
[0148]
Next, the frequency domain decoding unit 804 will be described with reference to FIG. FIG. 10 is a block diagram illustrating a configuration of the frequency domain decoding unit of the hierarchical decoding device according to the present embodiment. Transform coefficient decoding section 1002 decodes the quantized transform coefficients from the frequency domain coded code input from input terminal 1001. Next, time domain transform section 1003 performs time domain transform processing on the transform coefficients obtained from transform coefficient decoding section 1002 to generate a signal in the time domain. The signal in the time domain is subjected to processing such as superposition addition so as to prevent discontinuity between frames (or subframes). Then, time domain conversion section 1003 outputs this output signal from output terminal 1004.
[0149]
As described above, according to the hierarchical decoding device of the present embodiment, a coded signal obtained by coding the difference between the signal obtained by decoding the coding result of the upper layer and the input audio signal in the time domain and the frequency domain, By decoding a periodic signal in the time domain and a non-periodic signal in the frequency domain, it is possible to perform high-quality encoding and decoding at a low bit rate.
[0150]
Further, according to the hierarchical decoding device of the present embodiment, by setting the sampling frequency of the signal to be decoded in the lower layer higher than the sampling frequency of the signal to be decoded in the upper layer, the signal is made to correspond to various sampling frequencies. The encoded signal can be decoded.
[0151]
(Embodiment 3)
FIG. 11 is a block diagram showing a configuration of a hierarchical encoding device according to Embodiment 3 of the present invention. In the present embodiment, in an audio signal encoding method in which an n-th layer (2 ≦ n ≦ N) encoding unit is composed of a time-domain encoding unit and a frequency-domain encoding unit, a pitch determined by an upper layer It is characterized in that it has a time domain encoding unit that performs encoding using a cycle.
[0152]
According to the present embodiment, by using the pitch period obtained in the upper layer, it is possible to more efficiently encode the pitch period of the time domain encoding unit, and as a result, at a low bit rate Encoding can be performed with high quality. In FIG. 11, components having the same names as those in FIG. 2 have the same functions, and thus detailed description of such components will be omitted.
[0153]
From the input terminal 1108, the pitch period T obtained in the upper layer is input. Time domain coding section 1103 performs coding using the input pitch cycle of the upper layer. FIG. 12 shows the configuration of time domain encoding section 1103 in this case. FIG. 12 is a block diagram illustrating a configuration of the time domain encoding unit of the hierarchical encoding device according to the present embodiment. In FIG. 12, components having the same names as those in FIG. 3 have the same functions, and thus detailed description of such components will be omitted.
[0154]
The pitch period T of the lower layer input from the input terminal 1218 is provided to the search candidate determining unit 1219. The search candidate determination unit 1219 limits the candidates of the adaptive vector to be searched included in the adaptive codebook 1207 based on the pitch period T of the upper layer.
[0155]
Due to the above limitation, compared to the case where all the candidates included in the adaptive codebook 1207 are to be searched, according to this method, the number of candidates for the adaptive vector to be searched is reduced. Code amount can be reduced. Further, there can be obtained an advantage that the amount of calculation required for searching the adaptive codebook can be reduced.
[0156]
The search candidate determination unit 1219 can use the pitch period T of the upper layer to search for an adaptation vector corresponding to the pitch period included in the range represented by the following equation (14). However, when the sampling frequency of the upper layer is different from the sampling frequency of the layer, the pitch period T of the upper layer is corrected and used so as to be compatible with the sampling frequency of the layer.
[Equation 14]

Here, T (n) represents the pitch cycle of the layer (n-th layer). T (m) represents the pitch period of the upper layer, and the range of m is represented by 1 ≦ m <n. ΔT1 and ΔT2 represent constants that determine the range of the pitch period. The search for the adaptive vector is performed only for the adaptive vector corresponding to the pitch period T (n) included in Expression (14). As a result of the search, the relative pitch period ΔT is determined, and this information is used as an encoded code. It is provided to the multiplexing unit 1215.
[0157]
Further, in consideration of the case where the pitch period of the upper layer is double pitch or half pitch, a search candidate for an adaptive vector included in adaptive codebook 1207 may be determined according to the following equation (15).
(Equation 15)

Here, k is a variable that represents an integer multiple or a fraction of an integer, such as k = {..., 1/4, 1/3, 1/2, 1, 2, 3, 4,. The notations ΔT1 (k) and ΔT2 (k) indicate cases where the search range of the pitch cycle may be different depending on the value of k.
[0158]
As described above, according to the hierarchical coding apparatus of the present embodiment, in time domain coding lower than the second layer, the search is performed from the adaptive vector of the adaptive codebook using the pitch period obtained in the upper layer. By limiting the candidates of the adaptive vector to be subjected to, and performing encoding using the limited adaptive vector, the encoding of the pitch period of the time domain encoding can be performed more efficiently, and at a low bit rate. Can be encoded with high quality.
[0159]
(Embodiment 4)
FIG. 13 is a block diagram showing a configuration of a hierarchical decoding device according to Embodiment 4 of the present invention. In the present embodiment, in a hierarchical coding scheme in which an n-th layer (2 ≦ n ≦ N) coding section is composed of a time domain coding section and a frequency domain coding section, a pitch period determined in an upper layer It is characterized in that the coded code generated by the time domain coding unit that performs coding by using coded data can be decoded.
[0160]
According to the present embodiment, by using the pitch period obtained in the upper layer, it is possible to more efficiently encode the pitch period of the time domain coding unit, and as a result, at a low bit rate By decoding a coded code of a hierarchical coding method capable of performing high-quality coding, an effect that a high-quality decoded signal can be obtained can be obtained.
[0161]
In FIG. 13, components having the same names as those in FIG. 8 have the same functions, and thus detailed description of such components is omitted. The pitch period T decoded in the upper layer is input from the input terminal 1307 and is provided to the time domain decoding unit 1303.
[0162]
Time domain coding section 1303 performs decoding using the input pitch cycle of the upper layer. FIG. 14 shows the configuration of time domain decoding section 1303. FIG. 14 is a block diagram showing a configuration of the time domain decoding unit of the hierarchical decoding device according to the present embodiment. In FIG. 14, components having the same names as those in FIG. 9 have the same functions, and thus detailed descriptions of such components will be omitted.
[0163]
The pitch period T decoded in the upper layer input from the input terminal 1412 is provided to the adaptive vector determination unit 1413. Further, the relative pitch period ΔT is decoded by separation section 1402 and provided to adaptive vector determination section 1413.
[0164]
Using the pitch cycle T and the relative pitch cycle ΔT of the lower layer, the adaptive vector determination unit 1413 calculates the pitch cycle T of the layer according to the following equation (16).
(Equation 16)

Here, T (n) represents the pitch period of the layer (n-th layer), and T (m) represents the pitch period of the upper layer (1 ≦ m <n). When the search candidate of the adaptive vector is determined according to the equation (15), the pitch period of the layer is calculated according to the following equation (17).
[Equation 17]

Here, k is a variable that represents an integer multiple or a fraction of an integer, such as k = {..., 1/4, 1/3, 1/2, 1, 2, 3, 4,. The pitch cycle of the layer decoded in this way is provided to adaptive codebook 1403. The adaptive codebook 1403 outputs an adaptive vector corresponding to the decoded pitch period.
[0165]
As described above, according to the speech decoding apparatus of the present embodiment, in the time domain decoding lower than the second layer, the adaptive codebook of the adaptive codebook is utilized by using the pitch period obtained in the upper layer on the encoding side. By performing the decoding while limiting the candidates of the adaptive vector used for the decoding from the adaptive vector, the encoding and the decoding of the pitch period of the time domain encoding can be performed more efficiently, and the encoding can be performed at a low bit rate and at a high bit rate. Can be encoded to quality.
[0166]
(Embodiment 5)
In the fifth embodiment, an example in which parameters input from the input terminal 1108 of the third embodiment are different will be described. In the third embodiment, the pitch period obtained in the upper layer is input, but in the present embodiment, the LPC coefficient obtained in the upper layer is input.
[0167]
In the present embodiment, in a hierarchical coding scheme in which an n-th layer (2 ≦ n ≦ N) coding section is composed of a time-domain coding section and a frequency-domain coding section, LPC coefficients obtained in an upper layer It is characterized in that it has a time-domain coding unit that performs coding by using. According to the present embodiment, by using the LPC coefficient obtained in the upper layer, it is possible to more efficiently perform encoding of the LPC coefficient of the time domain encoding unit, and as a result, at a low bit rate. Encoding can be performed with high quality. In FIG. 11, components having the same names as those in FIG. 2 have the same functions, and thus detailed description of such components will be omitted.
[0168]
In FIG. 11, LPC coefficients obtained in an upper layer are input from an input terminal 1108 and provided to a time-domain coding unit 1103. Time domain coding section 1103 performs coding using the input lower layer LPC coefficients. FIG. 15 shows the configuration of time domain encoding section 1103 in this case. FIG. 15 is a block diagram illustrating a configuration of a time-domain coding unit of the hierarchical coding device according to the present embodiment. In FIG. 15, components having the same names as those in FIG. 3 have the same functions, and thus detailed description of such components is omitted.
[0169]
The upper layer LPC coefficients input from input terminal 1518 are provided to LPC quantizer 1503. The LPC quantizer 1503 efficiently encodes the LPC coefficient of the layer given from the LPC analyzer 1502 using the LPC coefficient of the upper layer. The configuration of LPC quantizer 1503 will be described with reference to FIG. FIG. 16 is a block diagram illustrating a configuration of an LPC quantizer of the hierarchical encoding device according to the present embodiment.
[0170]
From the input terminal 1609, the LPC coefficient of the layer determined by the LPC analyzer 1502 (not shown) is input. The LPC coefficient of the layer is expressed as {αp; p = 1 to NP (n)}. Here, NP (n) represents the order of the LPC coefficient of the layer (n-th layer).
[0171]
Next, LSF conversion section 1606 converts the LPC coefficients of the frame into LSF coefficients. The LSF coefficient is a parameter that can be mutually converted with the LPC coefficient, and has advantages such as easy determination of filter stability conditions, good parameter interpolation characteristics, and substantially constant parameter sensitivity to spectral distortion. Is widely used.
[0172]
Here, when the LSF coefficient is expressed as {Fp; p = 1 to NP (n)}, the LSF coefficient takes a value between 0 and 1 and has a relationship of Fp <Fp + 1. Similarly, the LPC coefficient of the upper layer input from the input terminal 1601 is represented as {βp; p = 1 to NP (m)}. Here, NP (m) represents the order of the LPC coefficient of the upper layer (m-th layer, m <n).
[0173]
Next, LSF conversion section 1602 converts the LPC coefficient {βp; p = 1 to NP (m)} of the upper layer into LSF coefficient {Gp; p = 1 to NP (m)}. Next, the correction unit 1603 multiplies the LSF coefficient of the lower layer by a constant so as to match the sampling frequency of the layer. This constant is represented by Fs (m) / Fs (n).
[0174]
The adder 1605 adds the LSF coefficient of the lower layer after conversion given from the correction unit 1603 and the delta LSF vector stored in the delta LSF codebook 1604. Subtractor 1607 subtracts the output vector of adder 1605 from the LSF coefficient of the layer, and outputs the error signal to searcher 1608.
[0175]
The searcher 1608 efficiently searches for the delta LSF vector stored in the delta LSF codebook 1604 that minimizes the energy of the error signal or the perceptually weighted energy, and outputs the index as an encoded code. Output from terminal 1610.
[0176]
Thus, according to the hierarchical coding apparatus of the present embodiment, in the time domain coding lower than the second layer, the LPC coefficient (or LSF coefficient) obtained in the upper layer and the LPC coefficient ( Or LSF coefficient), the optimum delta LSF vector can be searched for in consideration of the LPC coefficient obtained in the upper layer, and time domain coding can be performed more efficiently. And high quality coding at a low bit rate.
[0177]
(Embodiment 6)
In the sixth embodiment, an example in which parameters input from the input terminal 1307 of the fourth embodiment are different will be described. In the fourth embodiment, the pitch period obtained in the upper layer is input, whereas in the sixth embodiment, the LPC coefficient obtained in the upper layer is input.
[0178]
In the present embodiment, in an audio signal decoding method in which an n-th layer (2 ≦ n ≦ N) decoding unit includes a time-domain decoding unit and a frequency-domain decoding unit, LPC decoded by an upper layer It is characterized in that it has a time domain decoding unit that decodes the LPC coefficients of the layer using the coefficients. According to the present embodiment, by using the LPC coefficient decoded in the lower layer, it becomes possible to decode the encoded code of the time domain encoding unit that efficiently encodes the LPC coefficient, As a result, a high-quality decoded signal can be generated at a low bit rate. In FIG. 13, components having the same names as those in FIG. 8 have the same functions, and thus detailed description of such components is omitted.
[0179]
The LPC coefficient decoded in the upper layer is input from input terminal 1307 and provided to time domain decoding section 1303. Time domain decoding section 1303 performs decoding using the input LPC coefficients of the upper layer. FIG. 17 shows the configuration of time domain decoding section 1303 in this case. FIG. 17 is a block diagram illustrating a configuration of a time domain decoding unit of the hierarchical decoding device according to the present embodiment. In FIG. 17, components having the same names as those in FIG. 9 have the same functions, and thus detailed description of such components will be omitted.
[0180]
The LPC coefficient of the upper layer input from input terminal 1712 is provided to LPC decoder 1709. The LPC decoder 1709 decodes the LPC coefficient of the layer using the LPC coefficient of the upper layer. The configuration of LPC decoder 1709 will be described with reference to FIG. FIG. 18 is a block diagram illustrating a configuration of an LPC decoder of the hierarchical decoding device according to the present embodiment.
[0181]
An LPC coefficient {βp; p = 1 to NP (m)} of an upper layer is input from an input terminal 1801. The LSF conversion unit 1807 converts the LSF coefficient of the upper layer into {Gp; p = 1 to NP (m)}. The correction unit 1803 converts the sampling frequency Fs (m) of the upper layer and the constant Fs (m) / Fs (n) defined by the sampling frequency Fs (n) of the layer into the LSF coefficient of the upper layer ｛Gp; p = 1 ＮNP (m)}, and the result is given to the adder 1805.
[0182]
From an input terminal 1802, an encoded code representing a delta LSF vector is input. The delta LSF codebook 1804 decodes the delta LSF vector using the encoded code and supplies the decoded delta LSF vector to the adder 1805. Adder 1805 adds the corrected upper layer LSF coefficient and the decoded delta LSF vector, and provides the LSF vector after addition to LPC conversion section 1808. LPC conversion section 1808 converts the LSF vector into LPC coefficients, and outputs the result from output terminal 1806.
[0183]
As described above, according to the speech decoding apparatus of the present embodiment, in time domain decoding lower than the second layer, the optimal search performed in consideration of the LPC coefficient obtained in the upper layer on the encoding side is considered. By decoding using the delta LSF vector, LPC coefficients of time domain coding can be more efficiently encoded and decoded, and high-quality encoding can be performed at a low bit rate.
[0184]
(Embodiment 7)
Next, a seventh embodiment of the present invention will be described with reference to the drawings. FIG. 19 is a block diagram showing a configuration of a communication device according to Embodiment 7 of the present invention. The feature of this embodiment is that the signal processing apparatus 1903 in FIG. 19 is configured by one of the hierarchical coding apparatuses shown in the above-described first to sixth embodiments.
[0185]
As shown in FIG. 19, a communication device 1900 according to Embodiment 7 of the present invention includes an input device 1901, an A / D conversion device 1902, and a signal processing device 1903 connected to a network 1904.
[0186]
The A / D converter 1902 is connected to the output terminal of the input device 1901. An input terminal of the signal processing device 1903 is connected to an output terminal of the A / D conversion device 1902. The output terminal of the signal processing device 1903 is connected to the network 1904.
[0187]
The input device 1901 converts a sound wave that can be heard by a human ear into an analog signal that is an electric signal, and supplies the analog signal to the A / D converter 1902. An A / D converter 1902 converts an analog signal into a digital signal and provides the digital signal to a signal processor 1903. The signal processing device 1903 encodes the input digital signal to generate a code, and outputs the code to the network 1904.
[0188]
As described above, according to the communication apparatus of the embodiment of the present invention, it is possible to enjoy the effects shown in the above-described first to sixth embodiments in communication, and to efficiently encode an audio signal with a small number of bits. An encoding device can be provided.
[0189]
(Embodiment 8)
Next, an eighth embodiment of the present invention will be described with reference to the drawings. FIG. 20 is a block diagram showing a configuration of a communication device according to Embodiment 8 of the present invention. The feature of this embodiment is that the signal processing device 2003 in FIG. 20 is configured by one of the hierarchical decoding devices shown in the above-described first to sixth embodiments.
[0190]
As shown in FIG. 20, a communication device 2000 according to Embodiment 8 of the present invention includes a receiving device 2002, a signal processing device 2003, a D / A conversion device 2004, and an output device 2005 connected to a network 2001. are doing.
[0191]
The input terminal of the receiving device 2002 is connected to the network 2001. An input terminal of the signal processing device 2003 is connected to an output terminal of the receiving device 2002. The input terminal of the D / A conversion device 2004 is connected to the output terminal of the signal processing device 2003. The input terminal of the output device 2005 is connected to the output terminal of the D / A converter 2004.
[0192]
Receiving apparatus 2002 receives a digital coded acoustic signal from network 2001, generates a digital received acoustic signal, and provides the signal to signal processing apparatus 2003. The signal processing device 2003 receives the received audio signal from the receiving device 2002, performs a decoding process on the received audio signal, generates a digital decoded audio signal, and supplies the digital decoded audio signal to the D / A conversion device 2004. The D / A conversion device 2004 converts the digital decoded audio signal from the signal processing device 2003 to generate an analog decoded audio signal, and supplies the analog decoded audio signal to the output device 2005. The output device 2005 converts an analog decoded sound signal, which is an electric signal, into vibration of air and outputs the sound as a sound wave so that the sound can be heard by a human ear.
[0193]
As described above, according to the communication device of the present embodiment, it is possible to enjoy the effects shown in the above-described first to sixth embodiments in communication, and to efficiently decode an encoded audio signal with a small number of bits. Therefore, a good acoustic signal can be output.
[0194]
(Embodiment 9)
Next, a ninth embodiment of the present invention will be described with reference to the drawings. FIG. 21 is a block diagram showing a configuration of a communication device according to Embodiment 9 of the present invention. In the ninth embodiment of the present invention, the signal processing apparatus 2103 in FIG. 21 is configured by one of the acoustic encoding units shown in the first to sixth embodiments. There is a feature of the form.
[0195]
As shown in FIG. 21, a communication device 2100 according to Embodiment 9 of the present invention includes an input device 2101, an A / D conversion device 2102, a signal processing device 2103, an RF modulation device 2104, and an antenna 2105.
[0196]
The input device 2101 converts a sound wave that can be heard by a human ear into an analog signal, which is an electric signal, and supplies the analog signal to the A / D converter 2102. The A / D converter 2102 converts an analog signal into a digital signal and supplies the digital signal to the signal processor 2103. The signal processing device 2103 encodes the input digital signal to generate an encoded audio signal, and supplies the encoded audio signal to the RF modulation device 2104. The RF modulator 2104 modulates the coded audio signal to generate a modulated coded audio signal, and provides the modulated coded audio signal to the antenna 2105. The antenna 2105 transmits the modulated and coded acoustic signal as a radio wave.
[0197]
As described above, according to the communication apparatus of the present embodiment, it is possible to enjoy the effects shown in the above-described first to sixth embodiments in wireless communication, and to efficiently encode an audio signal with a small number of bits. it can.
[0198]
Note that the present invention can be applied to a transmission device, a transmission encoding device, or an audio signal encoding device that uses an audio signal. Further, the present invention can be applied to a mobile station device or a base station device.
[0199]
(Embodiment 10)
Next, a tenth embodiment of the present invention will be described with reference to the drawings. FIG. 22 is a block diagram showing a configuration of a communication device according to Embodiment 10 of the present invention. The tenth embodiment of the present invention is characterized in that the signal processing device 2203 in FIG. 22 is configured by one of the audio decoding units shown in the first to sixth embodiments. There is a feature of the form.
[0200]
As shown in FIG. 22, a communication device 2200 according to Embodiment 10 of the present invention includes an antenna 2201, an RF demodulation device 2202, a signal processing device 2203, a D / A conversion device 2204, and an output device 2205.
[0201]
The antenna 2201 receives a digital coded acoustic signal as a radio wave, generates a digital reception coded acoustic signal of an electric signal, and supplies the generated signal to the RF demodulation device 2202. The RF demodulation device 2202 demodulates the coded audio signal received from the antenna 2201, generates a demodulated coded audio signal, and provides the demodulated coded audio signal to the signal processing device 2203.
[0202]
The signal processing device 2203 receives the digital demodulated coded audio signal from the RF demodulation device 2202, performs a decoding process, generates a digital decoded audio signal, and supplies the digital decoded audio signal to the D / A conversion device 2204. The D / A converter 2204 converts the digital decoded audio signal from the signal processor 2203 to generate an analog decoded audio signal, and supplies the analog decoded audio signal to the output device 2205. The output device 2205 converts an analog decoded audio signal, which is an electric signal, into air vibration and outputs the sound as a sound wave so that it can be heard by human ears.
[0203]
As described above, according to the communication apparatus of the present embodiment, it is possible to enjoy the effects shown in the above-described first to sixth embodiments in wireless communication, and to decode an audio signal efficiently encoded with a small number of bits. Therefore, a good acoustic signal can be output.
[0204]
Note that the present invention can be applied to a receiving device, a receiving decoding device, or an audio signal decoding device that uses an audio signal. Further, the present invention can be applied to a mobile station device or a base station device.
[0205]
Further, the present invention is not limited to the above embodiment, and can be implemented with various modifications. For example, in the above-described embodiment, the case where the processing is performed as a signal processing apparatus is described. However, the present invention is not limited to this, and the signal processing method can be performed as software.
[0206]
For example, a program for executing the signal processing method may be stored in a ROM (Read Only Memory) in advance, and the program may be operated by a CPU (Central Processor Unit).
[0207]
Further, a program for executing the above signal processing method is stored in a computer-readable storage medium, and the program stored in the storage medium is recorded in a RAM (Random Access Memory) of the computer, and the computer is operated according to the program. You may make it do.
[0208]
In the above description, a case is described in which the discrete Fourier transform is used for the method of transforming from the time domain to the frequency domain. However, the present invention is not limited to this, and any orthogonal transform can be applied. For example, discrete cosine transform or MDCT (modified discrete cosine transform) can be applied.
[0209]
Note that the present invention can be applied to a receiving device, a receiving decoding device, or an audio signal decoding device that uses an audio signal. Further, the present invention can be applied to a mobile station device or a base station device.
[0210]
【The invention's effect】
As described above, according to the hierarchical encoding method and the hierarchical decoding method of the audio signal of the present invention, in the encoding of the second layer or lower, the signal obtained by decoding the encoding result of the upper layer and the input audio signal Is encoded in the time domain, and the difference between the residual signal that cannot be encoded by encoding in the time domain, that is, the difference between the difference signal and the decoded signal obtained by decoding the encoded signal in the time domain is encoded in the frequency domain. Accordingly, a signal having periodicity can be encoded in the time domain, and a signal having no periodicity can be encoded in the frequency domain, and high-quality encoding can be performed at a low bit rate.
[Brief description of the drawings]
FIG. 1 is a block diagram showing a configuration of a hierarchical encoding device according to Embodiment 1 of the present invention.
FIG. 2 is a block diagram showing a configuration of an n-th layer (2 ≦ n ≦ N) encoding unit of the hierarchical encoding device according to the embodiment.
FIG. 3 is a block diagram showing a configuration of a time domain coding unit of the hierarchical coding device according to the embodiment.
FIG. 4 is a block diagram illustrating a configuration of a target signal generation unit of the hierarchical encoding device according to the present embodiment.
FIG. 5 is a block diagram showing a configuration of a frequency domain encoding unit of the hierarchical encoding device according to the embodiment.
FIG. 6 is a block diagram showing a configuration of an auditory masking calculation unit of the hierarchical encoding device according to the embodiment.
FIG. 7 is a block diagram showing a configuration of a hierarchical decoding device according to Embodiment 2 of the present invention.
FIG. 8 is a block diagram showing a configuration of a decoding unit of the second layer and subsequent layers of the hierarchical decoding device according to the embodiment.
FIG. 9 is a block diagram illustrating a configuration of a time domain decoding unit of the hierarchical decoding device according to the above embodiment.
FIG. 10 is a block diagram illustrating a configuration of a frequency domain decoding unit of the hierarchical decoding device according to the above embodiment.
FIG. 11 is a block diagram showing a configuration of a hierarchical encoding device according to Embodiment 3 of the present invention.
FIG. 12 is a block diagram showing a configuration of a time-domain coding unit of the hierarchical coding device according to the embodiment.
FIG. 13 is a block diagram showing a configuration of a hierarchical decoding device according to Embodiment 4 of the present invention.
FIG. 14 is a block diagram showing a configuration of a time-domain decoding unit of the hierarchical decoding device according to the embodiment.
FIG. 15 is a block diagram showing a configuration of a time-domain coding unit of a hierarchical coding device according to Embodiment 5 of the present invention.
FIG. 16 is a block diagram showing a configuration of an LPC quantizer of the hierarchical encoding device according to the embodiment.
FIG. 17 is a block diagram showing a configuration of a time-domain decoding unit of the hierarchical decoding device according to Embodiment 6 in the present invention.
FIG. 18 is a block diagram illustrating a configuration of an LPC decoder of the hierarchical decoding device according to the above embodiment.
FIG. 19 is a block diagram showing a configuration of a communication device according to a seventh embodiment of the present invention.
FIG. 20 is a block diagram showing a configuration of a communication device according to an eighth embodiment of the present invention.
FIG. 21 is a block diagram showing a configuration of a communication device according to Embodiment 9 of the present invention.
FIG. 22 is a block diagram showing a configuration of a communication device according to a tenth embodiment of the present invention.
[Explanation of symbols]
102 DS1 part
103 first layer encoding section
104, 703 First layer decoding section
107, 113, 404 delay unit
105, 704 US1
106 DS2 part
108, 114, 405 Subtractor
109 second layer encoding section
110, 704 Second layer decoding section
111, 705, 709, 805, 1305 Adder
112,707 US2
115 third layer encoding section
203 time domain coding unit
204 Target signal generator
205 frequency domain coding unit
403 decryption unit
503 Frequency domain transform unit
504 Auditory masking calculation unit
505 Quantization unit
708 Third layer decoding unit
803, 1303 time domain decoding unit
804, 1304 frequency domain decoding unit
1002 Transform coefficient decoding unit
1003 Time domain converter

Claims (28)

A hierarchical encoding method for encoding an input audio signal, decoding a signal encoded in an upper layer, and encoding a difference between the decoded signal and the input signal, wherein the input audio signal is a frame unit having a predetermined length. And a second encoding step of encoding the difference between the signal obtained by decoding the encoding result of the upper layer and the input audio signal in one or more stages, In the second encoding step, a difference between a signal obtained by decoding an encoding result of an upper layer and an input audio signal is encoded in both a time domain and a frequency domain. The second encoding step includes a hierarchical decoding step of decoding a signal encoded in an upper layer to generate a decoded signal; an upsampling step of increasing a sampling frequency of the decoded signal; 2. The hierarchical encoding method according to claim 1, further comprising: a hierarchical subtraction step of generating a difference signal by subtracting the difference signal; and a hierarchical encoding step of encoding the difference signal. The hierarchical encoding step is a time-domain encoding step of encoding a difference between a signal obtained by decoding an encoding result of an upper layer and an input audio signal in a time domain, and the time-domain encoding step A time domain signal decoding step of decoding a signal to generate a decoded signal; a time domain signal subtracting step of subtracting the decoded signal from the difference signal to generate a second difference signal; 3. The hierarchical encoding method according to claim 2, further comprising a frequency domain encoding step of encoding. The time domain encoding step includes a search candidate determining step of limiting an adaptive vector to be used in encoding from a pitch period obtained in an upper layer, and searching an adaptive vector having the smallest difference from an input speech signal from the limited adaptive vector. 4. The hierarchical encoding method according to claim 3, further comprising: The time domain encoding step includes a quantization step of quantizing a pitch cycle, and the search candidate determining step modifies the pitch cycle obtained in an upper layer so as to match the sampling frequency of the layer, The hierarchical encoding method according to claim 4, wherein the quantization step quantizes a pitch period of the layer using the corrected pitch period. The time domain coding step includes an addition step of adding a parameter of the coding obtained in an upper layer and a parameter that is a search candidate of the layer, and a parameter having a smallest difference between a result of the addition step and an input audio signal. 4. The hierarchical encoding method according to claim 3, further comprising a search step of searching for The time domain coding step includes a conversion step of converting LPC coefficients obtained in an upper layer into LSF coefficients, and the adding step holds the LSF coefficient and the LSF codebook converted in the conversion step. The LSF coefficient that minimizes the difference between the added LSF coefficient and the LSF coefficient obtained from the input audio signal is searched for in the search step, wherein the LSF coefficient is determined. Hierarchical encoding method. An auditory masking step of calculating an auditory masking from an input acoustic signal, wherein the frequency domain encoding step encodes a second differential signal using a signal after the masking using the auditory masking. The hierarchical encoding method according to claim 3. A hierarchical decoding method for encoding an input audio signal, decoding a signal encoded in an upper layer, and decoding a signal in which a difference between the decoded signal and the input signal is encoded, comprising: A first decoding step of decoding a code, a second decoding step of decoding a coded code of a layer lower than a second layer in both a time domain and a frequency domain, and the first decoding step and the second decoding step. An addition step of adding a decoding result. An up-sampling step of up-sampling a sampling frequency of a decoding result of the first decoding step to a sampling frequency of a decoding result of the second decoding step, wherein the adding step includes the steps of: The hierarchical decoding method according to claim 9, wherein the decoding result of the second decoding step is added. 11. The method according to claim 9, wherein the second decoding step includes a time domain decoding step of decoding the layer using a pitch cycle or an LPC coefficient of an upper layer in time domain decoding. The hierarchical decoding method described in the above. 12. The method according to claim 9, wherein the second decoding step includes a frequency domain decoding step of decoding an encoded code obtained by encoding a frequency domain coefficient using auditory masking. 2. The hierarchical decoding method according to item 1. A hierarchical encoding device that encodes an input audio signal, decodes a signal encoded by an upper layer, and encodes a difference between the decoded signal and the input signal, wherein the input audio signal is a frame unit having a predetermined length. And a second encoder for encoding the difference between the signal obtained by decoding the encoding result of the upper layer and the input audio signal in one or more stages, The hierarchical encoding device according to claim 2, wherein the second encoding unit encodes a difference between a signal obtained by decoding an encoding result of an upper layer and an input audio signal in both a time domain and a frequency domain. The second encoding means decodes a signal encoded in an upper layer to generate a decoded signal; a hierarchical decoding means for increasing a sampling frequency of the decoded signal; an upsampling means for increasing a sampling frequency of the decoded signal; 14. The hierarchical encoding apparatus according to claim 13, further comprising: a hierarchical subtraction unit that generates a difference signal by subtracting the difference signal from the input signal; and a hierarchical encoding unit that encodes the difference signal. The hierarchical encoding unit includes a time domain encoding unit that encodes, in a time domain, a difference between a signal obtained by decoding an encoding result of an upper layer and an input audio signal, and a time domain encoding unit that encodes the difference. A time-domain signal decoding unit that decodes the signal to generate a decoded signal; a time-domain signal subtraction unit that subtracts the decoded signal from the difference signal to generate a second difference signal; 15. The hierarchical coding apparatus according to claim 14, further comprising: frequency domain coding means for coding. The time domain encoding means includes: a codebook for holding adaptive vectors generated in the past; a search candidate determining means for limiting adaptive vectors to be used in encoding from a pitch period obtained in an upper layer; and a limited adaptive vector. 16. The hierarchical coding apparatus according to claim 15, further comprising: searching means for searching for an adaptive vector having the smallest difference from the input speech signal. The time domain encoding means includes a quantization means for quantizing a pitch cycle, and the search candidate determining means modifies the pitch cycle obtained in an upper layer so as to match a sampling frequency of the layer, 17. The hierarchical coding apparatus according to claim 16, wherein the quantization means quantizes the pitch cycle of the layer using the corrected pitch cycle. The time domain encoding means includes an adding means for adding a parameter for encoding obtained in an upper layer and a parameter which is a search candidate for the layer, and a parameter having a smallest difference between a result of the adding means and an input audio signal. 16. The hierarchical encoding apparatus according to claim 15, further comprising: a search unit that searches for. The time domain coding unit includes an LSF codebook that holds LSF coefficients, and a conversion unit that converts LPC coefficients obtained in an upper layer into LSF coefficients, and the adding unit converts the LPC coefficients obtained by the conversion unit. The LSF coefficient added to the LSF coefficient held by the LSF codebook is added to the LSF coefficient, and the search means searches for an LSF coefficient that minimizes the difference between the added LSF coefficient and the LSF coefficient obtained from the input audio signal. 19. The hierarchical encoding device according to claim 18, wherein: An audio masking means for calculating an audio masking from an input audio signal, wherein the frequency domain encoding means encodes a second difference signal using a signal after the audio masking using the audio masking. 20. The hierarchical encoding device according to claim 15. A hierarchical decoding device that encodes an input audio signal, decodes a signal encoded in an upper layer, and decodes a signal obtained by encoding a difference between the decoded signal and the input signal. A first decoding unit that decodes a code, a second decoding unit that decodes an encoded code of a layer lower than a second layer in both a time domain and a frequency domain, and the first decoding unit and the second decoding unit. A hierarchical decoding device comprising: an adding unit that adds a decoding result. Up-sampling means for up-sampling the sampling frequency of the decoding result of the first decoding means to the sampling frequency of the decoding result of the second decoding means, wherein the adding means includes the first decoding means after up-sampling and the 22. The hierarchical decoding device according to claim 21, wherein the decoding result of the second decoding means is added. 23. The method according to claim 21, wherein the second decoding unit includes a time-domain decoding unit that decodes the layer using a pitch cycle or an LPC coefficient of an upper layer in the decoding of the time domain. A hierarchical decoding device according to claim 1. 24. The apparatus according to claim 21, wherein the second decoding unit includes a frequency domain decoding unit that decodes an encoded code obtained by encoding a frequency domain coefficient using auditory masking. 2. The hierarchical decoding device according to item 1. Audio input means for converting an audio signal into an electrical signal, A / D conversion means for converting a signal output from the audio input means into a digital signal, and encoding a digital signal output from the A / D conversion means 21. A hierarchical encoding device according to claim 13, wherein said encoding device outputs an encoded code output from said encoding device to a radio frequency signal. A transmitting antenna for converting an output signal into a radio wave and transmitting the radio wave; The receiving antenna according to any one of claims 21 to 24, wherein the receiving antenna receives a radio wave, RF demodulating means demodulates a signal received by the receiving antenna, and information obtained by the RF demodulating means is decoded. Hierarchical decoding device, D / A conversion means for converting a signal output from the decoding device into an analog signal, and audio output means for converting an electric signal output from the D / A conversion means into an audio signal And a sound signal receiving device. 27. A communication terminal device comprising at least one of the acoustic signal transmitting device according to claim 25 and the acoustic signal receiving device according to claim 26. A base station device comprising at least one of the acoustic signal transmitting device according to claim 25 and the acoustic signal receiving device according to claim 26.

Images