JP3951690B2 - Encoding apparatus and method, and recording medium - Google Patents

Abstract

PROBLEM TO BE SOLVED: To perform encoding with higher efficiency while deterring a listener from feeling physical disorder. SOLUTION: An adaptive mixing part 102 performs a mixing process for inputted signals Ln(t) and Rn(t) according to distortion rate information En(f) reported from a distortion rate detection part 106 and limits the operation times of MS stereo encoding and IS stereo encoding. Further, the adaptive mixing part 102 generates power correction information Pn,adj(t) corresponding to a mixing coefficient to make power corrections when reproduction is performed. An encoding control part 104 selects an encoding system for an encoding process that an encoding part 105 performs and indicates it to the encoding part 105. The encoding part 105 selects dual encoding, MS stereo encoding, or IS stereo encoding according to the indication from the encoding control part 104 and encodes a spectrum signal supplied from an area conversion part 103.

Description

【数２】

これらの式において、f₀は正規化された範囲の先頭スペクトル番号を、f₁は終端スペクトルの番号を、それぞれ示している。
【０１０６】
正規化スペクトル信号l_n(f)と正規化スペクトル信号r_n(f)が似ている（相関が高い）ほど、総和値Sが大きく、総和値Dが小さくなる。逆に、正規化スペクトル信号l_n(f)と正規化スペクトル信号r_n(f)が異なる（相関が低い）場合には、総和値Sと総和値Dは同じような値となるため、総和値S，Dの比(総和値比S/D)を算出することにより、正規化スペクトル信号l_n(f)と正規化スペクトル信号r_n(f)の相関関係を取得することができる。例えば、総和値比S/Dは、その値が「1」より大きい場合には、正規化スペクトル信号l_n(f)と正規化スペクトル信号r_n(f)の相関が高いことを示している。
【０１０７】
そして、比較部１４３は、ステップS５４において、ステップS５３で算出した総和値比S/Dが、分割された周波数帯域、または正規化された小領域毎に予め設定されている許容誤差レベル（閾値）Thrより小さいか否かを判定する。比較部１４３は、総和値比S/Dが許容誤差レベルThrより小さいと判定したとき、ステップS５５に進み、符号化部１０５に入力されているスペクトル信号L_n(f)，R_n(f)を、デュアル符号化により符号化することを選択し、符号化部１０５に通知する。すなわち、許容誤差レベルは、総和値比S/Dが所定のレベル以上であれば（正規化スペクトル信号l_n(f)と正規化スペクトル信号r_n(f)に所定レベル以上の相関があれば）、強制的にMSまたはISステレオ符号化により符号化すべく設定される。なお、本実施例では、上記総和値ＳおよびＤの比を用いて正規化スペクトル信号l_n(f)と正規化スペクトル信号r_n(f)の相関を判定しているが、相関の判定方法は当然これに限られるものではなく、l_n(f)とr_n(f)の絶対値を比較して相関係数を得るなど、他のパラメータを用いて判定してもよい。
【０１０８】
一方、ステップS５４において、比較部１４３は、総和値比S/Dが許容誤差レベルThrより大きいと判定したとき、その旨を比較部１４４に通知する。そして、比較部１４４は、ステップS５６において、対象としている帯域のスペクトルに対するd_n(f)の最大値が、復号装置１５１で実現可能な量子化精度レベルThqより大きいか否かを判定する。すなわち、比較部１４４は、差信号d_n(f)を符号化する必要がある場合にはMSステレオ符号化を選択し、符号化する必要がない場合にはISステレオ符号化を選択する。
【０１０９】
比較部１４４は、ステップS５６において、d_n(f)の最大値が量子化精度レベルThqより大きいと判定した場合、ステップS５７に進み、符号化部１０５に入力されるスペクトル信号L_n(f)，R_n(f)を、MSステレオ符号化により符号化することを選択し、符号化部１０５に通知する。また、比較部１４４は、ステップS５６で、d_n(f)の最大値が量子化精度レベルThqより小さいと判定した場合、ステップS５８に進み、符号化部１０５に入力されるスペクトル信号L_n(f)，R_n(f)を、ISステレオ符号化により符号化することを選択し、符号化部１０５に通知する。
【０１１０】
これにより、たとえ、正規化スペクトル信号l_n(f)と正規化スペクトル信号r_n(f)は高い相関があり、MSまたはISステレオ符号化よりも、デュアル符号化の方が、より高いSNRを実現できる可能性があっても、総和値比S/Dが、聴感上、ノイズとして聞き取ることができない閾値より高い場合には、MSまたはISステレオ符号化により入力信号が符号化される。
【０１１１】
また、差信号d_n(f)が符号化されない場合であっても、左右信号の正規化に関する情報は符号化されているため、ISステレオ符号化も、MSステレオ符号化と等価と考えることができる。これにより、MSステレオ符号化を行う処理部と、ISステレオ符号化を行う処理部を別々に設けることがなくなり、符号化装置９１を、より小さく構成することができる。
【０１１２】
なお、許容誤差レベルThrは、領域変換のブロック長や、ビットアロケーションなどの符号化システムの構成に応じて設定される。また、量子化精度レベルThqは、符号化装置９１で実現可能な最高量子化精度レベルを設定してもよいし、周波数帯域毎に量子化精度レベルThq(f)を設定するようにしてもよい。すなわち、量子化精度レベルThqも、許容誤差レベルThrと同様に、システムに応じて設定される。
【０１１３】
図１８は、符号化装置９１におけるセパレーションと信号対雑音比SNRとの関係（図１８(A)）、符号化（量子化）された信号の信号対雑音比SNRの時間的変化（図１８(B)）、MSステレオ符号化の動作時間確率P_MS、またはISステレオ符号化の動作時間確率P_ISの時間的変化（図１８(C)）、および、左右信号のセパレーションの時間的変化（図１８(D)）を示す図である。
【０１１４】
図１８(B)および図１８(C)に示すように、信号対雑音比SNRとMSステレオ符号化の動作時間確率P_MS、またはISステレオ符号化の動作時間確率P_ISは連動しているため、上述したようにミキシング係数を適宜変動させることにより、P_MS，P_ISを制御して、SNRを向上させることができる。これによって、音質を向上させることができる。
【０１１５】
また、図１８(A)に示すように、SNRが向上するにつれ、左右信号のセパレーションは悪化する（モノラル化する）ため、図１８(A)に示すSNRの変動に合わせて、セパレーションは、図１８(D)に示すように悪化するが、上述したように、パワー補正情報P_n,_adj(t)を生成し、復号する際にパワー調整を行うようにしたので、左右信号のセパレーションも改善することができる。なお、図１８(B)，(C)，(D)において、線L₁，L₃，L₅は、図８の符号化装置９１の特性を示し、線L₂，L₄，L₆は、従来の符号化装置の特性を示している。
【０１１６】
上述した例においては、符号化の歪み率を検出し、その値に応じてミキシング係数を設定し、次のタイミングの入力信号をミキシングするとしたが、歪み率が所定の閾値以下となるまで、所定の帯域の入力信号が、繰り返しミキシングされるような構成にしてもよい。この場合、フィルタバンク１０１―１により生成された信号L_n(t)およびフィルタバンク１０１−２により生成された信号Ｒ_n(t)は図示しないメモリ等に蓄積され、適応ミキシング部１０２にフィードバックされた歪み率情報E_n(f)により再度ミキシング、領域変換および符号化が施される。
【０１１７】
図１９は、本発明を適用した復号装置の構成例を示すブロック図である。
【０１１８】
デマルチプレクサ１６１は、図示せぬ伝送路を介して供給された符号列Cを所定の帯域毎の符号列C_n，C_n-1，…，C₁に分解し、それぞれの符号列C_iを対応する復号部（説明の便宜上、復号部１６２だけが示されている）に出力する。符号列C_nは、復号部１６２に供給される。
【０１１９】
復号部１６２は、入力された符号列C_nを、符号化方式に対応する復号方式により復号し、得られたスペクトル信号L'_n(f)を領域変換部１６３−１に、スペクトル信号R'_n(f)を領域変換部１６３−２に、それぞれ出力する。また、復号部１６２は、符号列C_nから取得したパワー補正情報P_n,_adj(t)をパワーウエイティング部１６４に供給する。
【０１２０】
領域変換部１６３は、入力されたスペクトル信号L'_n(f)，R'_n(f)に対して逆MDCTなどを利用して時間領域の信号に変換し、得られた信号L'_n(t)_mix，R'_n(t)_mixをパワーウエイティング部１６４に出力する。
【０１２１】
パワーウエイティング部１６４は、通知されてきたパワー補正情報P_n,_adj(t)に含まれるパワーウエイティング係数に基づいて、領域変換部１６３から供給されてきた信号L'_n(t)_mix，R'_n(t)_mixのパワー補正を実行し、生成した信号L'_n(t)をフィルタバンク１６５−１に、信号R'_n(t)をフィルタバンク１６５−２に、それぞれ出力する。
【０１２２】
フィルタバンク１６５は、パワーウエイティング部１６４から供給されてきた信号L'_n(t)，R'_n(t)を、他の帯域の信号L'_n-1(t)，…，L'₁(t)，R'_n-1(t)，…，R'₁(t)と合成し、生成した全帯域のオーディオ信号L'(t)，R'(t)を復号装置１５１の外部に出力する。
【０１２３】
図２０は、パワーウエイティング部１６４の詳細な構成例を示すブロック図である。
【０１２４】
パワーウエイティング係数設定部１７１は、供給されたパワー補正情報P_n,_adj(t)に含まれるパワーウエイティング係数cを乗算器１７２−１に、パワーウエイティング係数dを乗算器１７２−２に、それぞれ設定する。
【０１２５】
乗算器１７２−１は、入力された信号L'_n(t)_mixに対してパワーウエイティング係数cを乗算し、乗算器１７２−２は、入力されてきた信号R'_n(t)_mixに対してパワーウエイティング係数dを乗算し、得られた信号L'_n(t)，R'_n(t)がフィルタバンク１６５−１，１６５−２に出力される。
【０１２６】
図２１は、乗算器１７２−１の詳細な構成例を示すブロック図である（図示は省略するが、乗算器１７２−２も同様に構成される）。
【０１２７】
この例においては、バッファ１７２A，１７２Bが設けられており、現在（時刻t=0）、設定されているパワーウエイティング係数c(t₀)がバッファ１２４Aに、直前に設定されていた（時刻t=1のときに設定されていた）パワーウエイティング係数c(t₁)がバッファ１７２Bに、それぞれ記憶されるようになされている。
【０１２８】
すなわち、パワーウエイティング係数c(t)を変化させた場合、そのときに出力される信号に不連続点が生じることがある。そのため、図２２の線（１）乃至（３）に示すように、直線的または曲線的にパワーウエイティング係数c(t)を変化させることにより、不連続点が発生することを抑制することができる。なお、この例においては、バッファが２つ設けられることとしているが、３つ以上設けられるようにしてもよい。また、それぞれのパワーウエイティング係数を補間する補間関数は、１次でもよいし、２次または３次関数などにより補間するようにしてもよい。
【０１２９】
次に、図２３のフローチャートを参照して、図１９の復号装置１５１の復号処理について説明する。
【０１３０】
ステップS７１において、デマルチプレクサ１６１は、入力された符号列Cを所定の帯域数nの符号列C_n，C_n-1，…，C₁に分割し、対応する復号部に出力する。
【０１３１】
復号部１６２は、ステップS７２で、正規化情報、量子化精度情報、符号帳番号などの組合せから復号方式を選択し、入力されてきた符号列C_nを復号し、得られたスペクトル信号L'_n(f)を領域変換部１６３−１に、スペクトル信号R'_n(f)を領域変換部１６３−２に、それぞれ出力する。また、復号部１６２は、符号列C_nから取得したパワー補正情報P_n,_adj(t)をパワーウエイティング部１６４に出力する。
【０１３２】
ステップS７３において、領域変換部１６３−１，１６３−２は、入力されたスペクトル信号L'_n(f)，R'_n(f)に対して逆MDCTなどを利用して時間領域の信号に変換し、得られた信号L'_n(t)_mix，R'_n(t)_mixをパワーウエイティング部１６４に出力する。信号L_n'(t)_mix，R_n'(t)_mixは、符号化装置９１においてミキシングが施された可能性のある信号であり、ミキシング係数の設定によって、本来ステレオ信号であったものが、ほとんどモノラル信号となっている場合もある。そのため、ステップS７４において、パワーウエイティング部１６４は、通知されてきたパワー補正情報P_n,_adj(t)に基づいて、パワーウエイティング処理を実行して擬似的なステレオ信号を再生する。パワーウエイティング処理の詳細は、図２４のフローチャートを参照して後述する。
【０１３３】
パワーウエイティング処理により得られた信号L'_n(t)，R'_n(t)は、それぞれフィルタバンク１６５−１，１６５−２に出力される。以上の処理は、各帯域毎に行われる。
【０１３４】
そして、フィルタバンク１６５は、ステップS７５で、パワーウエイティング部１６４から供給された信号L'_n(t)，R'_n(t)を、他の帯域の信号L'_n-1(t)，…，L'₁(t)，R'_n-1(t)，…，R'₁(t)と合成し、合成した全帯域のオーディオ信号L'(t)，R'(t)を、復号装置１５１の外部に出力する。
【０１３５】
次に、図２４のフローチャートを参照して、図２３のステップS７４において実行されるパワーウエイティング処理について説明する。
【０１３６】
ステップS９１において、パワーウエイティング係数設定部１７１は、復号部１６２から通知されたパワー補正情報P_n,_adj(t)に含まれるパワーウエイティング係数に基づいて、乗算器１７２−１，１７２−２のパワーウエイティング係数c，dを設定する。
【０１３７】
ステップS９２において、乗算器１７２−１，１７２−２は、それぞれ、パワーウエイティング係数c，dを、入力された信号L'_n(t)_mix，R'_n(t)_mixに乗算し、生成した信号L'_n(t)，R'_n(t)をフィルタバンク１６５−１，１６５−２にそれぞれ出力する。
【０１３８】
例えば、上述したように、パワー補正部１２３において、パワー補正情報P_n,_adj(t)（パワーウエイティング係数）がc=1.25，d=0.05とされ、パワーウエイティング係数設定部１７１により、それぞれのパワーウエイティング係数c，dが設定されている場合、乗算器１７２−１は、入力信号L'_n(t)_mixのパワーを1.25倍し、生成した信号L'_n(t)をフィルタバンク１６５−１に出力する。また、乗算器１７２−２は、入力信号R'_n(t)_mixのパワーを0.05倍し、生成した信号R'_n(t)をフィルタバンク１６５−２に出力する。
【０１３９】
これにより、符号化する際に、左右の信号のセパレーションが悪化した場合であっても擬似的なステレオ信号を再生することができる。
【０１４０】
上述した一連の処理は、ハードウェアにより実行させることもできるが、ソフトウェアにより実行させることもできる。この場合、例えば、符号化装置９１は、図２５に示されるようなパーソナルコンピュータ１８１により構成される。
【０１４１】
図２５において、CPU（Central Processing Unit）１９１は、ROM（Read Only Memory）１９２に記憶されているプログラム、または、記憶部１９８からRAM（Random Access Memory）１９３にロードされたプログラムに従って各種の処理を実行する。RAM１９３にはまた、CPU１９１が各種の処理を実行する上において必要なデータなどが適宜記憶される。
【０１４２】
CPU１９１、ROM１９２、およびRAM１９３は、バス１９４を介して相互に接続されている。このバス１９４にはまた、入出力インタフェース１９５も接続されている。
【０１４３】
入出力インタフェース１９５には、キーボード、マウスなどよりなる入力部１９６、CRT，LCDなどよりなるディスプレイ、並びにスピーカなどよりなる出力部１９７、ハードディスクなどより構成される記憶部１９８、モデム、ターミナルアダプタなどより構成される通信部１９９が接続されている。通信部１９９は、ネットワークを介しての通信処理を行う。
【０１４４】
入出力インタフェース１９５にはまた、必要に応じてドライブ２００が接続され、磁気ディスク２０１、光ディスク２０２、光磁気ディスク２０３、あるいは半導体メモリ２０４などが、適宜、装着され、それから読み出されたコンピュータプログラムが、必要に応じて記憶部１９８にインストールされる。
【０１４５】
一連の処理をソフトウエアにより実行させる場合には、そのソフトウエアを構成するプログラムが、専用のハードウエアに組み込まれているコンピュータ、または、各種のプログラムをインストールすることで、各種の機能を実行することが可能な、例えば、汎用のパーソナルコンピュータ１８１などに、ネットワークや記録媒体からインストールされる。
【０１４６】
この記録媒体は、図２５に示すように、装置本体とは別に、ユーザにプログラムを提供するために配布される、プログラムが記録されている磁気ディスク２０１（フロッピディスクを含む）、光ディスク２０２（CD-ROM，DVD(Digital Versatile Disk)を含む）、光磁気ディスク２０３（MD(Mini-Disk)を含む）、もしくは半導体メモリ２０４などよりなるパッケージメディアにより構成されるだけでなく、装置本体に予め組み込まれた状態でユーザに提供される、プログラムが記録されているROM１９２や、記憶部１９８に含まれるハードディスクなどで構成される。
【０１４７】
なお、本明細書において、記録媒体に記録されるプログラムを記述するステップは、記載された順序に従って時系列的に行われる処理はもちろん、必ずしも時系列的に処理されなくとも、並列的あるいは個別に実行される処理をも含むものである。
【０１４８】
【発明の効果】
本発明の符号化装置および方法、並びに記録媒体のプログラムによれば、入力信号に基づいて符号化方式を選択し、選択した符号化方式に基づいて、入力信号を符号化し、検出した歪み率に対応して決定されるミキシング割合に基づいて、入力信号の左右成分をミキシングする。そして、ミキシングした入力信号に基づいて、符号化方式を選択するようにしたので、より高効率で、オーディオ信号を符号化することができる。
【図面の簡単な説明】
【図１】 MSステレオ符号化による従来のオーディオ信号伝送システムの構成例を示すブロック図である。
【図２】 ISステレオ符号化による従来のオーディオ信号伝送システムの構成例を示すブロック図である。
【図３】従来の符号化装置の構成例を示すブロック図である。
【図４】従来の他の符号化装置の構成例を示すブロック図である。
【図５】従来のさらに他の符号化装置の構成例を示すブロック図である。
【図６】従来の符号化装置の処理を説明するフローチャートである。
【図７】従来の符号化装置の動作と生成する信号の関係を示す図である
【図８】本発明を適用した符号化装置の構成例を示すブロック図である。
【図９】図８の適応ミキシング部の構成例を示すブロック図である。
【図１０】図９のミキシング係数設定部に記憶される情報の例を示す図である。
【図１１】図９のパワー補正部に記憶される情報の例を示す図である。
【図１２】図９の乗算器の構成例を示す図である。
【図１３】ミキシング係数の補間関数の例を示す図である。
【図１４】図８の符号化制御部の構成例を示すブロック図である。
【図１５】図８の符号化装置の処理を説明するフローチャートである。
【図１６】図１５のステップS１２において実行される処理の詳細を説明するフローチャートである。
【図１７】図１５のステップS１４において実行される処理の詳細を説明するフローチャートである。
【図１８】図８の符号化装置の動作と生成する信号の関係を示す図である
【図１９】本発明を適用した復号装置の構成例を示すブロック図である。
【図２０】図１９のパワーウエイティング部の構成例を示すブロック図である。
【図２１】図２０の乗算器の構成例を示すブロック図である。
【図２２】パワーウエイティング係数の補間関数の例を示す図である。
【図２３】図１９の復号装置の処理を説明するフローチャートである。
【図２４】図２３のステップS７４において実行される処理の詳細を説明するフローチャートである。
【図２５】パーソナルコンピュータの構成例を示すブロック図である。
【符号の説明】
９１ 符号化装置， １０１−１および１０１−２ フィルタバンク， １０２ 適応ミキシング部， １０３−１および１０３−２ 領域変換部， １０４符号化制御部， １０５ 符号化部， １０６ 歪み率検出部， １０７ マルチプレクサ， １２１ パワー算出部， １２２ ミキシング係数設定部， １２３ パワー補正部， １２４−１および１２４−２ 乗算器， １２４Aおよび１２４B バッファ， １２５−１および１２５−２ 乗算器， １２６−１および１２６−２ 加算器， １４１−１および１４１−２ 正規化部， １４２−１ 加算器， １４２−２ 減算器， １４３ 比較部， １４４ 比較部， １５１ 復号装置， １６１ デマルチプレクサ， １６２ 復号部， １６３−１および１６３−２ 領域変換部， １６４ パワーウエイティング部， １６５−１および１６５−２ フィルタバンク， １７１ パワーウエイティング係数設定部， １７２−１および１７２−２ 乗算器， １７２Aおよび１７２B バッファ[0001]
BACKGROUND OF THE INVENTION
The present invention relates to an encoding apparatus and method. , As well as recording media, especially with higher efficiency, Can be encoded Coding apparatus and method , And a recording medium.
[0002]
[Prior art]
In recent years, so-called “perceptual audio encoders (decoders)” have been developed, and a bit rate of about one-twelfth or less of the bit rate generally used in conventional CD-ROM (Compact Disk-Read Only Memory). Thus, transmission and storage of high-quality audio signals is possible.
[0003]
Such an encoder encodes using a waveform portion included in an audio signal that cannot be heard due to limitations of the human auditory system. For stereo audio signals, for example, MS stereo encoding is used. An encoder using (intermediate / side stereo encoding) and an encoder using IS stereo encoding (intensity stereo encoding) are known.
[0004]
FIG. 1 is a block diagram showing a configuration example of a conventional audio signal transmission system using MS stereo coding.
[0005]
The left signal L and the right signal R constituting the stereo audio signal are input to the arithmetic unit 1, added by the adder 1-1, and output to the multiplier 1-2. On the other hand, a difference signal between these signals is generated in the subtractor 1-3 and output to the multiplier 1-4. In multipliers 1-2 and 1-4, the output of adder 1-1 and subtractor 1-3 is multiplied by coefficient x to generate sum signal M and difference signal S, which are encoded by encoding unit 2. Thus, the data is output to the transmission path 3 including a recording medium or a network.
[0006]
The decoding unit 4 decodes the input code string to generate a sum signal M ′ and a difference signal S ′. The sum signal M ′ and the difference signal S ′ are added by an adder 5-1, multiplied by a coefficient y by a multiplier 5-2, and output as a left signal L ′. Further, the sum signal M ′ and the difference signal S ′ are subtracted by the subtractor 5-3, multiplied by the coefficient y by the multiplier 5-4, and output as the right signal R ′. For example, the coefficient x is 0.5, and the coefficient y is 1.0.
[0007]
It is the sum signal rather than the difference signal that affects human hearing. Thus, the sum signal M and the difference signal S are generated, and more data (number of bits) is assigned to the sum signal M. Thus, encoding can be performed with higher efficiency than encoding each of them independently (dual encoding). Note that MS stereo coding is effective for signals in a low frequency band.
[0008]
FIG. 2 is a block diagram showing a configuration example of a conventional audio signal transmission system using IS stereo coding.
[0009]
The right signal L and the left signal R input to the calculation unit 11 are added by an adder 11-1, and an intensity signal I obtained by the correlation between these signals is generated. Also, a power left signal P indicating the power of each of the left signal L and the right signal R _l , Power right signal P _r (Scaling signal describing the content of energy) is generated by the calculation unit 11. Intensity signal I, power left signal P _l , And power right signal P _r Is input to the encoding unit 12, encoded, and then output to the transmission path 13.
[0010]
The decoding unit 14 decodes the input signal, and obtains the obtained intensity signal I ′, power left signal P _l ', And power right signal P _r 'Is output to the calculation unit 15. In the arithmetic unit 15, the multiplier 15-1 includes an intensity signal I ′ and a power left signal P ′. _l , And the multiplier 15-2 is connected to the intensity signal I ′ and the power right signal P ′. _r Based on the above, the left signal L ′ and the right signal R ′ are reproduced and output to the outside.
[0011]
By encoding using IS stereo encoding, it is possible to use the characteristic that the position detection capability due to the human auditory time difference is lower for higher frequency signals. For example, in the high frequency band, the left and right signals are It is possible to encode at a data rate of about half compared to the case of encoding each independently.
[0012]
By the way, MS stereo coding and IS stereo coding do not always have the same effect for all input signals. For example, MS stereo coding is an effective means only when the difference signal S has less energy than the sum signal M. Otherwise, the left signal is obtained from the sum signal M ′ and the difference signal S ′. When playing back L 'and right signal R', the quantization noise generated by encoding or decoding (quantization / inverse quantization) may cause mutual interference, resulting in audible noise. .
[0013]
Also, in IS coding, a spectrum SP obtained by synthesizing high-frequency components of a stereo signal and converting it from the time domain to the frequency domain _m And the original power spectrum P _l , P _r If there is no high correlation in the envelope shape, for example, if the left signal L is a trumpet signal and the right signal R is a cymbal signal, the positional relationship of each sound source (instrument) should be preserved May not be able to be heard, and noise that is clearly audible may be generated.
[0014]
Therefore, as shown in FIG. 3, FIG. 4, and FIG. 5, dual encoding that encodes the left and right signals independently and MS or IS stereo encoding are combined, and encoding is appropriately performed according to the input signal. An encoding device that selects a method is considered.
[0015]
FIG. 3 is a block diagram illustrating a configuration example of a conventional encoding device that encodes an input signal in the time domain.
[0016]
The filter bank 31-1 converts the input left signal L (t) into a signal L in a predetermined frequency band. _n (t), L _n-1 (t), ..., L ₁ (t) (n is the number of divided bands), and each signal is output to the corresponding dual encoding unit 32 and MS / IS encoding unit 33. In FIG. 3, the signal L _n Only the dual encoder 32 and the MS / IS encoder 33 for processing (t) are shown, but the signal L _n-1 (t), L _n-2 (t), ..., L ₁ Each encoding unit corresponding to (t) is L _n It is provided in the same manner as the encoding unit that processes (t).
[0017]
Similarly to the filter bank 31-1, the filter bank 31-2 also converts the right signal R (t) into a signal R in a predetermined frequency band. _n (t), R _n-1 (t), ..., R ₁ The signal is divided into (t), and the respective signals are output to the dual encoding unit 32 and the MS / IS encoding unit 33. In the following description, the filter bank 31-1 and the filter bank 31-2 are collectively referred to as a filter bank 31 when it is not necessary to distinguish them individually. The same applies to other devices.
[0018]
The dual encoding unit 32 encodes the input signal by the dual encoding method (left signal L _n (t) and right signal R _n (t) is independently encoded), and the obtained data is output to the switch 35. The dual encoding unit 32 also includes necessary bit number information B, which is information related to the amount of encoded data. _n (t) ₁ , And distortion rate information E, which is information about distortion rate with sine wave when encoding _n (t) ₁ Is generated and supplied to the encoding control unit 34.
[0019]
The MS / IS encoder 33 encodes the input signal using the MS stereo encoding method or the IS stereo encoding method, and outputs the obtained data to the switch 35. Further, the MS / IS encoding unit 33 performs necessary bit number information B _n (t) ₂ , And distortion rate information E _n (t) ₂ Is generated and supplied to the encoding control unit 34.
[0020]
Based on the information supplied from the dual encoding unit 32 and the MS / IS encoding unit 33, the encoding control unit 34 performs encoding using an encoding method with a low distortion rate or an encoding method with a small number of necessary bits. The contact of the switch 35 is switched so as to select the converted code string. The code string selected by the switch 35 is input to the multiplexer 36.
[0021]
The multiplexer 36 is a code string C for each band divided by the filter bank 31. _n , C _n-1 , ..., C ₁ And the synthesized code string C is output to a device external to the coding device 21 such as a transmission path (not shown).
[0022]
FIG. 4 is a block diagram illustrating a configuration example of a conventional encoding device that encodes an input signal in the frequency domain.
[0023]
The region converting unit 51-1 performs spectrum conversion on the input left signal L (t) into the frequency domain, and generates the generated spectrum signal L _n (f) is output to the dual encoding unit 52 and the MS / IS encoding unit 53. Similarly to the region converting unit 51-1, the region converting unit 51-2 performs spectral conversion on the right signal R (t) and obtains the obtained spectrum signal R. _n (f) is output to the dual encoding unit 52 and the MS / IS encoding unit 53.
[0024]
The dual encoding unit 52 encodes the input signal by the dual encoding method, and outputs the obtained code string to the switch 55. In addition, the dual encoding unit 52 includes necessary number-of-bits information B that is information regarding the data amount of the encoded data. _n (f) ₁ , And distortion rate information E, which is information about distortion rate with sine wave when encoding _n (f) ₁ Is supplied to the encoding control unit 54.
[0025]
The MS / IS encoding unit 53 encodes the input signal by the MS stereo encoding method or the IS stereo encoding method, and outputs the obtained data to the switch 55. Further, the MS / IS encoding unit 53 performs necessary bit number information B _n (f) ₂ , And distortion rate information E _n (f) ₂ Is supplied to the encoding control unit 54.
[0026]
Based on the information supplied from the dual encoding unit 52 and the MS / IS encoding unit 53, the encoding control unit 54 uses an encoding method with a smaller distortion rate or an encoding method with a smaller number of necessary bits. The switch 55 is controlled so as to select the encoded code string.
[0027]
FIG. 5 is a block diagram showing a configuration example of a conventional encoding device configured by combining the encoding device 21 of FIG. 3 and the encoding device 31 of FIG.
[0028]
In other words, in this example, the input signals L (t) and R (t) are divided into a predetermined number of bands by the filter banks 71-2 and 71-2, and the divided signals are converted into area conversion units. The spectrum is converted by 72-1 and 72-2. The converted spectrum signal is encoded by the dual encoding unit 73 and the MS / IS encoding unit 74. In the encoding control unit 75 and the switch 76, the code sequence encoded by the dual encoding unit 73 and the MS / IS encoding unit 74 is more efficient (the distortion rate is smaller or the data amount is larger). A code string based on the encoding method is selected and output to the multiplexer 77. Then, the multiplexer 77 combines the input data of all the bands, and then outputs them to the outside of the encoding device 61.
[0029]
Next, processing of the encoding control unit 34 of the encoding device 21 of FIG. 3 will be described with reference to the flowchart of FIG. Although not described, the processes of the encoding control unit 54 in FIG. 4 and the encoding control unit 75 in FIG. 5 are similar processes. In this example, the encoding control unit 34 selects an encoding method based on the distortion rate.
[0030]
In step S <b> 1, the encoding control unit 34 receives the distortion rate information E notified from the dual encoding unit 32. _n (t) ₁ , And the distortion rate information E received from the MS / IS encoding unit 33 _n (t) ₂ Compare In step S2, the encoding control unit 34 determines whether or not the distortion rate notified from the dual encoding unit 32 is smaller than the distortion rate notified from the MS / IS encoding unit 33. If it is determined, the switch 35 is controlled in step S3 and the data encoded by the dual encoding unit 32 is output to the multiplexer 36.
[0031]
On the other hand, when the encoding control unit 34 determines in step S2 that the distortion rate notified from the dual encoding unit 32 is larger than the distortion rate notified from the MS / IS encoding unit 33, the encoding control unit 34 proceeds to step S4. Then, the switch 35 is controlled, and the data encoded by the MS / IS encoder 33 is output to the multiplexer 36.
[0032]
Similar processing is performed in other bands. As a result, a code string C encoded for each band using a more efficient encoding method is generated and output to the outside of the encoding device 21.
[0033]
As described above, by comparing the encoding efficiency of the respective encoding methods, and selecting the optimum method according to the result, higher efficiency than in the case of encoding with one encoding method. Encoded data can be acquired.
[0034]
FIGS. 7A to 7D show the operation time probability P of MS stereo coding in the coding apparatus of FIGS. _MS , Or IS stereo coding operating time probability P _IS FIG. 5 is a diagram illustrating an example of a relationship between a signal-to-noise ratio SNR (signal power to noise power ratio) of an encoded (quantized) signal and a separation between left and right signals.
[0035]
As shown in FIG. 7A, the probability P shown on the horizontal axis _MS Or P _IS And the SNR shown on the vertical axis is proportional, and the probability P _MS Or P _IS As the value approaches 100% (monaural), the SNR improves.
[0036]
FIG. 7B shows the probability P _MS Or P _IS FIG. 7C is a diagram showing a temporal change in SNR. As shown in these figures, each waveform has the same phase, and the probability P depends on the input signal. _MS Or P _IS Since the coding efficiency is improved by increasing the SNR, the SNR is also improved and the sound quality is improved. Therefore, from the viewpoint of coding efficiency, the probability P _MS Or P _IS Is preferably high.
[0037]
However, the probability P _MS A high value indicates that there is a high correlation between the left and right signals, and the probability P _IS A high value indicates that the intensity signal and spectrum to be encoded are equivalent to one channel even though the power level is different. That is, the probability P _MS Or P _IS A high value indicates that the stereo signal becomes monophonic, and as shown in FIG. _MS / P _IS Deteriorates with increasing.
[0038]
Also, probability P _MS Or P _IS Since SNR is linked, probability P _MS Or P _IS If the value of is high, the SNR may reach a perceptual limit noise level in the psychoacoustic model (a level that can be perceived as noise when the SNR is further reduced) due to the nature of the input signal or time changes. Therefore, overall, the probability P _MS Or P _IS A high value of is not necessarily preferred.
[0039]
[Problems to be solved by the invention]
By the way, in the encoding apparatus as shown in FIG. 3 to FIG. 5, which is the efficiency when encoded by MS stereo encoding or IS stereo encoding and the efficiency when encoded by dual encoding? The determination as to whether it is excellent or not is not known unless the two encoding processes are actually executed, and there is a problem that the processing amount of each encoding unit increases.
[0040]
Also, when MS stereo coding or IS stereo coding is ON, encoding efficiency can be increased (quantization noise is low), but such an effect cannot be obtained when it is OFF. Therefore, when MS stereo coding or IS stereo coding is in the ON state and when it is in the OFF state, the temporal sound quality fluctuation is large, and it may cause the listener to feel a great sense of discomfort. There is also.
[0041]
The present invention has been made in view of such circumstances, and is capable of encoding or decoding an audio signal with higher efficiency while suppressing the listener from feeling uncomfortable.
[0042]
[Means for Solving the Problems]
An encoding apparatus according to the present invention includes an encoding method selection unit that selects an encoding method based on an input signal, and an encoding method that encodes an input signal based on the encoding method selected by the encoding method selection unit. Means, a distortion rate detecting means for detecting a distortion rate of encoding by the encoding means, and a left / right component of the input signal based on a mixing ratio determined corresponding to the distortion rate detected by the distortion rate detecting means. Mixing means for mixing, and the encoding method selection means selects the encoding method based on the input signal mixed by the mixing means.
[0043]
Output correction information generating means for generating output correction information when decoding the input signal encoded by the encoding means can be further provided.
[0044]
The encoding method selection means can select the encoding method of the input signal based on a threshold value determined by the configuration of the encoding device.
[0045]
The encoding method selection means can select an encoding method from any one of a dual encoding method, an MS stereo encoding method, and an IS stereo encoding method.
[0046]
The encoding method selection means is , input signal Obtained by mixing Of left and right components of Choose to encode with dual coding when low Can be.
[0047]
The encoding method selection means is an input signal Obtained by mixing The left-right component correlation can be determined using the ratio of the sum of the sum signals to the sum of the difference signals of the left-right components.
[0048]
The encoding method selection means is an input signal Obtained by mixing When the left and right components of the input signal are highly correlated, the MS stereo encoding method or IS stereo encoding method is selected based on the maximum absolute value of the difference signal of the left and right components of the input signal. be able to.
[0049]
The mixing means stores the mixing ratio, and can change the mixing ratio based on the mixing ratio determined immediately before and the interpolation function of the mixing ratio currently determined.
[0050]
Input signal storage means for storing the input signal is further provided, and the mixing means can mix the left and right components of the same input signal again based on the distortion rate when the input signal is encoded.
[0051]
The encoding method of the encoding device of the present invention includes an encoding method selection step for selecting an encoding method based on an input signal, and an input signal based on the encoding method selected by the processing of the encoding method selection step. An encoding step for encoding the image, a distortion rate detection step for detecting a distortion rate of encoding by the processing of the encoding step, and a mixing ratio determined in accordance with the distortion rate detected by the processing of the distortion rate detection step And a mixing step for mixing the left and right components of the input signal, and the encoding method selection step processing selects the encoding method based on the input signal mixed by the mixing step processing. And
[0052]
The recording medium program according to the present invention encodes an input signal based on an encoding method selection step that selects an encoding method based on an input signal and an encoding method selected by the processing of the encoding method selection step. Encoding step, a distortion rate detecting step for detecting a distortion rate of encoding by the processing of the encoding step, and a mixing ratio determined corresponding to the distortion rate detected by the processing of the distortion rate detecting step And a mixing step for mixing the left and right components of the input signal, and the processing of the encoding method selection step selects the encoding method based on the input signal mixed by the processing of the mixing step.
[0056]
In the encoding apparatus and method and the recording medium program of the present invention, the encoding method is selected based on the input signal, the input signal is encoded based on the selected encoding method, and the detected distortion is detected. Based on the mixing ratio determined corresponding to the rate, the left and right components of the input signal are mixed. An encoding method is selected based on the mixed input signal.
[0058]
DETAILED DESCRIPTION OF THE INVENTION
FIG. 8 is a block diagram showing a configuration example of an encoding apparatus to which the present invention is applied.
[0059]
The filter bank 101-1 converts the left signal L (t) of the input audio signal into a signal L of n frequency bands. _n (t), L _n-1 (t), ..., L ₁ Generated signal L divided into (t) _n (t) is output to the adaptive mixing unit 102. Similarly to the filter bank 101-1, the filter bank 101-2 also converts the right signal R (t) of the input audio signals into a signal R of n frequency bands. _n (t), R _n-1 (t), ..., R ₁ Generated signal R divided into (t) _n (t) is output to the adaptive mixing unit 102. Although not shown, signal L _n-1 (t), ..., L ₁ (t), R _n-1 (t), ..., R ₁ Similarly, a corresponding processing unit is provided for (t).
[0060]
The adaptive mixing unit 102 receives the distortion rate information E notified from the distortion rate detection unit 106. _n Based on (f), the signal L _n (t), R _n (t) is mixed and the signal L _n (t) _mix , R _n (t) _mix (Details thereof will be described later with reference to FIG. 9). Generated signal L _n (t) _mix , And R _n (t) _mix Are supplied to the area conversion units 103-1 and 103-2, respectively. As will be described later, the distortion rate detection unit 106 determines distortion rate information E according to the result of encoding in the encoding unit 105. _n Since (f) is generated, the mixing ratio is set to 0 in the initial state of operation. That is, signal L ₀ (t), R ₀ No mixing process is applied to (t).
[0061]
The adaptive mixing unit 102 also corrects the power correction information P for correcting the output of the left and right signals. _n , _adj (t) is generated and output to the multiplexer 107.
[0062]
The area conversion unit 103-1 receives the supplied signal L _n (t) _mix For example, the domain signal such as MDCT (Modified Discrete Cosine Transform) _n (f) is output to the encoding control unit 104 and the encoding unit 105. Similarly, the region conversion unit 103-2 receives the supplied signal R _n (t) _mix The spectral signal R generated by transforming _n (f) is output to the encoding control unit 104 and the encoding unit 105.
[0063]
The encoding control unit 104 receives the spectrum signal L supplied from the region conversion unit 103. _n (f), R _n Based on (f), the encoding method of the encoding process performed in the encoding part 105 is selected, and the encoding part 105 is controlled.
[0064]
The encoding unit 105 selects dual encoding, MS stereo encoding, or IS stereo encoding based on the control from the encoding control unit 104, and the spectrum signal L supplied from the region conversion unit 103. _n (f), R _n Data sequence C obtained by encoding (f) _n Is output to the multiplexer 107. The above processing is applied to signals L in other frequency bands. _n-1 (t), ..., L ₁ (t), R _n-1 (t), ..., R ₁ The same operation is performed in (t).
[0065]
The multiplexer 107 is a code string C of a predetermined band supplied from the encoding unit 105. _n , Code string C of other bands _n-1 , ..., C ₁ And the synthesized audio data C is output to a device (not shown) provided outside the encoding device 91 or a network. The synthesized audio data C includes power correction information P notified from the adaptive mixing unit 102. _n , _adj Information such as (t) and which encoding method is used for encoding is also included.
[0066]
FIG. 9 is a block diagram illustrating a detailed configuration example of the adaptive mixing unit 102 of FIG.
[0067]
The power calculation unit 121 outputs the signal L divided into a predetermined band by the filter banks 101-1 and 101-2. _n (t), R _n From (t), the power value P of each signal _ln , P _rn Is output to the power correction unit 123.
[0068]
The mixing coefficient setting unit 122 receives distortion rate information E notified from the distortion rate detection unit 106. _n Based on (f), the mixing coefficient is extracted from the correspondence table stored in the built-in storage unit, the mixing coefficient a of the multipliers 124-1 and 124-2, and the multipliers 125-1 and 125-2 are calculated. Set the mixing coefficient b. In addition, the mixing coefficient setting unit 122 notifies the power correction unit 123 of the extracted mixing coefficients a and b.
[0069]
Each of the multipliers 124-1 and 124-2 uses the input signal L to the mixing coefficient a set by the mixing coefficient setting unit 122. _n (t), R _n Multiply (t) and output the obtained signal to adders 126-1 and 126-2. Each of the multipliers 125-1 and 125-2 receives the mixing coefficient b set by the mixing coefficient setting unit 122 as an input signal R. _n (t), L _n Multiply (t) and output the obtained signal to adders 126-1 and 126-2.
[0070]
The adder 126-1 is a left signal L multiplied by a coefficient a in the multiplier 124-1. _n (t) and the right signal R multiplied by the coefficient b by the multiplier 125-1. _n Add (t) and add the result to signal L _n (t) _mix Is output to the area conversion unit 103-1. Further, the adder 126-2 outputs the right signal R multiplied by the coefficient a by the multiplier 124-2. _n (t) and the left signal L multiplied by the coefficient b by the multiplier 125-2 _n Add (t) and signal R _n (t) _mix Is output to the area conversion unit 103-2.
[0071]
FIG. 10 shows distortion rate information E stored in a storage unit (not shown) of the mixing coefficient setting unit 122. _n It is a figure which shows the example of the corresponding | compatible table of (f) and mixing coefficient a, b.
[0072]
In this example, distortion rate information E _n (f) is shown as a percentage, and this value is hereinafter referred to as E. For example, E = 0% means that the perceptual noise is zero. Further, E = 100% means that the noise can be perceived in all spectral regions.
[0073]
In this example, mixing coefficients a = 1.00 and b = 0.00 are set corresponding to the distortion rate E = 0%. In this case, the left and right input signals L _n (t), R _n Since (t) is not mixed, encoding is performed in a completely separated state (complete stereo). Corresponding to the distortion rate E = 100%, mixing coefficients a = 0.50 and b = 0.50 are set. In this case, the left and right input signals L _n (t), R _n (t) is mixed at the same ratio and encoded in a completely unified state (complete monaural).
[0074]
The power correction unit 123 receives the signal L supplied from the power calculation unit 121. _n (t), R _n Power value P of (t) _ln , P _rn Based on the mixing coefficients a and b notified from the mixing coefficient setting unit 122, power correction information P used when power correction is performed in a decoding device 151 (FIG. 19) described later. _n , _adj (t) is generated and output to the multiplexer 107. That is, the power correction unit 123 stores power correction information P in a storage unit (not shown). _n , _adj (t), mixing coefficients a and b, and power value P _ln , P _rn A correspondence table describing the relationship is stored.
[0075]
FIG. 11 is a diagram illustrating an example of a correspondence table stored in the power correction unit 123.
[0076]
In this example, the power value P calculated by the power calculation unit 121. _ln , P _rn , Distortion rate information E _n (f), mixing coefficients a and b, and a signal L reproduced in the decoding device 151 _n '(t) _mix , R _n '(t) _mix Power value P _lnmix , P _rnmix And power correction information P _n , _adj (t) is associated. In this example, the power correction information P _n , _adj (t) is represented by power weighting coefficients c and d set in the decoding device 151.
[0077]
For example, as shown in the second row of FIG. _n The power value of (t) is P _ln = 1.0, signal R _n The power value of (t) is P _rn = 1.0 and distortion rate E = 0%, the mixing coefficients are set as a = 1.00 and b = 0.00 from the correspondence table shown in FIG. Signal L ′ in decoding device 151 _n (t) _mix Power value of P _lnmix = 1.0 and signal R ' _n (t) _mix Power value of P _rnmix = 1.0. Power correction information P _n , _adj Since (t) includes a coefficient that makes the reproduction signal close to the input signal, the signal L ′ _n (t) _mix The coefficient for correcting the power of the signal is c = 1.00, and the signal R ′ _n (t) _mix The coefficient for correcting the power of is d = 1.00.
[0078]
FIG. 12 is a block diagram illustrating a detailed configuration example of the multiplier 124-1 (not illustrated, but the multiplier 124-2 is configured similarly).
[0079]
In this example, buffers 124A and 124B are provided, and the currently set mixing coefficient a (t ₀ ) Is set in the buffer 124A immediately before (the time t = 1), the mixing coefficient a (t ₁ ) Are stored in the buffer 124B.
[0080]
When the mixing coefficient is changed, a discontinuous point may occur in the signal output at that time. Therefore, as shown by the curves (1) to (3) in FIG. 13, it is possible to suppress the occurrence of discontinuous points by changing the mixing coefficient linearly or curvedly. In this example, two buffers are provided, but three or more buffers may be provided. The interpolation function for interpolating each mixing coefficient may be a linear function, a quadratic function or a cubic function. Of course, the multipliers 125-1 and 125-2 are similarly provided with a buffer, the mixing coefficient b is stored, and the mixing coefficient can be changed based on the interpolation function.
[0081]
FIG. 14 is a block diagram illustrating a detailed configuration example of the encoding control unit 104 in FIG.
[0082]
The normalization unit 141-1 receives the spectrum signal L input from the region conversion unit 103-1. _n (f) is normalized for each divided frequency band or for each range of small regions obtained by collecting several spectra in the same divided frequency band. _n (f) is generated and output to the adder 142-1 and the subtractor 142-2. Similarly, the normalization unit 141-2 receives the spectrum signal R input from the region conversion unit 103-2. _n Normalize (f) and normalize spectral signal r _n (f) is generated and output to the adder 142-1 and the subtractor 142-2. Normalized spectral signal l _n (f), r _n (f) is a signal s generated by addition or subtraction on the spectrum in the adder 142-1 and subtractor 142-2, respectively. _n (f) (= | l _n (f) + r _n (f) |), d _n (f) (= | l _n (f) -r _n (f) |) is supplied to the comparison unit 143.
[0083]
The comparison unit 143 receives the input signal s _n (f), d _n The total value S, D for each divided frequency band of (f) is calculated, and the spectrum signal L performed in the encoding unit 105 based on the ratio S / D _n (f), R _n Select the encoding method of (f). Note that the comparison unit 143 determines whether or not to perform encoding by dual encoding, and uses either MS stereo encoding or IS stereo encoding to transmit the spectrum signal L. _n (f), R _n Whether to encode (f) is determined by the comparison unit 144 described later.
[0084]
The comparison unit 144 receives the normalized spectrum signal l notified from the comparison unit 143. _n (f), r _n Difference component d of (f) _n (f) (= l _n (f) -r _n Based on (f)), the spectrum signal L can be obtained by either MS stereo encoding or IS stereo encoding. _n (f), R _n Select whether to encode (f).
[0085]
Next, the operation of the encoding device 91 of FIG. 8 will be described with reference to the flowchart of FIG.
[0086]
In step S <b> 11, the filter bank 101 divides the input audio signal into predetermined frequency bands, and outputs the generated signal to the adaptive mixing unit 102. That is, the filter bank 101-1 divides the left signal L (t) into n bands and outputs the left signal L _n (t) is output to the adaptive mixing unit 102. Further, the filter bank 101-2 divides the right signal R (t) into n bands, and the left signal R _n (t) is output to the adaptive mixing unit 102.
[0087]
In step S12, the adaptive mixing unit 102 receives the distortion rate information E notified from the distortion rate detection unit 106. _n Based on (f), the input signal L _n (t), R _n A mixing process is performed on (t). Details of the mixing process will be described later with reference to the flowchart of FIG.
[0088]
Signal L generated by the mixing process _n (t) _mix , And R _n (t) _mix Is supplied to the region conversion unit 103, and is converted from a time domain to a frequency domain spectrum signal by MDCT or the like in step S13, and the converted spectrum signal L _n (f), R _n (f) is output to the encoding control unit 104 and the encoding unit 105.
[0089]
In step S14, the encoding control unit 104 transmits the spectrum signal L input to the encoding unit 105. _n (f), R _n A process for controlling the encoding method of (f) is performed. Details of the encoding control process will be described later with reference to the flowchart of FIG.
[0090]
In step S15, the encoding unit 105 selects dual encoding, MS stereo encoding, or IS stereo encoding based on an instruction from the encoding control unit 104, and the spectrum signal supplied from the region conversion unit 103 is selected. L _n (f), R _n (f) is encoded by the selected method, and the obtained code sequence C _n Is output to the multiplexer 107. It should be noted that, depending on which encoding method is used, for example, when encoding a spectrum signal, a combination of information for identifying a codebook to be referred to, information on quantization accuracy, or normalization information Thus, the decoding device 151 is uniquely determined.
[0091]
The distortion rate detection unit 106 detects the distortion rate of the encoding process executed by the encoding unit 105, and detects distortion rate information E. _n Generate (f). Generated distortion rate information E _n In step S16, (f) is notified to the adaptive mixing unit 102 and used for the subsequent processing. The above processing is performed in all bands.
[0092]
In step S17, the multiplexer 107 transmits the code string C supplied from the encoding unit 105. _n Is a code string C from an encoding unit of another band. _n-1 , C _n-2 , ..., C ₁ And the obtained composite code string C is output to a device (not shown) or a network provided outside the encoding device 91. The composite code string C includes power correction information P notified from the adaptive mixing unit 102. _n , _adj Information such as (t) is included.
[0093]
Next, the mixing process of the adaptive mixing unit 102 executed in step S12 of FIG. 15 will be described with reference to the flowchart of FIG.
[0094]
In step S31, the mixing coefficient setting unit 122 receives distortion rate information E from the distortion rate detection unit 106. _n It is determined whether or not (f) has been notified. When it is determined that the notification has been received, the process proceeds to step S32, and distortion rate information E _n Based on (f), the mixing coefficients a and b of the multipliers 124 and 125 are set. For example, when notified that the distortion rate E = 10%, the mixing coefficient setting unit 122 extracts the mixing coefficient a = 0.95, b = 0.05 from the correspondence table as shown in FIG. Is set to “0.95”, and the mixing coefficient b of the multiplier 125 is set to “0.05”. The mixing coefficient setting unit 122 notifies the power correction unit 123 of the set mixing coefficient.
[0095]
On the other hand, in step S31, the distortion rate information E is sent from the distortion rate detection unit 106. _n When it is determined that (f) has not been notified, in step S33, the mixing coefficient setting unit 122 sets the initial mixing coefficient in the multipliers 124 and 125, respectively. That is, as described above, in the initial state, the distortion rate E = 0%, and the mixing coefficients a = 1.00 and b = 0.00 are set.
[0096]
In step S34, the adder 126-1 uses the left signal L in the multiplier 124-1. _n The signal obtained by multiplying (t) by the mixing coefficient a and the right signal R by the multiplier 125-1. _n The signal obtained by multiplying (t) by the mixing coefficient b is added, and the mixing signal L _n (t) _mix Is generated and output to the area conversion unit 103-1.
[0097]
In step S35, the adder 126-2 uses the right signal R in the multiplier 124-2. _n The signal obtained by multiplying (t) by the mixing coefficient a and the left signal L in the multiplier 125-2 _n The signal obtained by multiplying (t) by the mixing coefficient b is added, and the mixing signal R _n (t) _mix Is output to the area conversion unit 103-2.
[0098]
That is, when the above-described mixing coefficients (a = 0.95, b = 0.05) are set in the multipliers 124 and 125 by steps S34 and S35, the left and right signals L _n (t), R _n One of (t) is output to the region conversion unit 103 after 5% of the other is mixed. In the initial state, left and right signal L _n (t), R _n (t) is output to the region conversion unit 103 in a completely stereo state without being mixed.
[0099]
In step S36, the power calculation unit 121 determines that the signal L divided into a predetermined band by the filter bank 101. _n (t), R _n Power value P of (t) _ln , P _rn Is calculated and notified to the power correction unit 123.
[0100]
In step S37, the power correction unit 123 receives the signal L supplied from the power calculation unit 121. _n (t), R _n Power value P of (t) _ln , P _rn Based on the mixing coefficients a and b notified from the mixing coefficient setting unit 122, power correction information P used when power correction is performed in a decoding device 151 (see FIG. 19) described later. _n , _adj (t) is generated and output to the multiplexer 107.
[0101]
For example, from the power calculation unit 121, the signal L _n Power value P of (t) _ln = 5.0, R _n Power value P of (t) _rn When 1.0 is notified and the mixing coefficient setting unit 122 notifies that the mixing coefficients a = 0.75 and b = 0.25 (when the distortion rate E = 50%), the fourth row from the top in FIG. As shown, power correction information P _n , _adj As (t) (power weighting coefficient), c = 1.25 and d = 0.50 are extracted. That is, in the decoding device 151, the signal L _n Signal L ′ obtained when data of (t) is decoded _n (t) _mix Is the power value P _lnmix = 4.0, signal R _n Signal R ′ obtained when data of (t) is decoded _n (t) _mix Is the power value P _rnmix Since the reproduction is performed at 2.0, power weighting coefficients c and d that are equal to the input signal when the reproduction signal is multiplied are extracted and output to the multiplexer 107.
[0102]
For example, when the distortion rate is high, adaptive mixing section 102 sets the mixing coefficient so as to change the left and right signals in a monaural manner, and increases the operation probability of MS stereo encoding or IS stereo encoding. As a result, the SNR can be increased and the distortion rate can be decreased. Further, as described above, the normalized spectrum l is set by setting the mixing coefficient based on the distortion rate information that has been fed back. _n (f), r _n In the region (f), a region having a high correlation is generated in a region having no high correlation. Further, in the decoding device, the power correction information P _n , _adj Since power correction is performed based on (t), the separation of the left and right signals is saved.
[0103]
Next, the encoding control process of the encoding control unit 104 executed in step S14 of FIG. 15 will be described with reference to the flowchart of FIG.
[0104]
In step S51, the normalization unit 141 normalizes the input signal for each divided frequency band or for each range of small regions obtained by collecting several spectra in the same divided frequency band. Generated normalized spectral signal l _n (f), r _n (f) is supplied to the adder 142-1 and the subtractor 142-2, and in step S52, the sum signal s of the normalized spectrum signal _n (f) (= | l _n (f) + r _n (f) |) is added to the difference signal d by the adder 142-1. _n (f) (= | l _n (f) -r _n (f) |) is respectively generated by the subtractor 142-2. Sum signal s of the generated normalized spectral signal _n (f) and difference signal d _n (f) is supplied to the comparison unit 143.
[0105]
In step S53, the comparison unit 143 receives the input signal s. _n The total value S of all bands in (f) is expressed by the following equation (1) as signal d _n The total value D of the normalized range of (f) is calculated by the following equation (2).
[Expression 1]

[Expression 2]

In these equations, f ₀ Is the first spectrum number in the normalized range, f ₁ Indicates the number of the terminal spectrum.
[0106]
Normalized spectral signal l _n (f) and normalized spectral signal r _n As (f) is more similar (the correlation is higher), the total value S is larger and the total value D is smaller. Conversely, the normalized spectral signal l _n (f) and normalized spectral signal r _n When (f) is different (correlation is low), the sum value S and the sum value D are the same value, so by calculating the ratio of the sum values S and D (sum value ratio S / D) , Normalized spectral signal l _n (f) and normalized spectral signal r _n The correlation of (f) can be acquired. For example, if the sum value ratio S / D is greater than `` 1 '', the normalized spectral signal l _n (f) and normalized spectral signal r _n (f) shows a high correlation.
[0107]
Then, in step S54, the comparison unit 143 sets an allowable error level (threshold value) in which the sum value ratio S / D calculated in step S53 is set in advance for each of the divided frequency bands or normalized small regions. Judge whether it is smaller than Thr. When the comparison unit 143 determines that the total value ratio S / D is smaller than the allowable error level Thr, the comparison unit 143 proceeds to step S55, and the spectral signal L input to the encoding unit 105 is obtained. _n (f), R _n (f) is selected to be encoded by dual encoding, and is notified to the encoding unit 105. That is, the allowable error level is a normalized spectrum signal l if the sum value ratio S / D is equal to or higher than a predetermined level. _n (f) and normalized spectral signal r _n If (f) has a correlation above a predetermined level), it is set to force encoding by MS or IS stereo encoding. In this embodiment, the normalized spectrum signal l is calculated using the ratio of the summation values S and D. _n (f) and normalized spectral signal r _n Although the correlation of (f) is determined, the correlation determination method is naturally not limited to this. _n (f) and r _n The determination may be made using other parameters, such as obtaining a correlation coefficient by comparing the absolute values of (f).
[0108]
On the other hand, in step S54, when the comparison unit 143 determines that the sum value ratio S / D is larger than the allowable error level Thr, the comparison unit 143 notifies the comparison unit 144 accordingly. Then, in step S56, the comparison unit 144 sets d for the spectrum of the target band. _n It is determined whether the maximum value of (f) is greater than the quantization accuracy level Thq that can be realized by the decoding device 151. That is, the comparison unit 144 outputs the difference signal d _n When it is necessary to encode (f), MS stereo encoding is selected, and when it is not necessary to encode, IS stereo encoding is selected.
[0109]
In step S56, the comparison unit 144 determines that d _n When it is determined that the maximum value of (f) is larger than the quantization accuracy level Thq, the process proceeds to step S57, where the spectrum signal L input to the encoding unit 105 is processed. _n (f), R _n (f) is selected to be encoded by MS stereo encoding, and is notified to the encoding unit 105. In addition, the comparison unit 144 determines that d is d at step S56. _n When it is determined that the maximum value of (f) is smaller than the quantization accuracy level Thq, the process proceeds to step S58, where the spectrum signal L input to the encoding unit 105 is processed. _n (f), R _n (f) is selected to be encoded by IS stereo encoding, and is notified to the encoding unit 105.
[0110]
This makes the normalized spectral signal l _n (f) and normalized spectral signal r _n (f) has a high correlation, and even if there is a possibility that higher SNR can be achieved with dual coding than with MS or IS stereo coding, the summation ratio S / D is noise in terms of hearing. If it is higher than a threshold that cannot be heard, the input signal is encoded by MS or IS stereo encoding.
[0111]
The difference signal d _n Even if (f) is not encoded, the information regarding the normalization of the left and right signals is encoded, so that IS stereo encoding can be considered equivalent to MS stereo encoding. As a result, a processing unit that performs MS stereo encoding and a processing unit that performs IS stereo encoding are not provided separately, and the encoding device 91 can be configured to be smaller.
[0112]
The allowable error level Thr is set according to the configuration of the coding system, such as the block length of area conversion and bit allocation. The quantization accuracy level Thq may be set to the highest quantization accuracy level that can be realized by the encoding device 91, or the quantization accuracy level Thq (f) may be set for each frequency band. . That is, the quantization accuracy level Thq is also set according to the system, like the allowable error level Thr.
[0113]
18 shows the relationship between the separation and the signal-to-noise ratio SNR in the encoding device 91 (FIG. 18A), and the temporal change in the signal-to-noise ratio SNR of the encoded (quantized) signal (see FIG. B)), MS stereo coding operating time probability P _MS , Or IS stereo coding operating time probability P _IS It is a figure which shows the time change (FIG.18 (C)) of this, and the time change (FIG.18 (D)) of the separation of a left-right signal.
[0114]
As shown in FIG. 18B and FIG. 18C, the signal-to-noise ratio SNR and the operating time probability P of MS stereo coding _MS , Or IS stereo coding operating time probability P _IS Are linked to each other. Therefore, as described above, P _MS , P _IS Can be controlled to improve the SNR. Thereby, sound quality can be improved.
[0115]
Further, as shown in FIG. 18A, as the SNR improves, the separation of the left and right signals deteriorates (monaural), so that the separation is illustrated in accordance with the fluctuation of the SNR shown in FIG. 18 (D), but as described above, the power correction information P _n , _adj Since (t) is generated and power adjustment is performed at the time of decoding, the separation of the left and right signals can be improved. In FIGS. 18B, 18C, and 18D, the line L ₁ , L _Three , L _Five Shows the characteristics of the encoding device 91 of FIG. ₂ , L _Four , L ₆ These show the characteristics of a conventional encoding device.
[0116]
In the above-described example, the encoding distortion rate is detected, the mixing coefficient is set according to the value, and the input signal at the next timing is mixed. However, it is predetermined until the distortion rate becomes a predetermined threshold value or less. The input signal in the band may be configured to be repeatedly mixed. In this case, the signal L generated by the filter bank 101-1 _n (t) and the signal R generated by the filter bank 101-2 _n (t) is stored in a memory or the like (not shown), and distortion rate information E fed back to the adaptive mixing unit 102 _n By (f), mixing, region conversion, and encoding are performed again.
[0117]
FIG. 19 is a block diagram illustrating a configuration example of a decoding device to which the present invention is applied.
[0118]
The demultiplexer 161 converts the code string C supplied via a transmission path (not shown) into a code string C for each predetermined band. _n , C _n-1 , ..., C ₁ Each code string C _i Are output to a corresponding decoding unit (for the sake of convenience, only the decoding unit 162 is shown). Code string C _n Is supplied to the decoding unit 162.
[0119]
The decoding unit 162 receives the input code string C. _n Is decoded by a decoding method corresponding to the encoding method, and the obtained spectrum signal L ′ _n (f) is transmitted to the region conversion unit 163-1 and the spectrum signal R ′. _n (f) is output to the area conversion unit 163-2. In addition, the decoding unit 162 receives the code string C. _n Power correction information P obtained from _n , _adj (t) is supplied to the power weighting unit 164.
[0120]
The region converting unit 163 receives the input spectrum signal L ′. _n (f), R ' _n The signal L ′ obtained by converting to (f) using the inverse MDCT or the like to a time domain signal _n (t) _mix , R ' _n (t) _mix Is output to the power weighting unit 164.
[0121]
The power weighting unit 164 receives the notified power correction information P _n , _adj Based on the power weighting coefficient included in (t), the signal L ′ supplied from the region conversion unit 163 _n (t) _mix , R ' _n (t) _mix The generated signal L ' _n (t) to the filter bank 165-1 and the signal R ′ _n (t) is output to the filter bank 165-2.
[0122]
The filter bank 165 receives the signal L ′ supplied from the power weighting unit 164. _n (t), R ' _n (t), signal L 'of other band _n-1 (t), ..., L ' ₁ (t), R ' _n-1 (t), ..., R ' ₁ The audio signals L ′ (t) and R ′ (t) generated by combining with (t) are output to the outside of the decoding device 151.
[0123]
FIG. 20 is a block diagram illustrating a detailed configuration example of the power weighting unit 164.
[0124]
The power weighting coefficient setting unit 171 receives the supplied power correction information P _n , _adj The power weighting coefficient c included in (t) is set in the multiplier 172-1, and the power weighting coefficient d is set in the multiplier 172-2.
[0125]
The multiplier 172-1 receives the input signal L ′. _n (t) _mix Is multiplied by the power weighting coefficient c, and the multiplier 172-2 receives the input signal R ′. _n (t) _mix Is multiplied by the power weighting coefficient d and the resulting signal L ′ _n (t), R ' _n (t) is output to the filter banks 165-1 and 165-2.
[0126]
FIG. 21 is a block diagram illustrating a detailed configuration example of the multiplier 172-1 (not shown, but the multiplier 172-2 is configured similarly).
[0127]
In this example, buffers 172A and 172B are provided, and the currently set power weighting coefficient c (t ₀ ) Is set in the buffer 124A immediately before (it was set when the time t = 1). The power weighting coefficient c (t ₁ ) Are stored in the buffer 172B.
[0128]
That is, when the power weighting coefficient c (t) is changed, a discontinuous point may occur in the signal output at that time. Therefore, as shown in lines (1) to (3) in FIG. 22, it is possible to suppress the occurrence of discontinuous points by changing the power weighting coefficient c (t) linearly or curvedly. it can. In this example, two buffers are provided, but three or more buffers may be provided. Further, the interpolation function for interpolating each power weighting coefficient may be linear, or may be interpolated by a quadratic or cubic function.
[0129]
Next, the decoding process of the decoding device 151 in FIG. 19 will be described with reference to the flowchart in FIG.
[0130]
In step S71, the demultiplexer 161 converts the input code string C into a code string C having a predetermined number of bands n. _n , C _n-1 , ..., C ₁ And output to the corresponding decoding unit.
[0131]
In step S72, the decoding unit 162 selects a decoding method from a combination of normalization information, quantization accuracy information, codebook number, and the like, and receives the input code string C. _n And the resulting spectral signal L ′ _n (f) is transmitted to the region conversion unit 163-1 and the spectrum signal R ′. _n (f) is output to the area conversion unit 163-2. In addition, the decoding unit 162 receives the code string C. _n Power correction information P obtained from _n , _adj (t) is output to the power weighting unit 164.
[0132]
In step S73, the region conversion units 163-1 and 163-2 receive the input spectrum signal L ′. _n (f), R ' _n The signal L ′ obtained by converting to (f) using the inverse MDCT or the like to a time domain signal _n (t) _mix , R ' _n (t) _mix Is output to the power weighting unit 164. Signal L _n '(t) _mix , R _n '(t) _mix Is a signal that may have been mixed in the encoding device 91. Depending on the setting of the mixing coefficient, a signal that was originally a stereo signal may be almost a monaural signal. Therefore, in step S74, the power weighting unit 164 receives the notified power correction information P. _n , _adj Based on (t), a power weighting process is executed to reproduce a pseudo stereo signal. Details of the power weighting process will be described later with reference to the flowchart of FIG.
[0133]
Signal L 'obtained by the power weighting process _n (t), R ' _n (t) is output to the filter banks 165-1 and 165-2, respectively. The above processing is performed for each band.
[0134]
In step S75, the filter bank 165 receives the signal L ′ supplied from the power weighting unit 164. _n (t), R ' _n (t), signal L 'of other band _n-1 (t), ..., L ' ₁ (t), R ' _n-1 (t), ..., R ' ₁ (t) and the synthesized audio signals L ′ (t) and R ′ (t) of the entire band are output to the outside of the decoding device 151.
[0135]
Next, the power weighting process executed in step S74 in FIG. 23 will be described with reference to the flowchart in FIG.
[0136]
In step S91, the power weighting coefficient setting unit 171 determines the power correction information P notified from the decoding unit 162. _n , _adj Based on the power weighting coefficient included in (t), the power weighting coefficients c and d of the multipliers 172-1 and 172-2 are set.
[0137]
In step S92, the multipliers 172-1 and 172-2 respectively use the power weighting coefficients c and d as input signals L ′. _n (t) _mix , R ' _n (t) _mix The signal L 'generated by multiplying by _n (t), R ' _n (t) is output to the filter banks 165-1 and 165-2, respectively.
[0138]
For example, as described above, in the power correction unit 123, the power correction information P _n , _adj (t) When the power weighting coefficients are set to c = 1.25 and d = 0.05, and the power weighting coefficients c and d are set by the power weighting coefficient setting unit 171, the multiplier 172-1 is set. Is the input signal L ' _n (t) _mix Is multiplied by 1.25 and the generated signal L ' _n (t) is output to the filter bank 165-1. Further, the multiplier 172-2 receives the input signal R ′. _n (t) _mix Is multiplied by 0.05 and the generated signal R ' _n (t) is output to the filter bank 165-2.
[0139]
Thereby, when encoding, a pseudo stereo signal can be reproduced even when the separation of the left and right signals is deteriorated.
[0140]
The series of processes described above can be executed by hardware, but can also be executed by software. In this case, for example, the encoding device 91 is configured by a personal computer 181 as shown in FIG.
[0141]
25, a CPU (Central Processing Unit) 191 performs various processes according to a program stored in a ROM (Read Only Memory) 192 or a program loaded from a storage unit 198 into a RAM (Random Access Memory) 193. Execute. Further, the RAM 193 appropriately stores data necessary for the CPU 191 to execute various processes.
[0142]
The CPU 191, ROM 192, and RAM 193 are connected to each other via a bus 194. An input / output interface 195 is also connected to the bus 194.
[0143]
The input / output interface 195 includes an input unit 196 including a keyboard and a mouse, a display including a CRT and an LCD, an output unit 197 including a speaker, a storage unit 198 including a hard disk, a modem, a terminal adapter, and the like. A configured communication unit 199 is connected. The communication unit 199 performs communication processing via a network.
[0144]
A drive 200 is connected to the input / output interface 195 as necessary, and a magnetic disk 201, an optical disk 202, a magneto-optical disk 203, a semiconductor memory 204, or the like is appropriately mounted, and a computer program read therefrom is read. If necessary, it is installed in the storage unit 198.
[0145]
When a series of processing is executed by software, a program constituting the software executes various functions by installing a computer incorporated in dedicated hardware or various programs. For example, it is installed in a general-purpose personal computer 181 from a network or a recording medium.
[0146]
As shown in FIG. 25, this recording medium includes a magnetic disk 201 (including a floppy disk) on which a program is recorded and an optical disk 202 (CD -ROM, DVD (including Digital Versatile Disk)), magneto-optical disk 203 (including MD (Mini-Disk)), or package media consisting of semiconductor memory 204, etc., as well as pre-installed in the device body In this state, the program is configured by a ROM 192 in which a program is recorded and a hard disk included in the storage unit 198.
[0147]
In the present specification, the step of describing the program recorded in the recording medium is not limited to the processing performed in time series according to the described order, but is not necessarily performed in time series, either in parallel or individually. The process to be executed is also included.
[0148]
【The invention's effect】
According to the encoding apparatus and method and the recording medium program of the present invention, the encoding method is selected based on the input signal, the input signal is encoded based on the selected encoding method, and the detected distortion rate is obtained. Based on the correspondingly determined mixing ratio, the left and right components of the input signal are mixed. Since the encoding method is selected based on the mixed input signal, the audio signal can be encoded with higher efficiency.
[Brief description of the drawings]
FIG. 1 is a block diagram illustrating a configuration example of a conventional audio signal transmission system using MS stereo encoding.
FIG. 2 is a block diagram illustrating a configuration example of a conventional audio signal transmission system using IS stereo encoding.
FIG. 3 is a block diagram illustrating a configuration example of a conventional encoding device.
FIG. 4 is a block diagram illustrating a configuration example of another conventional encoding apparatus.
FIG. 5 is a block diagram illustrating a configuration example of still another conventional encoding apparatus.
FIG. 6 is a flowchart illustrating processing of a conventional encoding device.
FIG. 7 is a diagram illustrating a relationship between an operation of a conventional encoding device and a signal to be generated.
FIG. 8 is a block diagram illustrating a configuration example of an encoding device to which the present invention has been applied.
9 is a block diagram illustrating a configuration example of an adaptive mixing unit in FIG. 8. FIG.
10 is a diagram illustrating an example of information stored in a mixing coefficient setting unit in FIG. 9;
11 is a diagram illustrating an example of information stored in a power correction unit in FIG. 9;
12 is a diagram illustrating a configuration example of a multiplier in FIG. 9;
FIG. 13 is a diagram illustrating an example of an interpolation function of a mixing coefficient.
14 is a block diagram illustrating a configuration example of an encoding control unit in FIG. 8. FIG.
FIG. 15 is a flowchart for explaining processing of the encoding device in FIG. 8;
FIG. 16 is a flowchart illustrating details of processing executed in step S12 of FIG.
FIG. 17 is a flowchart for explaining details of processing executed in step S14 of FIG. 15;
18 is a diagram illustrating a relationship between an operation of the encoding device in FIG. 8 and a signal to be generated.
FIG. 19 is a block diagram illustrating a configuration example of a decoding device to which the present invention has been applied.
20 is a block diagram illustrating a configuration example of a power weighting unit in FIG. 19;
FIG. 21 is a block diagram illustrating a configuration example of a multiplier in FIG. 20;
FIG. 22 is a diagram illustrating an example of an interpolation function of a power weighting coefficient.
FIG. 23 is a flowchart for describing processing of the decoding device in FIG. 19;
FIG. 24 is a flowchart illustrating details of processing executed in step S74 of FIG.
FIG. 25 is a block diagram illustrating a configuration example of a personal computer.
[Explanation of symbols]
91 Coding Device, 101-1 and 101-2 Filter Bank, 102 Adaptive Mixing Unit, 103-1 and 103-2 Region Conversion Unit, 104 Coding Control Unit, 105 Coding Unit, 106 Distortion Rate Detection Unit, 107 Multiplexer , 121 power calculation unit, 122 mixing coefficient setting unit, 123 power correction unit, 124-1 and 124-2 multiplier, 124A and 124B buffer, 125-1 and 125-2 multiplier, 126-1 and 126-2 addition 141-1 and 141-2 normalization unit, 142-1 adder, 142-2 subtractor, 143 comparison unit, 144 comparison unit, 151 decoding device, 161 demultiplexer, 162 decoding unit, 163-1 and 163 -2 area conversion unit, 164 power weighting unit, 165-1 and 165-2 filterbank, 171 power weighting coefficient setting unit, 172-1 and 172-2 multipliers, 172A and 172B buffer

Claims (11)

An encoding method selection means for selecting an encoding method based on an input signal;
Encoding means for encoding the input signal based on the encoding method selected by the encoding method selection means;
Distortion rate detecting means for detecting a distortion rate of encoding by the encoding means;
Mixing means for mixing left and right components of the input signal based on a mixing ratio determined corresponding to the distortion rate detected by the distortion rate detection means,
The encoding apparatus, wherein the encoding method selection unit selects the encoding method based on the input signal mixed by the mixing unit. The encoding apparatus according to claim 1, further comprising output correction information generation means for generating output correction information when the input signal encoded by the encoding means is decoded. The encoding apparatus according to claim 1, wherein the encoding method selection unit selects the encoding method of the input signal based on a threshold value determined by a configuration of the encoding apparatus. The encoding apparatus according to claim 1, wherein the encoding method selection unit selects the encoding method from any one of a dual encoding method, an MS stereo encoding method, and an IS stereo encoding method. 5. The encoding method selecting unit, when the left and right components of a signal obtained by mixing the input signal have a low correlation, selects encoding by the dual encoding method. Encoding device. The encoding method selecting means determines a correlation between left and right components of a signal obtained by mixing the input signal, using a ratio of a sum of sum signals to a sum of difference signals of the left and right components. The encoding device according to claim 5. The encoding method selection means determines whether to select the MS stereo encoding method or the IS stereo encoding method when the correlation between the left and right components of the signal obtained by mixing the input signal is high. The encoding device according to claim 5, wherein the encoding device is determined based on a maximum absolute value of a difference signal between the left and right components. The mixing means stores the mixing ratio,
The encoding apparatus according to claim 1, wherein the mixing ratio is changed based on an interpolation function of the mixing ratio determined immediately before and the mixing ratio currently determined. Further comprising input signal storage means for storing the input signal;
The mixing means mixes the left and right components of the input signal stored in the input signal storage means at least once based on a mixing ratio determined corresponding to a distortion rate when the input signal is encoded. The encoding device according to claim 1, wherein: An encoding method selection step of selecting an encoding method based on an input signal;
An encoding step for encoding the input signal based on the encoding method selected by the processing of the encoding method selection step;
A distortion rate detecting step of detecting a distortion rate of encoding by the processing of the encoding step;
A mixing step of mixing left and right components of the input signal based on a mixing ratio determined corresponding to the distortion rate detected by the processing of the distortion rate detection step,
The encoding method selecting step selects the encoding method based on the input signal mixed by the mixing step processing. An encoding method selection step of selecting an encoding method based on an input signal;
An encoding step for encoding the input signal based on the encoding method selected by the processing of the encoding method selection step;
A distortion rate detecting step of detecting a distortion rate of encoding by the processing of the encoding step;
A mixing step of mixing left and right components of the input signal based on a mixing ratio determined corresponding to the distortion rate detected by the processing of the distortion rate detection step,
In the encoding method selection step, the encoding method is selected based on the input signal mixed in the mixing step. The computer-readable program is recorded. Medium .

Images