JP4297635B2 - Information reproducing method, information reproducing apparatus, and information recording / reproducing apparatus - Google Patents

Description

【０１０２】
もっとも、表１に示す判定結果を更に調査したところ、先行テーブルを参照して判定した結果と、続行テーブルを参照して判定した結果とが一致しない場合があることが判った。その現象は、表１中のＮｏ．２５の例に示されており、先行テーブルを参照した判定結果は「３」レベルで不正解、続行テーブルを参照した判定結果は「２」レベルで正解であることが判る。このような現象を数多く調べたところ、次の傾向があることが判った。
【０１０３】
３つの連続するセル１の組合せを（Ｘ，Ｙ，Ｚ）とし、Ｙを注目セルの多値レベルデータとする時、
▲１▼先行テーブル参照でＹのレベル判定正解確率が高いのは、Ｘ＞Ｚの関係の場合
▲２▼続行テーブル参照でＹのレベル判定正解確率が高いのは、Ｘ＜Ｚの関係の場合
である。但し、Ｘ，Ｙ，Ｚで表記された多値レベル値は、セル１に記録されたマークの面積が大きいほど、多値レベルの数値が大きいものとする。表１のＮｏ．２５では、（Ｘ，Ｙ，Ｚ）＝（２，２，７）であり、Ｘ＜Ｚで▲２▼のケースに相当する。この場合、続行テーブルでの判定結果は「２」レベルとなり、注目セルに関して正解となる多値レベル判定を選択できる。これが、本発明の「第４の原理」となる。この現象は、「注目セルに関して先行セルと続行セルとの関係で、記録マークの面積が大きいセルの方が、光学的な影響として現れる符号間干渉量が大きい」ためである。
【０１０４】
次にケース▲１▼▲２▼で示された関係以外の場合、即ち、Ｘ＝Ｚの場合のレベル判定の原理について説明する。表１中のＮｏ．２２９が、（Ｘ，Ｙ，Ｚ）＝（６，１，６）であり、Ｘ＝Ｚのケースになる。この場合、先行テーブルによる判定結果は「１」レベルで不正解、続行テーブルによる判定結果は「０」レベルで正解となっている。
【０１０５】
前述の図３及び図４に示した各関係で、例えば、前レベルが「０」で、注目セルのレベル（図中では「中心レベル」と表記）が「０」レベルの組合せの分布は、続行セルの記録マークの影響により変動幅を持っている。続行セルの記録マークの判定値が判らない場合の学習結果（先行テーブル）として、続行セルの記録マークの判定値を無視した統計処理で、これらの分布（Ｘ＝０，Ｙ＝０，Ｚは任意）の平均値（正規化された再生信号値）を学習結果として取り扱うことにする。具体的な計算式は、
ＬＲ（Ｘｉ,Ｙｊ）＝｛（Ｘｉ,Ｙｊ）組合せの正規化された再生信号値の総和｝／（組合せ数）
で表現される。但し、ＸｉとＹｊは、多値レベル信号の取り得る範囲で、８値記録の場合であれば、０≦Ｘｉ，Ｙｊ≦７になる。
【０１０６】
同様にして、続行テーブルの場合は、
ＬＲ（Ｙｉ,Ｚｊ）＝｛（Ｙｉ,Ｚｊ）組合せの正規化された再生信号値の総和｝／（組合せ数）
となる。
【０１０７】
これらの学習結果は、予め、３つの連続するセル１の既知の全組合せパタンを数回繰り返して記録された領域を再生し、統計演算処理することによって得ることが望ましい。
【０１０８】
この学習結果で、予め得られた結果
先行テーブル学習結果：ＬＲ（Ｘｉ,Ｙｊ）
続行テーブル学習結果：ＬＲ（Ｙｉ,Ｚｊ）
を今後、「理想値」として表現する。Ｘ＝Ｚの場合、実測された再生信号値をｙとし、各テーブルの学習結果から算出した理想値を先行テーブルではｙ１、続行テーブルではｙ２とするとき、理想値に対する誤差が小さい方の判定結果を採用することで、Ｘ＝Ｚの場合でも、高い正解確率で判定することができることが判った。この追加原理を、本発明の「第５の原理」と呼ぶ。
【０１０９】
具体的には、実測データから作成した表１中のＮｏ．２２９の場合、ｙ＝0.84590、ｙ１＝0.900420、ｙ２＝0.894110であり、各々の誤差の大小関係は、
先行テーブル結果 絶対値｜ｙ−ｙ１｜＝0.054521＞ 続行テーブル結果 絶対値｜ｙ−ｙ１２｜＝0.048209
となり、理想値に対する誤差が小さい続行テーブルの判定結果が選択できる。このように第４の原理と第５の原理とを利用して注目セルの多値レベルを判定することにより、先行テーブルだけ、或いは、続行テーブルだけを参照する場合に比べると、多値レベルの判定確率を更に高めることが可能になる。
【０１１０】
また、図２に示したように再生信号を８値レベルに判定しようとすると一意的に決定できない場合でも、図６（ａ）（ｂ）に示すように記録マークの多値数を減らすことにより、一意的な決定が可能になる。図６（ａ）は、３ビット中、最下位ビットを０に固定することにより、８値を４値に減らした例を示し、図６（ｂ）は、３ビット中、最下位ビットを1に固定することにより、８値を４値に減らした例を示している。
【０１１１】
１つ前のセルの情報を元に注目セルのレベルの再生を繰り返したような場合には、１つ前のセルを欠陥等で読み誤ると次のセルにも誤りが連鎖し、次々と読み誤ってしまう可能性があるので、図１（ｃ）に示すように、所定のセル毎に予め多値レベル数を下げたマークを記録し、多値レベル数を下げたセルは、単独で再生し、その他のセルは隣接する前後のセルのレベル情報を元に再生することにより、読み誤りの連鎖を防止することができる。図１中に示す例では、８値、８値、４値の組合せでデータが記録され、８値、８値、４値の組合せで再生される（実際には、４値セルは“０００”“０１０”“１００”“１１０”の４種類のどれかが記録され、再生時には図１（ｄ）に示すように最下位ビットの“０”が省かれて再生される。）この効果で、連鎖的なエラーを防ぐことができる。
【０１１２】
このように判定の連鎖的なエラーを防止し、セルの円周方向長さを、約０．４５μｍと更に記録密度を高めた結果、本発明の第４の原理と第５の原理とに矛盾する結果が出始めることが判った。この原因を調査したところ、各セル１の多値レベル信号をＡ／Ｄ変換する際のサンプリングタイミング誤差が原因であった。サンプリング誤差のため、セル１の中心でサンプリングができず、符号間干渉のバランスが理想の状態からずれたためで、注目セルだけで判定したセルのレベル判定結果に誤差が生じやすくなっている。このような状態の基で判定確率を上げるために、第４の原理を次のように改良した。
【０１１３】
前述したような先行テーブルと続行テーブルの学習から、下記の結果を求める。
【０１１４】
先行テーブル学習結果：ＬＲ（Ｘｉ,Ｙｊ）
続行テーブル学習結果：ＬＲ（Ｙｉ,Ｚｊ）
そして、実測された再生信号値をｙとし、各テーブルの学習結果から算出した理想値を先行テーブルではｙ１、続行テーブルではｙ２とするとき、理想値に対する誤差が小さい方の判定結果を採用することで、高い正解確率で判定することができることが判った。この第４の原理の改良原理を、本発明において「第６の原理」と呼ぶ。
【０１１５】
【表２】

【０１１６】
具体的には、実測データから作成した表２中のＮｏ．１０６の場合、ｙ＝0.82589、ｙ１＝0.878425、ｙ２＝0.798644であり、各々の誤差の関係は、
先行テーブル結果 絶対値｜ｙ−ｙ１｜＝0.052538＞ 続行テーブル結果 絶対値｜ｙ−ｙ２｜＝0.027242
となり、理想値に対する誤差が小さい続行テーブルの判定結果が選択でき、判定結果は正解となる。また、表２のＮｏ．２０２の場合、ｙ＝0.84643、ｙ１＝0.878425、ｙ２＝0.821804であり、各々の誤差の関係は、
先行テーブル結果 絶対値｜ｙ−ｙ１｜＝0.031990＞ 続行テーブル結果 絶対値｜ｙ−ｙ２｜＝0.024631
となり、理想値に対する誤差が小さい続行テーブルの判定結果が選択でき、判定結果は正解となる。
【０１１７】
この２つの例で判るように、第６の原理を利用して注目セルの多値レベルを判定することにより、第４の原理と第５の原理との組合せに対し、多値レベルの判定確率を更に高めることが可能になる。但し、第４の原理と第５の原理との組合せに比べ、常に誤差判定を行う必要があり計算量が増えるため、高い計算処理速度が必要になる。また、本発明の原理方法で再判定を行ってフィードバックした結果を更に再判定に用いることで、誤判定の確率を下げられることは言うまでもない。
【０１１８】
つづいて、上述した本発明の実施の形態の原理に基づく実施の形態を数例挙げて説明する。
【０１１９】
本発明の第一の形態を図７ないし図９に基づいて説明する。本実施の形態は、第１の原理を利用したものである。
【０１２０】
まず、本実施の形態に用いられる情報記録再生装置の構成例及び動作の概要を図７を参照して説明する。スピンドルモータ２により記録媒体３を回転させながら、ＬＤ（レーザダイオード）４から出射されたレーザビームが光学ヘッド５により記録媒体３上のトラックに沿って照射される。記録媒体３からの反射光はＰＤ（フォトディテクタ）６により検出され、ＯＰアンプ７、サーボ回路８により、ビームを情報トラック上に安定して照射するためにフォーカス、トラッキング制御される。またＰＤ６により検出された再生信号ａは、ＲＦアンプ９により増幅された後、セクタ同期検出手段１０とＡ／Ｄ変換手段１１とよりなるデジタルデータ変換手段１２によりデジタルデータｃに変換される。
【０１２１】
なお、図７中には示していないが、デジタルデータｃは、再生されたデジタルデータの値が変動しないように前述したような正規化処理がなされる。図1では分かりやすいように最大レベルが1となるように正規化したが、図７ではその後のデータ処理がしやすいようにデジタルデータが整数値になるように正規化される。１０ｂｉｔのデジタルデータとして処理する場合には、
｛（ＲＦ−min）／（max−min）｝×１０２４ ………（２）
で正規化される。
【０１２２】
このデジタルデータcを多値レベル判定手段１３により検出し、各セル１の多値レベルに相当するセルレベルデータeに変換し、セルデータ変換手段１４によりセルレベルデータを再生情報gに変換することにより、多値記録された記録媒体３から情報を再生する。
【０１２３】
また、記録時には、入力情報ｈを多値レベル変換手段１５によりセルレベルデータeに変換し、記録パルス生成手段１６により記録、消去パルスの光量とタイミング信号を生成してＬＤ駆動回路１７に供給することにより、ＬＤ４で記録媒体３上に記録マークを記録する。
【０１２４】
このような構成の下、本実施の形態では、第１の原理を利用し、1つ前のセルレベルで注目セルの判定基準を判定基準変更手段により切換え変更させることにより、レベル判定誤りを防止するようにしたものである。このため、多値レベル判定手段１３は、図８に示すように、セルレベルの判定を行う演算手段１８と、１つ前のセルのセルレベルデータeを保持するセルレベル保持手段（多値レベル保持手段）１９とにより構成されている。演算手段１８は、入力されるデジタルデータ（再生レベル）毎のレベルデータが記録されたテーブルＲＡＭ２０と、１つ前のセルレベルfとデジタルデータｃの値とに基づきテーブルＲＡＭ２０からセルレベルデータｅを読み出すアドレス生成手段２１とにより構成されている。
【０１２５】
判定基準変更手段として機能するテーブルＲＡＭ２０には、図９に示すようにデジタルデータの値に相当するアドレスにセルレベルデータが、１つ前のセル（先行セル）のセルレベルデータ毎に分離されて保存されている。よって、例えば、１つ前のセルのセルレベルデータが“２”で、デジタルデータが“１３１”の場合、“２＊１０２４＋１３１”のアドレスを読めばセルレベルデータ1が読み出される。即ち、１つ前のセルのセルレベルデータから生成したオフセットを加えたデジタルデータのアドレスを読み出すことにより、１つ前のセル毎にセルレベルの判定基準を変えてセルレベルを生成できる。この1つのセルから再生される３bitのセルレベルデータ列をセルデータ変換手段１４により集めて８bitのデータ列の再生情報ｇとなる。これにより、従来では、１つの再生ビーム内に３セル入ってしまい、1セル毎の再生では符号間干渉の影響により再生誤りを起こすような微小セルでも、誤りなく再生できるものである。
【０１２６】
なお、テーブルＲＡＭ２０のデータは、記録媒体３のリードインエリア等、ユーザデータの記録領域とは別の領域に、３セルの全ての組合せ（８＊８＊８＝５１２通り）のデータを予め記録し、これを再生して各々の場合のセルレベルを求め、再生前にテーブルＲＡＭ２０に記録したものである。もっとも、テーブルを予め用意しておき、０レベル等でオフセット値を調整することで、反射率変動等をキャンセルしても良い。
【０１２７】
本発明の第二の実施の形態を図１０ないし図１９に基づいて説明する。本実施の形態では、図１に示した例のように、８値の多値記録を基本とし、８値−８値−４値の組合せでユーザデータが記録されている場合への適用例を示す。本実施の形態は、多値レベルの判定に、第４の原理及び第５の原理を利用したもので、1つ前のセル（先行セル）のセルレベルと１つ後のセル（続行セル）のセルレベルとを参照して注目セルのセルレベルの判定基準を切換えることにより、レベル判定誤りの防止を図ったものである。
【０１２８】
本実施の形態の情報記録再生装置の構成例を図１０を参照して説明する。基本的には、図７に示した情報記録再生装置の場合と同様であるが、ＰＤ６により検出された再生信号ａは、ＲＦアンプ９により増幅され、波形等化処理回路（図示せず）で符号間干渉が抑圧された後、サンプリング同期検出手段２２とＡ／Ｄ変換手段２３とにより構成されたデジタル化手段２４により、デジタルデータｃに変換される。ここには記載していないが、デジタルデータｃは、再生されたデジタルデータの値が変動しないように前述のような正規化処理がなされる。この場合も、前述の（２）式により正規化される。
【０１２９】
このデジタルデータｃをデジタル保持手段（デジタルデータ保持手段）２５により一時保存して、サンプリングされたセル群の信号データｅをＣＰＵ２６によって１セル単位で選択的に引き出せるようにする。
【０１３０】
ここで、多値データを再生するために、多値レベル判定を行う基準テーブルを３つ（注目セル判定テーブル、先行セル参照テーブル、続行セル参照テーブル）作成する。８値記録では、３つの連続するセルの組合せは、８^３＝５１２通り存在する。そこで、３つのマーク（セル）を１つのグループとする全ての組合せパタンを既知の多値レベルの学習パタン情報として、学習パタン記録手段（図示せず）により、記録媒体３のユーザデータが記録されない最内周領域に記録した後、学習手段（図示せず）により再生し、３連続セルの中心のセルでサンプリングされたデジタルデータｃをデジタル保持手段２５により一時保存する。保存されたデータは、図１２で示すような配列で、図１１に示す一時保存用ＲＡＭ２７に保存されており、アドレス生成手段２８により一時保存用ＲＡＭ２７のアドレスデータを指定することで、選択的にデータを読み出せる状態になっている。
【０１３１】
まず、一時保存用ＲＡＭ２７に保存されているこの情報を基に、先行セルのレベル（前レベル）と続行セルのレベル（後レベル）を無視した注目セル判定テーブル２９を作成する。例えば、注目セルのセルレベルが“０”では、前レベル数８値×後レベル数８値＝６４通りの組合せがある。各レベルの判定に利用する基準信号ｙ0は、これらの６４通りの平均値とする。注目セルのセルレベルが“１”〜“７”でも同じように計算することで、図１２中に示すような注目セル判定テーブル２９を作成できる。図１０に示した構成において、一時保存用ＲＡＭ２７に一時保存された学習パタン（５１２通り）を順次読込み、ＣＰＵ２６により先述の方法で計算処理した結果をＲＡＭ３０に記録した。この処理が、基準テーブル作成手段の機能としてＣＰＵ２６により実行される。
【０１３２】
なお、４値のセルに関する注目セル判定テーブル３２ａ，３２ｂも同様に作成される。図１３（ａ）に示す注目セル判定テーブル３２ａは図６（ａ）に対応する偶数値の場合、図１３（ｂ）に示す注目セル判定テーブル３２ｂは図６（ｂ）に対応する奇数値の場合に各々対応する。
【０１３３】
次に、先行セル参照テーブル３３の作成方法について説明する。注目セル判定テーブル２９の場合と同じように、３連続セルの全ての組合せ記録から得られ一時保存されたセルの信号データｅを選択的に読み出す。先行セル参照テーブル３３では、続行するセルのレベルを無視したテーブルを作成する。
【０１３４】
図１４を参照してその作成方法を説明する。例えば、先行セルのセルレベルが“０”で、注目セルのセルレベルが“０”の組合せでは、図１２の場合と同じ配列で一時保存用ＲＡＭ２７に一時保存されたデータから、○で囲んだ８個のデータを平均値処理し、セルレベルの判定に利用する基準信号ｙ1を算出する。この結果、続行セルのレベルを無視した６４通りの組合せを有する先行セル参照テーブル３３を作成できる。注目セルのセルレベルが“１”〜“７”でも同じように計算することで、図１４に示すようなテーブルが作成できる。図１０に示した構成において、一時保存用ＲＡＭ２７に一時保存された学習パタン（５１２通り）を順次読込み、ＣＰＵ２６により先述の方法で計算処理した結果をＲＡＭ３０に記録した。この処理が、基準テーブル作成手段の機能としてＣＰＵ２６により実行される。
【０１３５】
なお、４値のセルに関する先行セル参照テーブル３４ａ，３４ｂも同様に作成される。図１５（ａ）に示す先行セル参照テーブル３４ａは図６（ａ）に対応する偶数値の場合、図１５（ｂ）に示す先行セル参照テーブル３４ｂは図６（ｂ）に対応する奇数値の場合に各々対応する。
【０１３６】
さらに、続行セル参照テーブル３５の作成方法について説明する。注目セル判定テーブル２９の場合と同じように、３連続セルの全ての組合せ記録から得られ一時保存されたセルの信号データeを選択的に読み出す。続行セル参照テーブル３５では、先行するセルのレベルを無視したテーブルを作成する。
【０１３７】
図１６を参照してその作成方法を説明する。例えば、続行セルのセルレベルが“０”で、注目セルのセルレベルが“０”の組合せでは、図１２の場合と同じ配列で一時保存用ＲＡＭ２７に一時保存されたデータから、○で囲んだ８個のデータを平均値処理し、レベルの判定に利用する基準信号ｙ２を算出する。この結果、先行セルのセルレベルを無視した６４通りの組合せを有する続行セル参照テーブル３５を作成できる。注目セルのセルレベルが“１”〜“７”でも同じように計算することで、図１６に示すようなテーブルが作成できる。図１０に示した構成において、一時保存用ＲＡＭ２７に一時保存された学習パタン（５１２通り）を順次読込み、ＣＰＵ２６により先述の方法で計算処理した結果をＲＡＭ３０に記録した。この処理が、基準テーブル作成手段の機能としてＣＰＵ２６により実行される。
【０１３８】
なお、４値のセルに関する続行セル参照テーブル３６ａ，３６ｂも同様に作成される。図１７（ａ）に示す続行セル参照テーブル３６ａは図６（ａ）に対応する偶数値の場合、図１７（ｂ）に示す続行セル参照テーブル３６ｂは図６（ｂ）に対応する奇数値の場合に各々対応する。
【０１３９】
このようにして得られた学習結果（注目セル判定テーブル２９、先行セル参照テーブル３３、続行セル参照テーブル３５）が記録されたＲＡＭ３０を参照しながら、第４の原理と第５の原理との組合せを使って、多値レベルを判定する方法について説明する。第４の原理と第５の原理との組合せでＣＰＵ２６により実行される多値レベルを判定する処理例を図１８の概略フローチャートに示す。
【０１４０】
再生したセル１の信号データ列は、セル単位でデジタル化され、一時保存用ＲＡＭ２７に一時保存される。このデジタル化された信号値をｙとする。ここに、注目セルＣｎに先行する先行セルＣn-1のデジタルデータＳＬn-1の読込み（ステップＳ１）、そのセルレベルＣＬn-1の判定（Ｓ２）、注目セルＣｎのデジタルデータＳＬnの読込み（Ｓ３）、そのセルレベルＣＬnの判定（Ｓ４）、注目セルＣｎに続行する続行セルＣn+1のデジタルデータＳＬn+1の読込み（Ｓ５）、そのセルレベルＣＬn+1の判定（Ｓ６）を順次行う。ここに、ステップＳ４では、注目セル判定テーブル２９でｙ値に最も近いセルレベルＣＬnを決定するもので、このステップＳ４の処理が一次演算手段の機能として実行される。また、ステップＳ６の処理が続行セルレベル演算手段の機能として実行される。そして、セルレベルデータＣＬnと信号値ｙとを、データの関係付けが判るようにアドレスを指定して判定値保持手段としてのＲＡＭ３１に書き込む。
【０１４１】
次に、ＲＡＭ３１から連続する３つのセルレベル情報（ＣＬn-1，ＣＬn，ＣＬn+1）と注目セルＣnのデジタル化された信号値ｙとを読み出し、第４の原理と第５の原理とに基づいて、セルレベルＣＬn-1，ＣＬn+1の大小関係を比較し（Ｓ７）、注目セルＣnに関する再判定処理を行う。
【０１４２】
ここに、比較の結果、ＣＬn-1＞ＣＬn+1であれば、先行セル参照テーブル３３を参照することにより、デジタルデータＳＬnに最も近い値ｙ１を持つ注目セルレベル値を注目セルＣnのレベル判定値ＤＬとして決定する（Ｓ８）。また、比較の結果、ＣＬn-1＜ＣＬn+1であれば、続行セル参照テーブル３５を参照することにより、デジタルデータＳＬnに最も近い値ｙ２を持つ注目セルレベル値を注目セルＣnのレベル判定値ＤＬとして決定する（Ｓ９）。
【０１４３】
一方、比較の結果、ＣＬn-1＝ＣＬn+1であれば、一旦、先行セル参照テーブル３３を参照してデジタルデータＳＬnに最も近い値ｙ１を持つ注目セルレベル値を注目セルＣnのレベル判定値ＤＬ１とするとともに、続行セル参照テーブル３５を参照してデジタルデータＳＬnに最も近い値ｙ２を持つ注目セルレベル値を注目セルＣnのレベル判定値ＤＬ２とする（Ｓ１０）。そして、これらのレベル判定値ＤＬ１，ＤＬ２の大小関係を比較する。ＤＬ１＝ＤＬ２であれば、注目セルＣnのレベル判定値ＤＬとしてＤＬ１又はＤＬ２を任意に選択設定し（Ｓ１１）、注目セルＣnのセルレベルＣＬｎをＤＬに確定し、確定されたＤＬをＲＡＭ３１に書換える（Ｓ１２）。
【０１４４】
また、ＤＬ１≠ＤＬ２の場合には、｜ｙ−ｙ１｜と｜ｙ−ｙ２｜との大小関係に応じて、｜ｙ−ｙ１｜＜｜ｙ−ｙ２｜であれば、注目セルＣnのレベル判定値ＤＬとしてＤＬ１を選択設定し、｜ｙ−ｙ１｜＞｜ｙ−ｙ２｜であれば、注目セルＣnのレベル判定値ＤＬとしてＤＬ２を選択設定し（Ｓ１３）、注目セルＣnのセルレベルＣＬｎをＤＬに確定し、確定されたＤＬをＲＡＭ３１に書換える（Ｓ１２）。
【０１４５】
これらのステップＳ７〜Ｓ１３の処理が参照判定値選択手段及び再判定演算手段の機能として実行される。
【０１４６】
なお、４値レベルのセルに関しては、図１９に示すように、注目セルのデジタルデータＳＬnを読込み（Ｓ２１）、注目セル判定テーブル３２ａ又は３２ｂを参照してｙに最も近いｙ０値を持つセルレベル値を判定値ＤＬnとする（Ｓ２２）ものであり、注目セル自身のみで判定する。
【０１４７】
そして、このように書換えられた多値データで次以降のセルの多値判定を順次行うことにより、誤判定の確率を更に下げることができる。
【０１４８】
なお、各ＲＡＭ２７，３０，３１への書き込み及び読み込み、演算処理のタイミングは、図１０中に示すサンプリング同期検出手段２２によるサンプリングタイミングに同期して処理される。
【０１４９】
本発明の第三の実施の形態を図２０に基づいて説明する。本実施の形態では、図１に示した例のように、８値の多値記録を基本とし、８値−８値−４値の組合せでユーザデータが記録されている場合への適用例を示す。本実施の形態は、多値レベルの判定に、第６の原理を利用することで、レベル判定誤りの防止を図ったものである。また、第二の実施の形態の場合と同様に、図１０に示した情報記録再生装置を使用し、かつ、多値レベル判定を行うための基準テーブルを３つ（注目セル判定テーブル２９、先行セル参照テーブル３３、続行セル参照テーブル３５）作成するものとする。
【０１５０】
学習結果（注目セル判定テーブル２９、先行セル参照テーブル３３、続行セル参照テーブル３５）が記録されたＲＡＭ３０を参照しながら、第６の原理に基づき多値レベルを判定する方法について図２０に示すフローチャートを参照して説明する。ステップＳ１〜Ｓ６の処理は、図１８の場合と同様に行なわれる。
【０１５１】
ステップＳ６の後、仮判定処理として、一旦、先行セル参照テーブル３３を参照してデジタルデータＳＬnに最も近い値ｙ１を持つ注目セルレベル値を注目セルＣnのレベル判定値ＤＬ１とするとともに、続行セル参照テーブル３５を参照してデジタルデータＳＬnに最も近い値ｙ２を持つ注目セルレベル値を注目セルＣnのレベル判定値ＤＬ２とする（Ｓ３１）。この後、セルレベルＣＬn-1，ＣＬn+1の大小関係を比較し（Ｓ３２）、注目セルＣnに関する再判定処理を行う。
【０１５２】
比較の結果、｜ｙ−ｙ１｜＝｜ｙ−ｙ２｜であれば、レベル判定値は当初のＣＬnのままとする（Ｓ３３）。｜ｙ−ｙ１｜＜｜ｙ−ｙ２｜であれば、注目セルＣnのレベル判定値ＤＬとしてＤＬ１を選択設定し（Ｓ３４）、注目セルＣnのセルレベルＣＬｎをＤＬに確定し、確定されたＤＬをＲＡＭ３１に書換える（Ｓ３５）。｜ｙ−ｙ１｜＞｜ｙ−ｙ２｜であれば、注目セルＣnのレベル判定値ＤＬとしてＤＬ２を選択設定し（Ｓ３６）、注目セルＣnのセルレベルＣＬｎをＤＬに確定し、確定されたＤＬをＲＡＭ３１に書換える（Ｓ３５）。
【０１５３】
そして、このように書換えられた多値データで次以降のセルの多値判定を順次行うことにより、誤判定の確率を更に下げることができる。
【０１５４】
なお、各ＲＡＭ２７，３０，３１への書き込み及び読み込み、演算処理のタイミングは、図１０中に示すサンプリング同期検出手段２２によるサンプリングタイミングに同期して処理される。
【０１５５】
これらの第二及び第三の実施の形態の効果の度合いを、注目セルのみによりレベル判定だけの場合と比較した結果を表３に示す。表３中に記述した誤り率は、セル単位での誤り率である。また、記録密度はセル長で決まるが、セル長が短いほど記録密度が高い。
【０１５６】
【表３】

【０１５７】
表３に示す結果によれば、記録密度によって、誤り率の低下効果は変わるが、第４の原理と第５の原理との組合せではセル単位の誤判定率を約０．２〜０．６倍に低下、第６の原理では誤判定率を約０．１〜０．４倍に低下できることが判る。
【０１５８】
このように、注目セルの前後に位置する先行セルや続行セルのセル情報と関連付けてレベル判定を行う場合、「第４の原理と第５の原理との組合せ」と「第６の原理」で、大幅に誤判定を改善できることが確認できたものである。
【０１５９】
本発明の第四の実施の形態を図２１及び図２２に基づいて説明する。本実施の形態では、図１に示した例のように、８値の多値記録を基本とし、８値−８値−４値の組合せでユーザデータが記録されている場合への適用例を示すが、多値レベルの判定方法としては、前述の如く、参照テーブルを切換える方式ではなく、３連続の記録マークの組合せ全てのパタンを学習した結果に基づいて判定を行うようにしたものである。本実施の形態の学習テーブル３７の構成例を図２１に示す。この学習テーブル３７は、先述の３つのテーブル（注目セル判定テーブル２９、先行セル参照テーブル３３、続行セル参照テーブル３５）を演算処理する前の状態のものである。以降、この学習テーブル３７を３連続記録マーク学習テーブルと記す。
【０１６０】
学習結果（３連続記録マーク学習テーブル３７）が記録されたＲＡＭ３０を参照しながら、多値レベルを判定する方法について図２２に示すフローチャートにより説明する。ステップＳ１〜Ｓ６の処理は前述の場合と同様に行なわれる。
【０１６１】
ステップＳ６の処理の後、注目セルのレベルの再判定を行う上で、前後のセルレベルデータＣＬn-1，ＣＬn+1の組合せからＣＬｎの８個の候補値を図２１に示すように選択し、学習テーブル３７より理想値ｙ０〜ｙ７を読み込み（Ｓ４１）、実測値ｙに最も近い値を持つＣＬｎを注目セルＣnの判定値ＤＬとする（Ｓ４２）。図２１では、ＣＬn-1＝１，ＣＬn+1＝５の組合せ時の、注目セルレベルの参照値８個を学習テーブル３７より抜き出している様子を示している。このような再判定によりセルレベルＣＬｎとしてＤＬを確定し、ＲＡＭ３１においてセルレベルＣＬｎを確定されたＤＬに書換える（Ｓ４３）。
【０１６２】
そして、このように書換えられた多値データで次以降のセルの多値判定を順次行うことにより、誤判定の確率を更に下げることができる。
【０１６３】
なお、各ＲＡＭ２７，３０，３１への書き込み及び読み込み、演算処理のタイミングは、図１０中に示すサンプリング同期検出手段２２によるサンプリングタイミングに同期して処理される。
【０１６４】
本実施の形態の方法によるセル単位の誤り率を測定したところ、セル長０．４０μｍで、約４％程度と注目セルによる判定に対し、誤り率が改善されることが判った。
【０１６５】
本発明の第五の実施の形態を図２３ないし図２７に基づいて説明する。前述した各実施の形態では、先行セル、続行セル、或いは、先行セル及び続行セルと注目セルとの関係を予め学習し、その学習結果を示す学習テーブルを利用することで注目セルの判定を行うようにしたが、本実施の形態では、このような判定処理を複数の連続するセル列で実行して複数の正解候補列を選び出し、その中で最も誤差が小さくなる組合せを注目セルの判定値とする判定方法を提供するものである。
【０１６６】
本実施の形態の考え方を、図２３を参照して説明する。まず、初期値として多値レベルが既知のものから開始し、この既知の初期値から先行セル参照テーブル３３を参照して、３つの候補（ｙとｙiの誤差が最も小さいセルレベル値と、そのセルレベル値に対して＋１のレベル値、そのセルレベル値に対して−１のレベル値）を選択する。同じように、これらの３個の候補を既知値として、この既知値から先行セル参照テーブル３３を参照して、更に３個の候補を選択する。ここまでの演算で、注目セルレベルと続行セルレベルとの組合せとして９個の候補の組合せが抽出できる。
【０１６７】
この９個の候補の組合せの各々に対し、学習テーブルと観測値との誤差の２乗和Σ（σi0^２＋σij^２）を計算する。但し、１≦ｉ，ｊ≦３である（図２４参照）。この誤差の２乗和が最小となる組合せを計算し、この時の注目セルレベル判定値を正解とする。この演算を逐次繰返し行うことで、更に正解の確率を上げることができる。また、本実施の形態では、候補選びを注目セルと続行セルの組合せまで行ったが、図２６及び図２７に示すように、続行セルに続く次のセルまで含めて繰り返せば、更に正解の確率が上がることは言うまでもない。
【０１６８】
本実施の形態では、先行セルレベルによる参照テーブル（先行セル参照テーブル３３）を用いたが、注目セルだけの判定をまず行い、この判定結果に基づき、前述した如く、先行セル参照テーブル３３と続行セル参照テーブル３５とを選択切換える方式、又は、３連続記録マーク学習テーブル３７を用いて、候補値及び誤差計算を行えば、同じように判定の正解率を上げることが可能である。
【０１６９】
３連続記録マーク学習テーブル３７を用い、注目セルと続行セルの候補値の組合せを抽出し、正解の組合せを選択するアルゴリズムを図２５のフローチャートに示す。ステップＳ１〜Ｓ６，Ｓ４１の処理は図２２により前述した場合と同様である。ステップＳ４１の処理の後、ｙとｙｉとの比較に基づき候補値Ｘ１０，Ｘ２０，Ｘ３０を選択し（Ｓ５１）、ｙとｙｉとの誤差σ１０，σ２０，σ３０を算出する（Ｓ５２）。その後、一つの候補値Ｘ１０を仮にＣＬnとしてＣＬn-1とＣＬn+1との関係から注目セルレベルの比較参照値ｙ０〜ｙ７を抜き出す（Ｓ５３）。そして、ｙとｙｉとの比較から候補値Ｘ１１，Ｘ１２，Ｘ１３を選択し（Ｓ５４）、ｙとｙｉとの誤σ１１，σ１２，σ１３を算出する（Ｓ５５）。
【０１７０】
別の候補値Ｘ２０に関しても、仮にＣＬnとしてＣＬn-1とＣＬn+1との関係から注目セルレベルの比較参照値ｙ０〜ｙ７を抜き出す（Ｓ５６）。そして、ｙとｙｉとの比較から候補値Ｘ２１，Ｘ２２，Ｘ２３を選択し（Ｓ５７）、ｙとｙｉとの誤σ２１，σ２２，σ２３を算出する（Ｓ５８）。同様に、候補値Ｘ３０に関しても、仮にＣＬnとしてＣＬn-1とＣＬn+1との関係から注目セルレベルの比較参照値ｙ０〜ｙ７を抜き出す（Ｓ５９）。そして、ｙとｙｉとの比較から候補値Ｘ３１，Ｘ３２，Ｘ３３を選択し（Ｓ６０）、ｙとｙｉとの誤σ３１，σ３２，σ３３を算出する（Ｓ６１）。
【０１７１】
そして、これらの算出結果に基づき、誤差の２乗和が最小となる組合せパスを正解とし、この時のＣＬnを判定値ＤＬnとして決定する（Ｓ６２）。
【０１７２】
なお、計算量を減らす方法としては、前述の誤差の２乗和が最小となる組合せとして、注目セルだけでなく、この組合せの続行セルも正解として処理すればよい。即ち、次の処理で続行セルを既知として取扱い、逐次演算を開始すれば良い。また、本実施の形態では、３つの候補の組合せ「ｙとｙiとの誤差が最も小さいセルレベル値と、そのセルレベル値に対して＋１のレベル値、そのセルレベル値に対して−１のレベル値」を利用したが、ｙに近い２つの候補を使えば、計算量を減らすことができる。この場合、注目セルと続行セルの候補の組合せとして４つとなり、計算量を半分以下にできる。
【０１７３】
なお、これらの実施の形態では、ユーザが情報を記録再生できるシステムについて説明したが、記録媒体として再生専用のＲＯＭディスクにおいても同じである。例えば、注目セル判定テーブル２９、先行セル参照テーブル３３、続行セル参照テーブル３５を作成するための学習パタンを、ユーザが使用するＲＯＭ情報と合わせて、位相ピット（ディスク基板に凹凸として形成されたピット）として形成しておく。この学習パタンを、先述と同じように読み出すことで、ＲＯＭ専用のディスクにおいても、レベルの誤判定を防止することが可能になる。
【０１７４】
また、前述した実施の形態では、学習領域をユーザデータが記録されない最内周領域に設けたが、ユーザデータがブッロク単位（例えば、２５６セル単位＝２５６＊３bits単位）で記録される先頭に設けたりしても構わない。この方が、学習しながら記録再生ができるので、より誤判定の確率を下げることができる。
【０１７５】
【発明の効果】
請求項１記載の発明によれば、注目セルの再生信号の出力レベルが先行セルや続行セルのマークの影響を受ける場合でも、これらの先行セルや続行セルの多値判定値を参照して注目セルの多値レベルを判定することで、注目セルの多値レベルの誤判定の可能性を減らすことができ、この際、記録媒体のフォーマットに制約を加える等の不具合もない。又、注目セルの前後に位置する先行セルや続行セルのマークが、注目セルの多値判定に与える影響を、符号間干渉の大きさから判断して、用いる多値判定値を選択切換えして注目セルの多値レベルを再判定するようにしたので、再生信号が前後に位置するセルのマークの影響を受ける場合でも、注目セルの多値レベルの誤判定の可能性を減らすことができ、多値判定の信頼性を高めることができる。
【０１８０】
請求項２記載の発明によれば、請求項１記載の情報再生方法において、注目セルの前後に位置する先行セルや続行セルのマークが、注目セルの多値判定に与える符号間干渉量の大きさを判断するために、各セルからの再生信号で多値判定が行われた判定結果に基づき予め判断するようにしたので、先行セル、続行セル、又は、先行セル及び続行セルの多値判定値のうちで何れかの多値判定値を参照するかの選択を適正に行うことができ、多値判定の信頼性を高めることができる。
【０１８１】
請求項３記載の発明によれば、請求項２記載の情報再生方法において、符号間干渉の程度を有効に活用して、先行セルの多値判定値を参照する場合の選択を適正に行うことができ、多値判定の信頼性を高めることができる。
【０１８２】
請求項４記載の発明によれば、請求項２又は３記載の情報再生方法において、符号間干渉の程度を有効に活用して、続行セルの多値判定値を参照する場合の選択を適正に行うことができ、多値判定の信頼性を高めることができる。
【０１８３】
請求項５記載の発明によれば、請求項２，３又は４記載の情報再生方法において、符号間干渉の程度を有効に活用して、符号間干渉量がほぼ同等な場合の判定処理を適正に行うことができる。
【０１８４】
請求項６記載の発明によれば、請求項２ないし５の何れか一記載の情報再生方法において、先行セルや続行セルのマークが、注目セルのマークの判定値に与える符号間干渉量の大きさを、学習結果から予測される信号値と実測値との誤差を比較することで判断するようにしたので、各セルのマークからの再生信号で多値判定が行われた判定値で符号間干渉量の大きさを予め判断した結果があいまいであっても、注目セルのマークの多値判定を適正に行うことができ、多値判定の信頼性を高めることができる。
【０１８５】
請求項７記載の発明によれば、請求項１記載の情報再生方法において、注目セルの前後に位置する先行セルや続行セルのマークが、注目セルの多値判定に与える符号間干渉量の大きさを判断するために、学習結果から予測される信号値と実測値との誤差の比較結果を利用するようにしたので、先行セル、続行セル、又は、先行セル及び続行セルの多値判定値のうちで何れかの多値判定値を参照するかの選択を適正に行うことができ、多値判定の信頼性を高めることができる。
【０１８６】
請求項８記載の発明によれば、請求項７記載の情報再生方法において、符号間干渉の程度を有効に活用して、注目セルの多値判定値の決定を適正に行うことができ、多値判定の信頼性を高めることができる。
【０１８７】
請求項９記載の発明によれば、請求項７又は８記載の情報再生方法において、符号間干渉の程度を有効に活用して、注目セルの多値判定値の決定を適正に行うことができ、多値判定の信頼性を高めることができる。
【０１８８】
請求項１０記載の発明によれば、請求項１ないし９の何れか一記載の情報再生方法において、再判定した結果をフィーバックしているので、注目セルの前後に位置する先行セルや続行セルのマークが、注目セルのマークの判定値に与える符号間干渉量の大きさを判断する結果の信頼性が更に向上し、多値判定の信頼性を高めることができる。
【０１８９】
請求項１１記載の発明によれば、請求項１ないし１０の何れか一記載の情報再生方法において、学習パタン情報から得られる再生信号のレベル値情報に基づき作成された基準テーブルを用いて注目セルの多値判定を行うことにより、請求項５ないし１４記載の情報再生方法を容易に実現でき、また、予め既知の多値レベルの学習パタン情報が記録されている情報領域を再生し統計演算処理した結果により基準テーブルを作成しているので、記録・再生光パワーの変動や基板歪などの外乱影響を学習でき、この外乱影響による誤判定が修正可能な基準テーブルを作成することができる。
【０１９０】
請求項１２記載の発明によれば、請求項１１記載の情報再生方法において、基準テーブルとして、注目セル判定テーブルと、先行セル参照テーブルと、続行セル参照テーブルとを用いることにより、符号間干渉量の大きさ等の判断を適正に行うことができ、結果として、多値判定の信頼性を高めることができる。
【０１９１】
請求項１３記載の発明によれば、請求項１２記載の情報再生方法において、極力少ない演算処理で基準テーブルを作成することがてきる。
【０１９５】
請求項１４記載の発明の情報再生装置によれば、注目セルの前後に位置する先行セルや続行セルのマークが、注目セルの多値判定に与える影響を、符号間干渉の大きさから判断して、用いる多値判定値を選択切換えして注目セルの多値レベルを再判定演算手段により再判定することにより、再生信号が前後に位置するセルのマークの影響を受ける場合でも、注目セルの多値レベルの誤判定の可能性を減らすことができ、多値判定の信頼性を高めることができる。
【０１９６】
請求項１５記載の発明によれば、請求項１４記載の情報再生装置において、注目セルの前後に位置する先行セルや続行セルのマークが、注目セルの多値判定に与える符号間干渉量の大きさを判断するために、各セルからの再生信号で多値判定が行われた判定結果に基づき予め判断するようにしているので、先行セル、続行セル、又は、先行セル及び続行セルの多値判定値のうちで何れかの多値判定値を参照するかの選択を適正に行うことができ、多値判定の信頼性を高めることができる。
【０１９７】
請求項１６記載の発明によれば、請求項１５記載の情報再生装置において、符号間干渉の程度を有効に活用して、先行セルの多値判定値を参照する場合の選択を適正に行うことができ、多値判定の信頼性を高めることができる。
【０１９８】
請求項１７記載の発明によれば、請求項１５又は１６記載の情報再生装置において、符号間干渉の程度を有効に活用して、続行セルの多値判定値を参照する場合の選択を適正に行うことができ、多値判定の信頼性を高めることができる。
【０１９９】
請求項１８記載の発明によれば、請求項１５，１６又は１７記載の情報再生装置において、符号間干渉の程度を有効に活用して、符号間干渉量がほぼ同等な場合の判定処理を適正に行うことができる。
【０２００】
請求項１９記載の発明によれば、請求項１５ないし１８の何れか一記載の情報再生装置において、先行セルや続行セルのマークが、注目セルのマークの判定値に与える符号間干渉量の大きさを、学習結果から予測される信号値と実測値との誤差を比較することで判断しているので、各セルのマークからの再生信号で多値判定が行われた判定値で符号間干渉量の大きさを予め判断した結果があいまいであっても、注目セルのマークの多値判定を適正に行うことができ、多値判定の信頼性を高めることができる。
【０２０１】
請求項２０記載の発明によれば、請求項１５記載の情報再生装置において、注目セルの前後に位置する先行セルや続行セルのマークが、注目セルの多値判定に与える符号間干渉量の大きさを判断するために、学習結果から予測される信号値と実測値との誤差の比較結果を利用しているので、先行セル、続行セル、又は、先行セル及び続行セルの多値判定値のうちで何れかの多値判定値を参照するかの選択を適正に行うことができ、多値判定の信頼性を高めることができる。
【０２０２】
請求項２１記載の発明によれば、請求項２０記載の情報再生装置において、符号間干渉の程度を有効に活用して、注目セルの多値判定値の決定を適正に行うことができ、多値判定の信頼性を高めることができる。
【０２０３】
請求項２２記載の発明によれば、請求項２０又は２１記載の情報再生装置において、符号間干渉の程度を有効に活用して、注目セルの多値判定値の決定を適正に行うことができ、多値判定の信頼性を高めることができる。
【０２０４】
請求項２３記載の発明によれば、請求項１５ないし２２の何れか一記載の情報再生装置において、再判定した結果をフィーバックしているので、注目セルの前後に位置する先行セルや続行セルのマークが、注目セルのマークの判定値に与える符号間干渉量の大きさを判断する結果の信頼性が更に向上し、多値判定の信頼性を高めることができる。
【０２０５】
請求項２４記載の発明によれば、請求項１５ないし２３の何れか一記載の情報再生装置において、学習パタン情報から得られる再生信号のレベル値情報に基づき作成された基準テーブルを用いて注目セルの多値判定を行うことにより、請求項１５ないし２３記載の情報再生装置を容易に実現でき、また、予め既知の多値レベルの学習パタン情報が記録されている情報領域を再生し統計演算処理した結果により基準テーブルを作成しているので、記録・再生光パワーの変動や基板歪などの外乱影響を学習でき、この外乱影響による誤判定が修正可能な基準テーブルを作成できる。
【０２０６】
請求項２５記載の発明よれば、請求項２４記載の情報再生装置において、基準テーブルとして、注目セル判定テーブルと、先行セル参照テーブルと、続行セル参照テーブルとを作成することにより、符号間干渉量の大きさ等の判断を適正に行うことができ、結果として、多値判定の信頼性を高めることができる。
【０２０７】
請求項２６記載の発明によれば、請求項２５記載の情報再生装置において、極力少ない演算処理で基準テーブルを作成することができる。
【０２０８】
請求項２７記載の発明の情報記録再生装置によれば、学習パタン情報から得られる再生信号のレベル値情報に基づき作成された基準テーブルを用いて注目セルの多値判定を行うことにより、請求項１５ないし２３記載の情報再生装置相当の情報記録再生装置を容易に実現でき、また、予め既知の多値レベルの学習パタン情報を実際に記録媒体に記録しこれを再生し統計演算処理した結果により基準テーブルを作成しているので、記録・再生光パワーの変動や基板歪などの外乱影響を学習でき、この外乱影響による誤判定が修正可能な基準テーブルを作成することができる。
【０２０９】
請求項２８記載の発明によれば、請求項２７記載の情報記録再生装置において、ユーザデータ領域に支障を及ぼすことなく、予め既知の多値レベルの学習パタン情報を実際に記録媒体に記録して基準テーブルの作成に狂することができる。
【０２１０】
請求項２９記載の発明によれば、請求項２７又は２８記載の情報記録再生装置において基準テーブルとして、注目セル判定テーブルと、先行セル参照テーブルと、続行セル参照テーブルとを作成することにより、符号間干渉量の大きさ等の判断を適正に行うことができ、結果として、多値判定の信頼性を高めることができる。
【０２１１】
請求項３０記載の発明によれば、請求項２９記載の情報記録再生装置において、極力少ない演算処理で基準テーブルを作成することができる。
【図面の簡単な説明】
【図１】本発明の実施の形態の原理を説明するための情報記録再生方式の一例を示すマーク形状、再生信号等の説明図である。
【図２】前後セルの影響による注目セルのレベル変化を示す特性図である。
【図３】先行セルのレベルに応じた注目セルのレベル変化を示す特性図である。
【図４】続行セルのレベルに応じた注目セルのレベル変化を示す特性図である。
【図５】先行セル及び続行セルのレベルに応じた注目セルのレベル変化を示す特性図である。
【図６】４値の場合の様子を示す特性図である。
【図７】本発明の第一の実施の形態の情報記録再生装置を示すブロック構成図である。
【図８】その多値レベル判定手段の構成例を示すブロック図である。
【図９】テーブルＲＡＭのデータ例を示す説明図である。
【図１０】本発明の第二の実施の形態の情報記録再生装置を示すブロック構成図である。
【図１１】そのデジタル保持手段の構成例を示すブロック図である。
【図１２】注目セル判定テーブル及びその作成方法を示す説明図である。
【図１３】４値の場合の注目セル判定テーブルを示す説明図である。
【図１４】先行セル参照テーブル及びその作成方法を示す説明図である。
【図１５】４値の場合の先行セル参照テーブルを示す説明図である。
【図１６】続行セル参照テーブル及びその作成方法を示す説明図である。
【図１７】４値の場合の続行セル参照テーブルを示す説明図である。
【図１８】判定処理例を示す概略フローチャートである。
【図１９】注目セルだけの処理例を示すフローチャートである。
【図２０】本発明の第三の実施の形態の判定処理例を示す概略フローチャートである。
【図２１】本発明の第四の実施の形態の３連続記録マーク学習テーブル及びそのデータ抽出方法を示す説明図である。
【図２２】判定処理例を示す概略フローチャートである。
【図２３】本発明の第五の実施の形態の判定方式を模式的に示す説明図である。
【図２４】その誤差集計アルゴリズムを示す説明図である。
【図２５】判定処理例を示す概略フローチャートである。
【図２６】判定方式の変形例を模式的に示す説明図である。
【図２７】その誤差集計アルゴリズムを示す説明図である。
【符号の説明】
１ セル
３ 記録媒体
１１ Ａ／Ｄ変換手段
１８ 演算手段
１９ 多値レベルデータ保持手段
２０ 判定基準変更手段
２３ Ａ／Ｄ変換手段
２５ デジタルデータ保持手段
２９ 注目セル判定テーブル
３１ 判定値保持手段
３３ 先行セル参照テーブル
３５ 続行セル参照テーブル
Ｓ４ 一次演算手段
Ｓ６ 続行セルレベル演算手段
Ｓ７〜Ｓ１３ 参照判定値選択手段、再判定演算手段[0001]
BACKGROUND OF THE INVENTION
The present invention relates to an information reproducing method, an information reproducing apparatus, an information recording / reproducing apparatus, and a recording medium for handling data in which marks representing multilevel information are recorded in cells.
[0002]
[Prior art]
Conventionally, in an optical disc, binary digital data is formed on a spiral or concentric track by forming embossed uneven pits (ROM disc), holes in inorganic / organic recording films (recordable disc), crystal Recording is performed by a difference in state (phase change disk), a magnetization direction (magneto-optical disk), or the like. When reproducing these recorded data, a track is irradiated with a laser beam to detect a difference in intensity of reflected light, a difference in polarization direction due to the magnetic Kerr effect, and the like to obtain a reproduced RF signal. Then, the obtained reproduction RF signal is processed with a certain threshold value, for example, to detect binary data.
[0003]
In order to increase the recording density of information, a method of recording multilevel data of three or more values instead of binary data is considered. However, in an optical disc, a difference in reflectivity between various optical discs, Level fluctuations and amplitude fluctuations occur in the playback signal due to various factors, such as differences in playback frequency characteristics between the inner and outer circumferences of the optical disc. There is a problem that the signal may be detected with an incorrect value. In particular, when the data is ternary, quaternary, etc., there is a problem that there are a plurality of threshold values, and the possibility of erroneous detection of multilevel data is increased.
[0004]
According to Japanese Patent Laid-Open No. 5-54391, in order to solve the above problem, a reproduction signal from a recording medium on which multilevel data is recorded is A / D converted based on a reference clock, and this A / D When multilevel data is detected by comparing the signal level data obtained by the conversion with a threshold value, the signal level data obtained by A / D conversion of the reproduction signal is stored in the memory means for each predetermined signal unit. Then, the distribution information of the signal level data is obtained for each predetermined signal unit stored in the memory means, and based on the distribution information, the threshold value used when the signal level data in the signal unit is multi-valued Is set. In this case, it is assumed that a format in which the ratio of the level value of the recorded data is determined is adopted.
[0005]
For example, signal level data obtained by A / D converting the reproduction signal in a predetermined signal unit, such as a reproduction signal in a sector unit in a recording format on an optical disc, is temporarily stored in a memory, and the occurrence frequency distribution information using the CPU As a result, the optimum threshold value is obtained as a multi-value standard, and the signal level data stored using the threshold value is multi-valued to detect data. For this reason, even when the level fluctuation or amplitude fluctuation of the reproduction signal occurs, there is an effect that the possibility of erroneous detection of multi-value data is reduced and an accurate multi-value detection operation is realized. High density recording data is promoted.
[0006]
In Japanese Patent Laid-Open No. 5-54391, in order to cope with fluctuations in a long section such as signal level fluctuations on the inner and outer circumferences of the optical disc, for example, the threshold value for multi-value is optimized for each sector. Also, it is assumed that the format is constrained so that each multi-value level including variations can be clearly separated.
[0007]
[Problems to be solved by the invention]
By the way, in order to improve the information recording density, it is necessary to reduce the unit of data recording / reproduction (hereinafter referred to as “information cell” or simply “cell”) along with the multi-valued data. When the information cell is made smaller and the reproduction light beam spot includes a plurality of information cells (see the description of FIG. 1 described later), the reproduction signal output level of the information cell is affected by the preceding and following information cells. Thus, the output level of the reproduction signal corresponding to the same multi-level shifts depending on which multi-level information cell is recorded before and after, and apparently the variation (spread) of each multi-level increases. . For this reason, if the data is tabulated in a long section such as a sector unit, adjacent multilevel levels have overlapping distributions. Although details will be described later, FIG. 2 shows an example of a reproduction signal level distribution when a multilevel level = 8 data levels are recorded in one information cell and the size of the reproduction light beam spot includes three information cells. It is. In this example, there is an overlap in distribution between levels 1 and 2, levels 2 and 3, and levels 4 and 5.
[0008]
Considering the application of a detection method such as that disclosed in Japanese Patent Laid-Open No. 5-54391 to such a reproduction signal, the level of the information cell with respect to distribution overlap is erroneously detected no matter how the threshold is determined. There is a problem that the possibility becomes high. Furthermore, there is a problem in that the degree of freedom of the format is limited.
[0009]
Accordingly, the present invention provides an information reproduction method, an information reproduction apparatus, and an information reproduction method capable of reducing the possibility of erroneous determination of a multilevel level when reproducing a cell in which multilevel information is recorded.as well asInformation recording and playback equipmentPlaceThe purpose is to provide.
[0010]
In particular, the present invention relates to an information reproduction method, an information reproduction device, and an information reproduction apparatus capable of preventing erroneous determination due to intersymbol interference between a cell preceding and continuing with respect to a target cell of interest.as well asInformation recording and playback equipmentPlaceThe purpose is to provide.
[0011]
Furthermore, the present invention relates to an information reproduction method and information reproduction capable of making good multilevel determination by effectively utilizing the degree of intersymbol interference between the preceding cell and the continuing cell with respect to the target cell of interest. apparatus,as well asInformation recording and playback equipmentPlaceThe purpose is to provide.
[0012]
[Means for Solving the Problems]
The invention described in claim 1 is an information reproducing method for reproducing the multi-level from a reproduction signal of a recording medium in which marks representing multi-level information are recorded in the cell, preceding the target cell of interest. The multilevel level of the cell of interest is determined by referring to the preceding cell, the continuing cell to continue, or the multilevel determination value of the preceding cell and the continuing cell.Further, according to the magnitude of the amount of intersymbol interference from the preceding cell with respect to the reproduction signal of the attention cell or the magnitude of the amount of intersymbol interference from the continuation cell with respect to the reproduction signal of the attention cell. A multilevel decision value is selected from among the multilevel decision values of a cell, the continuation cell, or the preceding cell and the continuation cell, and the multilevel level of the target cell is selected with reference to the multilevel decision value Re-determined.
[0013]
Therefore, even when the output level of the reproduction signal is affected by the marks of the preceding cell and the continuation cell, the multilevel level of the cell of interest is determined by referring to the multilevel determination values of these preceding cells and continuation cells. The possibility of misjudgment of the multilevel level of the cell of interest can be reduced. At this time, there is no problem such as restriction on the format of the recording medium.In addition, the influence of the preceding cell and continuation cell marks positioned before and after the target cell on the multilevel determination of the target cell is judged from the magnitude of intersymbol interference, and the multilevel determination value to be used is selectively switched. By re-determining the multi-value level of the cell of interest, the possibility of misjudgment of the multi-value level of the cell of interest can be reduced even when the playback signal is affected by the marks of the cells located in the front and back. The reliability of determination can be increased.
[0022]
Claim2The described invention is claimed.1In the information reproducing method described above, the magnitude of the intersymbol interference amount is determined based on the multilevel determination result based on the level of the reproduction signal from each cell.
[0023]
Therefore, in order to determine the amount of intersymbol interference that the marks of the preceding cell and the continuation cell positioned before and after the target cell give to the multilevel determination of the target cell, the multilevel determination is performed on the reproduction signal from each cell. Since it is determined in advance based on the determination result that has been made, it is possible to select whether to refer to any of the multilevel determination values of the preceding cell, the continuation cell, or the multilevel determination values of the preceding cell and the continuation cell. Can be performed appropriately, and the reliability of multi-level determination can be improved.
[0024]
Claim3The described invention is claimed.2In the information reproduction method described above, when the amount of intersymbol interference from the preceding cell with respect to the reproduction signal of the target cell is larger than the amount of intersymbol interference from the continuation cell with respect to the reproduction signal of the target cell In this case, the multi-value level of the target cell is re-determined with reference to the multi-value determination value of the preceding cell.
[0025]
Therefore, the degree of intersymbol interference can be effectively used to appropriately select the reference for the multilevel determination value of the preceding cell, and the reliability of the multilevel determination can be improved.
[0026]
Claim4The described invention is claimed.2Or3In the information reproduction method described above, when the amount of intersymbol interference from the preceding cell with respect to the reproduction signal of the target cell is smaller than the amount of intersymbol interference from the continuation cell with respect to the reproduction signal of the target cell In this case, the multi-value level of the target cell is re-determined with reference to the multi-value determination value of the continuing cell.
[0027]
Therefore, it is possible to appropriately utilize the degree of intersymbol interference and appropriately select the reference for the multilevel determination value of the continuation cell, and to improve the reliability of the multilevel determination.
[0028]
Claim5The described invention is claimed.2,3Or4In the described information reproduction method, the amount of intersymbol interference from the preceding cell with respect to the reproduction signal of the target cell is substantially the same as the amount of intersymbol interference from the continuation cell with respect to the reproduction signal of the target cell, The multi-value determination value of the target cell re-determined with reference to the multi-value determination value of the preceding cell is Y1, and the multi-value determination of the target cell re-determined with reference to the multi-value determination value of the continuation cell When these values are Y2 and the multilevel determination values Y1 and Y2 match, the multilevel determination value of the target cell is determined to be Y1 or Y2.
[0029]
Therefore, it is possible to appropriately perform the determination process when the amount of intersymbol interference is substantially equal by effectively utilizing the degree of intersymbol interference.
[0030]
Claim6The described invention is claimed.2Or5The amount of intersymbol interference from the preceding cell with respect to the reproduction signal of the target cell and the amount of intersymbol interference from the continuation cell with respect to the reproduction signal of the target cell in the information reproduction method according to any one of The target cell re-determined by referring to the multi-value determination value of the attention cell Y1 which is substantially the same, and re-determined with reference to the multi-value determination value of the preceding cell, and the multi-value determination value of the continuing cell. If the multi-value determination values Y1 and Y2 of the multi-value determination value Y1 and Y2 do not coincide with each other, the learned level value of the reproduction signal predicted from the multi-value determination value Y1 of the re-determined cell is obtained. When y1, the learned level value of the reproduction signal predicted from the multi-value determination value Y2 of the re-determined cell of interest is y2, and the level value of the actually measured reproduction signal of the cell of interest is y, -Y1 | and | y-y2 | If the multilevel determination value of the cell of interest is | y−y1 | <| y−y2 |, the multilevel determination value Y1 and | y−y1 |> | y−y2 | The multi-value determination value Y2 is determined.
[0031]
Therefore, the amount of intersymbol interference that the mark of the preceding cell or continuation cell gives to the determination value of the mark of the target cell is determined by comparing the error between the signal value predicted from the learning result and the actual measurement value. Therefore, even if the result of pre-determining the amount of intersymbol interference is ambiguous with the decision value obtained by performing multi-value decision on the reproduction signal from the mark of each cell, the multi-value decision of the mark of the cell of interest Can be performed appropriately, and the reliability of multi-level determination can be improved.
[0032]
Claim7The described invention is claimed.1In the information reproduction method described above, after performing multilevel determination at the level of a reproduction signal from each cell, prediction is performed from the multilevel determination value of the target cell re-determined with reference to the multilevel determination value of the preceding cell A learned signal level value of the reproduced signal, a learned signal level value of the reproduced signal predicted from the multi-valued decision value of the cell of interest re-determined with reference to the multi-valued decision value of the continuing cell, and the attention The magnitude of the amount of intersymbol interference is determined by comparison with the level value of the reproduced signal actually measured in the cell.
[0033]
Therefore, in order to determine the amount of intersymbol interference that the marks of the preceding cell and the continuation cell positioned before and after the cell of interest give to the multilevel determination of the cell of interest, the signal value predicted from the learning result and the actual measurement value Since the error comparison result is used, the selection of which one of the multilevel judgment values of the preceding cell, the continuation cell, or the multilevel judgment values of the preceding cell and the continuation cell is to be referred to is properly performed. This can be performed, and the reliability of multi-value determination can be improved.
[0034]
Claim8The described invention is claimed.7In the information reproducing method described above, the learned level value of the reproduction signal predicted from the multi-value determination value Y1 of the re-determined target cell is y1, and the multi-value determination value Y2 of the re-determined target cell is As a result of comparing the magnitude relationship between | y−y1 | and | y−y2 |, where y2 is the learned learned level value of the reproduced signal and y is the level value of the actually measured reproduction signal of the target cell. , | Y−y1 | <| y−y2 |, the multilevel determination value of the target cell is determined as the multilevel determination value Y1.
[0035]
Therefore, it is possible to appropriately determine the multilevel determination value of the cell of interest by effectively utilizing the degree of intersymbol interference, and to improve the reliability of multilevel determination.
[0036]
Claim9The described invention is claimed.7Or8In the information reproducing method described above, the learned level value of the reproduction signal predicted from the multi-value determination value Y1 of the re-determined target cell is y1, and the multi-value determination value Y2 of the re-determined target cell is As a result of comparing the magnitude relationship between | y−y1 | and | y−y2 |, where y2 is the learned learned level value of the reproduced signal and y is the level value of the actually measured reproduction signal of the target cell. , | Y−y1 |> | y−y2 |, the multilevel determination value of the cell of interest is determined as the multilevel determination value Y2.
[0037]
Therefore, it is possible to appropriately determine the multilevel determination value of the cell of interest by effectively utilizing the degree of intersymbol interference, and to improve the reliability of multilevel determination.
[0038]
Claim10The described invention is claimed.1Or9In the information reproduction method according to any one of the above, when the multi-value determination value of the target cell is rewritten by re-determination, the rewritten multi-value determination value is used for multi-value determination after the continuation cell.
[0039]
Therefore, since the result of re-determination is fed back, the result of determining the amount of intersymbol interference that the marks of the preceding cell and the continuation cell positioned before and after the cell of interest give to the determination value of the mark of the cell of interest The reliability of the multi-value determination can be improved.
[0040]
Claim11The described invention is claimed.1Or10In the information reproducing method according to any one of the above, the learning pattern information is reproduced from a recording medium in which all combination patterns having three marks as one group are recorded in advance as learning pattern information of a known multilevel level. A reference table is created on the basis of the level value information of the reproduction signal obtained from the learning pattern information, and this reference table is used for multi-value determination of the target cell.
[0041]
Therefore, by performing multilevel determination of the cell of interest using a reference table created based on the level value information of the reproduction signal obtained from the learning pattern information,1Or10The described information reproducing method can be easily realized. In addition, since a reference table is created based on the result of statistical calculation processing by reproducing an information area in which known multilevel learning pattern information is recorded in advance, disturbances such as fluctuations in recording / reproducing optical power and substrate distortion A reference table can be created in which the influence can be learned and the erroneous determination due to the influence of the disturbance can be corrected.
[0042]
Claim12The described invention is claimed.11In the described information reproduction method, the reference table includes a target cell determination table, a preceding cell reference table, and a continuation cell reference table.
[0043]
Therefore, by using the attention cell determination table, the preceding cell reference table, and the continuation cell reference table as the reference table, it is possible to appropriately determine the magnitude of the intersymbol interference amount, etc. Reliability of value determination can be improved.
[0044]
Claim13The described invention is claimed.12In the information reproduction method described above, the target cell determination table is created by an arithmetic process ignoring the level values of the preceding cell and the continuation cell among the level value information of the reproduction signal obtained from the learning pattern information, and the preceding cell The reference table is created by a calculation process ignoring the level value of the continuation cell in the level value information of the reproduction signal obtained from the learning pattern information, and the continuation cell reference table is a reproduction signal obtained from the learning pattern information. The level value information is created by an arithmetic process ignoring the level value of the preceding cell.
[0045]
Therefore, it is possible to create the reference table with as little arithmetic processing as possible.
[0052]
Claim14The invention described is an information reproducing apparatus for reproducing the multi-level from a reproduction signal of a recording medium in which marks representing multi-level information are recorded in a cell, and converting the reproduction signal into digital data. D conversion means, digital data holding means for holding the converted digital data, and primary calculation means for determining based on the reproduction signal of only the mark of the target cell regarding the multi-value level of the converted digital data Determination value holding means for temporarily holding the multi-value determination value by the primary calculation means, and continuation cell level calculation means for determining the multi-value level of the continuation cell continuing to the attention cell prior to the calculation of the attention cell. , The amount of intersymbol interference from the preceding cell with respect to the reproduction signal of the target cell, or between the codes from the continuation cell with respect to the reproduction signal of the target cell Reference determination value selection means for selecting any one of the multi-value determination values of the preceding cell, the continuation cell, or the multi-value determination values of the preceding cell and the continuation cell according to the amount of interference And a re-determination calculating means for re-determining the multi-value level of the cell of interest with reference to the multi-value determination value selected by the reference determination value selecting means.
[0053]
Therefore, the influence of the marks of the preceding cell and the continuation cell located before and after the target cell on the multilevel determination of the target cell is judged from the magnitude of intersymbol interference, and the multilevel determination value to be used is selectively switched. By re-determining the multi-value level of the cell of interest by the re-determination computing means, the possibility of erroneous determination of the multi-value level of the cell of interest is reduced even when the reproduction signal is affected by the marks of the cells located before and after It is possible to improve the reliability of multi-valued determination.
[0054]
Claim15The described invention is claimed.14In the information reproducing apparatus described above, the reference determination value selection unit determines the magnitude of the intersymbol interference amount based on a multilevel determination result based on a level of a reproduction signal from each cell.
[0055]
Therefore, in order to determine the amount of intersymbol interference that the marks of the preceding cell and the continuation cell positioned before and after the target cell give to the multilevel determination of the target cell, the multilevel determination is performed on the reproduction signal from each cell. Since it is determined in advance based on the determination result that has been made, it is possible to select whether to refer to any of the multilevel determination values of the preceding cell, the continuation cell, or the multilevel determination values of the preceding cell and the continuation cell. Can be performed appropriately, and the reliability of multi-level determination can be improved.
[0056]
Claim16The described invention is claimed.15In the information reproducing apparatus described above, the reference determination value selecting unit is configured such that the amount of intersymbol interference from the preceding cell with respect to the reproduction signal of the target cell is equal to the code interval from the continuation cell with respect to the reproduction signal of the target cell. When the amount of interference is larger than the amount of interference, the multi-value determination value of the preceding cell is selected.
[0057]
Therefore, the degree of intersymbol interference can be effectively used to appropriately select the reference for the multilevel determination value of the preceding cell, and the reliability of the multilevel determination can be improved.
[0058]
Claim17The described invention is claimed.15Or16In the information reproducing apparatus described above, the reference determination value selecting unit is configured such that the amount of intersymbol interference from the preceding cell with respect to the reproduction signal of the target cell is equal to the code interval from the continuation cell with respect to the reproduction signal of the target cell. When the amount of interference is smaller than the amount of interference, the multi-value determination value of the continuing cell is selected.
[0059]
Therefore, it is possible to appropriately utilize the degree of intersymbol interference and appropriately select the reference for the multilevel determination value of the continuation cell, and to improve the reliability of the multilevel determination.
[0060]
Claim18The described invention is claimed.15,16Or17In the information reproducing apparatus described above, the reference determination value selection unit includes: an amount of intersymbol interference from the preceding cell with respect to the reproduction signal of the attention cell; and an amount of intersymbol interference from the continuation cell with respect to the reproduction signal of the attention cell. If the size of the preceding cell is substantially the same, the multi-value determination value of the preceding cell and the multi-value determination value of the continuation cell are selected, and the re-determination calculation means refers to the multi-value determination value of the preceding cell. These multi-valued determination values Y1 are defined as Y1 as the multi-value determination value of the target cell re-determined in Y, and Y2 as the multi-value determination value of the target cell re-determined with reference to the multi-value determination value of the continuation cell. , Y2 match, the multi-value determination value of the target cell is determined as Y1 or Y2.
[0061]
Therefore, it is possible to appropriately perform the determination process when the amount of intersymbol interference is substantially equal by effectively utilizing the degree of intersymbol interference.
[0062]
Claim19The described invention is claimed.15Or18In the information reproduction apparatus according to any one of the above, the reference determination value selection unit includes a magnitude of an intersymbol interference amount from the preceding cell with respect to the reproduction signal of the attention cell and a value from the continuation cell with respect to the reproduction signal of the attention cell. If the magnitude of the intersymbol interference amount is substantially the same, the multi-value determination value of the preceding cell and the multi-value determination value of the continuation cell are selected, and the re-determination computing means determines the multi-value determination of the preceding cell. The multi-value determination value of the target cell re-determined with reference to the value is Y1, and the multi-value determination value of the target cell re-determined with reference to the multi-value determination value of the continuation cell is Y2. When the value determination values Y1 and Y2 do not match, the learned level value of the reproduction signal predicted from the multi-value determination value Y1 of the re-determined target cell is y1, and the re-determined target cell of the re-determined target cell Predicted from multi-value judgment value Y2 When the learned level value of the raw signal is y2 and the level value of the measured reproduction signal of the target cell is y, the magnitude relationship between | y−y1 | and | y−y2 | If the multi-value determination value is | y−y1 | <| y−y2 |, the multi-value determination value Y1 is determined. If | y−y1 |> | y−y2 |, the multi-value determination value Y2 is determined.
[0063]
Therefore, the amount of intersymbol interference that the mark of the preceding cell or continuation cell gives to the determination value of the mark of the target cell is determined by comparing the error between the signal value predicted from the learning result and the actual measurement value. Therefore, even if the result of pre-determining the amount of intersymbol interference is ambiguous with the decision value obtained by performing multi-value decision on the reproduction signal from the mark of each cell, the multi-value decision of the mark of the cell of interest Can be performed appropriately, and the reliability of multi-level determination can be improved.
[0064]
Claim20The described invention is claimed.15In the information reproducing apparatus described above, the reference determination value selection unit performs the multi-value determination at the level of the reproduction signal from each cell, and then re-determines the attention by referring to the multi-value determination value of the preceding cell The learned level value of the reproduction signal predicted from the multilevel determination value of the cell and the reproduction signal predicted from the multilevel determination value of the target cell re-determined with reference to the multilevel determination value of the continuing cell The magnitude of the intersymbol interference amount is determined by comparing the learned level value with the level value of the reproduced signal actually measured in the target cell.
[0065]
Therefore, in order to determine the amount of intersymbol interference that the marks of the preceding cell and the continuation cell positioned before and after the cell of interest give to the multilevel determination of the cell of interest, the signal value predicted from the learning result and the actual measurement value Since the error comparison result is used, the selection of which one of the multilevel judgment values of the preceding cell, the continuation cell, or the multilevel judgment values of the preceding cell and the continuation cell is to be referred to is properly performed. This can be performed, and the reliability of multi-value determination can be improved.
[0066]
Claim21The described invention is claimed.20In the information reproducing apparatus described above, the re-determination calculating means sets the learned level value of the reproduction signal predicted from the multi-value determination value Y1 of the re-determined attention cell to y1, and the re-determination of the re-determination of the attention cell. When y2 is the learned level value of the reproduction signal predicted from the multi-level determination value Y2, and y is the level value of the reproduction signal actually measured in the cell of interest, | y−y1 | and | y−y2 | If | y−y1 | <| y−y2 | as a result of the comparison of the magnitude relationships, the multilevel determination value of the target cell is determined as the multilevel determination value Y1.
[0067]
Therefore, it is possible to appropriately determine the multilevel determination value of the cell of interest by effectively utilizing the degree of intersymbol interference, and to improve the reliability of multilevel determination.
[0068]
Claim22The described invention is claimed.20Or21In the information reproducing apparatus described above, the re-determination calculating means sets the learned level value of the reproduction signal predicted from the multi-value determination value Y1 of the re-determined attention cell to y1, and the re-determination of the re-determination of the attention cell. When y2 is the learned level value of the reproduction signal predicted from the multi-level determination value Y2, and y is the level value of the reproduction signal actually measured in the cell of interest, | y−y1 | and | y−y2 | If | y−y1 |> | y−y2 | is determined as a result of the comparison of the magnitude relationships, the multilevel determination value of the target cell is determined as the multilevel determination value Y2.
[0069]
Therefore, it is possible to appropriately determine the multilevel determination value of the cell of interest by effectively utilizing the degree of intersymbol interference, and to improve the reliability of multilevel determination.
[0070]
Claim23The described invention is claimed.15Or22In the information reproducing apparatus according to any one of the above, when the multi-value determination value of the cell of interest is rewritten by re-determination, the re-determination calculation unit uses the re-written multi-value determination value as a multi-value after the continuation cell. Used for judgment.
[0071]
Therefore, since the result of re-determination is fed back, the result of determining the amount of intersymbol interference that the marks of the preceding cell and the continuation cell positioned before and after the cell of interest give to the determination value of the mark of the cell of interest The reliability of the multi-value determination can be improved.
[0072]
Claim24The described invention is claimed.15Or23In the information reproducing apparatus according to any one of the above, the learning pattern information is reproduced from a recording medium in which all combination patterns having three marks as one group are recorded in advance as learning pattern information of a known multilevel level. Learning means and reference table creating means for creating a reference table based on the level value information of the reproduction signal obtained from the learning pattern information by the learning means, and this reference table is used for multi-value determination of the cell of interest. I used it.
[0073]
Therefore, by performing multilevel determination of the cell of interest using a reference table created based on the level value information of the reproduction signal obtained from the learning pattern information,15Or23The described information reproducing apparatus can be easily realized. In addition, since a reference table is created based on the result of statistical calculation processing by reproducing an information area in which known multilevel learning pattern information is recorded in advance, disturbances such as fluctuations in recording / reproducing optical power and substrate distortion A reference table can be created in which the influence can be learned and the erroneous determination due to the influence of the disturbance can be corrected.
[0074]
Claim25The described invention is claimed.24In the described information reproducing apparatus, the reference table includes a target cell determination table, a preceding cell reference table, and a continuation cell reference table.
[0075]
Therefore, by creating an attention cell determination table, a preceding cell reference table, and a continuation cell reference table as a reference table, it is possible to appropriately determine the amount of intersymbol interference, and as a result, The reliability of multilevel determination can be improved.
[0076]
Claim26The described invention is claimed.25In the information reproducing apparatus described above, the reference table creating unit obtains the target cell determination table by a calculation process ignoring the level values of the preceding cell and the continuation cell among the level value information of the reproduction signal obtained from the learning pattern information. The level of the reproduced signal obtained from the learning pattern information is created by creating the preceding cell reference table by calculation processing ignoring the level value of the continuation cell among the level value information of the reproduced signal obtained from the learning pattern information. The continuation cell reference table is created by calculation processing ignoring the level value of the preceding cell in the value information.
[0077]
Therefore, the reference table can be created with as little arithmetic processing as possible.
[0078]
Claim27The invention described is an information recording / reproducing apparatus for reproducing the multi-level from a reproduction signal of a writable recording medium in which marks representing multi-level information are recorded in a cell, wherein the reproduction signal is converted into digital data. A / D conversion means for conversion, digital data holding means for holding the converted digital data, and determination based on the reproduction signal of only the mark of the target cell regarding the multi-value level of the converted digital data Primary calculation means, determination value holding means for temporarily holding a multi-value determination value by the primary calculation means, and a continuation cell for determining a multi-value level of a continuation cell continuing to the attention cell prior to the calculation of the attention cell Level calculation means, the amount of intersymbol interference from the preceding cell with respect to the reproduction signal of the target cell, or the continuation with respect to the reproduction signal of the target cell A reference for selecting one of the multilevel determination values of the preceding cell, the continuation cell, or the multilevel determination values of the preceding cell and the continuation cell according to the amount of intersymbol interference from the cell. A determination value selection unit, a re-determination operation unit that re-determines the multi-value level of the cell of interest with reference to the multi-value determination value selected by the reference determination value selection unit, and three marks as one group A learning pattern recording means for recording all combination patterns as known multi-level learning pattern information on the recording medium, a learning means for reproducing the learning pattern information from the recording medium, and these learning patterns by the learning means. And a reference table creating means for creating a reference table used for multi-value determination of the cell of interest based on level value information of the reproduction signal obtained from the information.
[0079]
Therefore, by performing multilevel determination of the cell of interest using a reference table created based on the level value information of the reproduction signal obtained from the learning pattern information,15Or23An information recording / reproducing apparatus equivalent to the described information reproducing apparatus can be easily realized. In addition, since a reference table is created based on the result of statistical calculation processing by actually recording learning pattern information of a known multi-value level in advance on a recording medium and reproducing it, fluctuations in recording / reproducing optical power, substrate distortion, etc. Can be learned, and a reference table capable of correcting erroneous determination due to this disturbance effect can be created.
[0080]
Claim28The described invention is claimed.27In the information recording / reproducing apparatus described above, the learning pattern recording unit records all combination patterns outside the user data area of the recording medium.
[0081]
Therefore, it is possible to actually record the learning pattern information of a known multi-value level in advance on the recording medium without disturbing the user data area, so that the creation of the reference table can be mistaken.
[0082]
Claim29The described invention is claimed.27Or28In the described information recording / reproducing apparatus, the reference table includes a target cell determination table, a preceding cell reference table, and a continuation cell reference table.
[0083]
Therefore, by creating an attention cell determination table, a preceding cell reference table, and a continuation cell reference table as a reference table, it is possible to appropriately determine the amount of intersymbol interference, and as a result, The reliability of multilevel determination can be improved.
[0084]
Claim30The described invention is claimed.29In the information recording / reproducing apparatus described above, the reference table creating unit performs the attention cell determination table by a calculation process in which the level values of the preceding cell and the continuation cell are ignored among the level value information of the reproduction signal obtained from the learning pattern information. And the preceding cell reference table is created by calculation processing ignoring the level value of the continuation cell among the level value information of the reproduction signal obtained from the learning pattern information, and the reproduction signal obtained from the learning pattern information The continuation cell reference table is created by calculation processing ignoring the level value of the preceding cell in the level value information.
[0085]
Therefore, the reference table can be created with as little arithmetic processing as possible.
[0090]
DESCRIPTION OF THE PREFERRED EMBODIMENT
First, the principle of the embodiment of the present invention will be described.
[0091]
FIG. 1 is an explanatory diagram showing an example of an information recording / reproducing system of the present invention. A recording medium having a tracking groove formed on the surface and provided with a recording material capable of recording and erasing records information as a mark at predetermined time intervals. An area corresponding to the predetermined time interval is a unit of multi-value information recording / reproduction called information cell (or simply cell) 1. For example, when a phase change (PC) recording film is used as a recordable erasable recording material, the cell 1 is irradiated with laser light, and the recording and erasing light quantity and their timing are adjusted to change the shape of the recording mark. As a result, recording marks of a plurality of reproduction levels are formed. FIG. 1A shows an example in which cells 1 having eight values from 0 level where no recording mark is formed to 7 level where the largest mark is formed are recorded as multilevel levels. That is, each cell 1 has 8 = 2³Therefore, 3-bit information is recorded. If the multilevel level of each cell 1 can be reproduced as recorded, a normal optical disk is recorded with 2 bits, so that a recording capacity three times as large can be obtained. Further, in order to increase the storage capacity, it is necessary to make the cell 1 small. If it is made small, the mark of 2 to 3 cells 1 is included in the reproduction beam diameter as shown in FIG. Become.
[0092]
In addition to the phase change film, a magneto-optical (MO) recording film can also be used as a recordable erasable recording material. In addition to the laser beam described above, the shape of the recording mark can be obtained by cooperating with a magnetic field from a magnetic head (not shown). Is changed to form recording marks of a plurality of reproduction levels. Furthermore, it is also possible to apply a recording material that can only be recorded and reproduced, and an organic dye or a metal film can be used as the recording material. By irradiating the cell with laser light and adjusting the recording light amount and their timing, By changing the shape of the recording mark, recording marks of a plurality of reproduction levels are formed. Similarly, even in a read-only recording medium, the recording mark can be formed on the substrate as a concavo-convex shape called a phase pit, and multi-level recording can be performed by modulating the area of the phase pit or the optical depth of the phase pit. Is possible.
[0093]
On the premise of such multi-level recording, the principle of the embodiment of the present invention will be described using an example in which a phase change recording film is used.
[0094]
Here, an example of multi-level recording / reproduction by the configuration of a general rewritable DVD recording apparatus will be shown. The recording / reproducing wavelength λ = 650 nm, the numerical aperture NA = 0.65 of the objective lens, and the beam diameter focused on the optical information recording medium (recording medium) is about 0.8 μm. As the optical information recording medium, a rewritable type of phase change material AgInSbTe was used. The recording mark is formed on a groove called a groove, the interval (track pitch) between adjacent grooves is 0.74 μm, the groove width is about 0.4 μm, and the linear velocity of recording / reproducing is about 3.5 m / s. Set to. The circumferential length of the cell 1 is about 0.6 μm. The erasing power is set to a value 0.5 to 0.6 times the recording power. The number of levels of multi-value information was recorded as 8 values.
[0095]
First, random information is recorded and reproduced, and the result is shown in FIG. In FIG. 2, the vertical axis indicates the reproduction level at the center of each cell, which is the level with the highest reflectance (max) where no mark is recorded and the level (min) where the mark is recorded so that the amount of reflected light is the lowest. Normalized by difference. If playback signal = RF,
Playback signal (normalized multilevel signal) = (RF−min) / (max−min) (1)
It is represented by
[0096]
According to the result shown in FIG. 2, the reproduced signal is not completely divided into eight values, and if the reproduced signal is determined to be a level from 0 to 7, one level is recorded depending on the reproduced signal. Nevertheless, there are cases where playback is performed at two levels. This is because the inter-symbol interference phenomenon occurs because the recording marks of the front and rear cells 1 are included in the reproduction beam, and the reproduction level varies depending on the size of the recording marks of the front and rear cells 1.
[0097]
On the other hand, FIG. 3 shows the data shown in FIG. 2 separated for each known multi-value level (horizontal axis in FIG. 3: previous level) of the previous cell (preceding cell). It shows how the playback level of the target cell of interest changes depending on the previous level. In the result shown in FIG. 2, the multi-value level could not be uniquely determined from the reproduction signal, but the relationship with the multi-value level of the preceding cell as shown in FIG. It can be seen that the multi-value level of the cell of interest can be uniquely determined by possessing it as a judgment criterion and disassembling each previous level of the preceding cell. The “first principle” of the present invention is based on this principle. “A standard for determining the multilevel level of the cell of interest is prepared for each multilevel data of the previous cell, and the previous cell By switching the determination criterion of the cell of interest after determination of the multi-value level ”, it is possible to prevent a level determination error caused by inter-code interference that occurs in the state of FIG.
[0098]
Similarly, in FIG. 4, the data shown in FIG. 2 is separated for each multi-value level (horizontal axis in FIG. 4: subsequent level) of the cell after the cell of interest (continuation cell). This shows how the reproduction level of the target cell of interest changes depending on the previous level. As shown in FIG. 4, the relationship with the multilevel level of the preceding cell is learned in advance, held as a multilevel level criterion, and decomposed for each subsequent level of the continuation cell, so that the It can be seen that the value level can be determined. The “second principle” of the present invention is based on this principle. “A standard for determining the multi-level of the target cell is prepared for each multi-level data of the next cell, and the cell after By switching the determination criterion of the cell of interest after determination of the multi-value level ”, it is possible to prevent a level determination error caused by inter-code interference that occurs in the state of FIG.
[0099]
Furthermore, when the multilevel levels of cells before and after the cell of interest (preceding cell and continuation cell) are known, the level of the multilevel signal to be reproduced hardly varies as shown in FIG. It can be seen that the multi-level can be uniquely determined from the reproduction signal. FIG. 5 shows the multi-value level of the cell of interest when the rear level = 0 and the previous level = 0-7. The “third principle” of the present invention is based on this, and “a standard for determining the multi-level of the cell of interest is prepared for each multi-level data of the preceding and following cells, and the cell level of the previous cell is prepared. By switching the determination criteria of the cell of interest after determining the cell level of the next cell, it is possible to prevent a level determination error caused by inter-code interference that occurs in the state of FIG. (Combination of the first principle and the second principle).
[0100]
Table 1 shows the results of actual level determination based on the review of the determination criteria based on the first to third principles. No. 2 and No. No. 8 is a result of the determination based on the reproduction signal value of the target cell, and both of the determinations are “0” level, and the correct answer is “1” level. However, if the value of the preceding table in which the determination criterion is set according to the previous cell level is referred to, or the value of the continuation table in which the determination criterion is set according to the next cell level is referred to, the determination is “1”. You can see that the answer is correct at the level.
[0101]
[Table 1]

[0102]
However, further investigation of the determination results shown in Table 1 revealed that the results determined with reference to the preceding table may not match the results determined with reference to the continuation table. The phenomenon is shown in No. 1 in Table 1. 25, the determination result referring to the preceding table is incorrect at the “3” level, and the determination result referring to the continuation table is correct at the “2” level. When many such phenomena were examined, it was found that there was the following tendency.
[0103]
When the combination of three consecutive cells 1 is (X, Y, Z), and Y is the multilevel data of the cell of interest,
(1) Referencing the preceding table, Y level judgment correct probability is high when X> Z
(2) Y level judgment correct probability is high when referring to the continuation table when X <Z
It is. However, it is assumed that the value of the multilevel level expressed by X, Y, Z is larger as the area of the mark recorded in the cell 1 is larger. No. in Table 1 25, (X, Y, Z) = (2, 2, 7), which corresponds to the case of (2) when X <Z. In this case, the determination result in the continuation table is “2” level, and it is possible to select multi-level determination that is correct for the cell of interest. This is the “fourth principle” of the present invention. This phenomenon is because “a cell with a larger recording mark area has a larger amount of intersymbol interference that appears as an optical effect due to the relationship between the preceding cell and the continuation cell with respect to the cell of interest”.
[0104]
Next, the principle of level determination in cases other than the relationship shown in case (1) (2), that is, in the case of X = Z will be described. No. in Table 1 229 is (X, Y, Z) = (6, 1, 6), and X = Z. In this case, the determination result based on the preceding table is incorrect at the “1” level, and the determination result based on the continuation table is correct at the “0” level.
[0105]
In each of the relationships shown in FIG. 3 and FIG. 4 described above, for example, the distribution of combinations in which the previous level is “0” and the level of the cell of interest (indicated as “center level” in the figure) is “0” level is There is a fluctuation range due to the effect of the recording mark of the continuation cell. As a learning result (preceding table) when the judgment value of the recording mark of the continuation cell is not known, these distributions (X = 0, Y = 0, Z are An arbitrary average value (normalized reproduction signal value) is handled as a learning result. The specific formula is
LR (Xi, Yj) = {sum of normalized reproduction signal values of (Xi, Yj) combination} / (number of combinations)
It is expressed by However, Xi and Yj are ranges that can be taken by the multilevel signal, and in the case of 8-level recording, 0 ≦ Xi and Yj ≦ 7.
[0106]
Similarly, for a continuation table:
LR (Yi, Zj) = {sum of normalized reproduction signal values of (Yi, Zj) combination} / (number of combinations)
It becomes.
[0107]
These learning results are desirably obtained by reproducing a region recorded in advance by repeating all known combination patterns of three consecutive cells 1 several times and performing statistical calculation processing.
[0108]
In this learning result, the result obtained in advance
Previous table learning result: LR (Xi, Yj)
Continue table learning result: LR (Yi, Zj)
Will be expressed as “ideal value” in the future. When X = Z, when the measured reproduction signal value is y, and the ideal value calculated from the learning result of each table is y1 in the preceding table and y2 in the continuation table, the determination result with the smaller error with respect to the ideal value By adopting, it was found that even when X = Z, it can be determined with a high probability of correct answer. This additional principle is called the “fifth principle” of the present invention.
[0109]
Specifically, No. 1 in Table 1 created from the measured data. In the case of 229, y = 0.84590, y1 = 0.900420, y2 = 0.894110, and the magnitude relation of each error is
Preceding table result absolute value | y−y1 | = 0.054521> Continuing table result absolute value | y−y12 | = 0.048209
Thus, the determination result of the continuation table with a small error with respect to the ideal value can be selected. In this way, by determining the multi-value level of the cell of interest using the fourth principle and the fifth principle, the multi-value level is compared with the case where only the preceding table or the continuation table is referred to. The determination probability can be further increased.
[0110]
Further, even if it is not possible to determine uniquely when the reproduction signal is determined to be an 8-level level as shown in FIG. 2, the number of recording marks is reduced as shown in FIGS. 6 (a) and 6 (b). Allows for a unique decision. FIG. 6A shows an example in which the eight values are reduced to four by fixing the least significant bit to 0 in 3 bits, and FIG. 6B shows the least significant bit in 1 in which the least significant bit is 1 In this example, 8 values are reduced to 4 values.
[0111]
When reproduction of the level of the target cell is repeated based on the information of the previous cell, if the previous cell is mistakenly read due to a defect or the like, errors will be chained to the next cell and read one after another. As shown in FIG. 1 (c), a mark with a lower number of multi-levels is recorded in advance for each predetermined cell, and a cell with a lower number of multi-levels is reproduced independently. However, other cells can be reproduced based on the level information of adjacent cells before and after, thereby preventing a reading error chain. In the example shown in FIG. 1, data is recorded in a combination of 8 values, 8 values, and 4 values, and is reproduced in a combination of 8 values, 8 values, and 4 values (actually, the 4-value cell is “000”). Any one of the four types of “010”, “100”, and “110” is recorded, and at the time of reproduction, the least significant bit “0” is omitted and reproduced as shown in FIG. 1 (d). Can prevent cascading errors.
[0112]
As a result of preventing chain errors in determination and increasing the recording density to about 0.45 μm in the circumferential direction of the cell, there is a contradiction between the fourth principle and the fifth principle of the present invention. It turns out that the result to begin to come out. When this cause was investigated, the cause was a sampling timing error when A / D converting the multilevel signal of each cell 1. Due to the sampling error, sampling cannot be performed at the center of the cell 1 and the balance of intersymbol interference deviates from the ideal state, so that an error is likely to occur in the cell level determination result determined only by the target cell. In order to increase the determination probability based on such a state, the fourth principle was improved as follows.
[0113]
The following result is obtained from learning of the preceding table and the continuation table as described above.
[0114]
Previous table learning result: LR (Xi, Yj)
Continue table learning result: LR (Yi, Zj)
Then, when the actually measured reproduction signal value is y and the ideal value calculated from the learning result of each table is y1 in the preceding table and y2 in the continuation table, the determination result with the smaller error with respect to the ideal value is adopted. Thus, it was found that it can be judged with a high probability of correct answer. The improvement principle of the fourth principle is referred to as “sixth principle” in the present invention.
[0115]
[Table 2]

[0116]
Specifically, No. in Table 2 created from the measured data. In the case of 106, y = 0.82589, y1 = 0.78425, and y2 = 0.798644.
Preceding table result absolute value | y−y1 | = 0.052538> Continuing table result absolute value | y−y2 | = 0.027242
Thus, the determination result of the continuation table with a small error with respect to the ideal value can be selected, and the determination result is correct. In Table 2, No. In the case of 202, y = 0.84643, y1 = 0.878425, y2 = 0.821804, and the relationship between the errors is as follows:
Preceding table result absolute value | y−y1 | = 0.031990> Continuing table result absolute value | y−y2 | = 0.024631
Thus, the determination result of the continuation table with a small error with respect to the ideal value can be selected, and the determination result is correct.
[0117]
As can be seen from these two examples, by using the sixth principle to determine the multi-value level of the cell of interest, the determination probability of the multi-value level for the combination of the fourth principle and the fifth principle Can be further increased. However, as compared with the combination of the fourth principle and the fifth principle, it is necessary to always perform error determination and the amount of calculation increases, so that a high calculation processing speed is required. Further, it goes without saying that the probability of erroneous determination can be lowered by further using the result of the re-determination by the principle method of the present invention and feeding back to the re-determination.
[0118]
Next, several embodiments based on the principle of the above-described embodiment of the present invention will be described.
[0119]
A first embodiment of the present invention will be described with reference to FIGS. The present embodiment utilizes the first principle.
[0120]
First, a configuration example and an outline of operation of the information recording / reproducing apparatus used in the present embodiment will be described with reference to FIG. While rotating the recording medium 3 by the spindle motor 2, a laser beam emitted from an LD (laser diode) 4 is irradiated along a track on the recording medium 3 by the optical head 5. Reflected light from the recording medium 3 is detected by a PD (photo detector) 6, and focus and tracking are controlled by an OP amplifier 7 and a servo circuit 8 in order to stably irradiate a beam onto an information track. The reproduction signal a detected by the PD 6 is amplified by an RF amplifier 9 and then converted into digital data c by a digital data conversion means 12 comprising a sector synchronization detection means 10 and an A / D conversion means 11.
[0121]
Although not shown in FIG. 7, the digital data c is subjected to normalization processing as described above so that the value of the reproduced digital data does not fluctuate. In FIG. 1, normalization is performed so that the maximum level is 1 for easy understanding, but in FIG. 7, digital data is normalized so as to be an integer value so that subsequent data processing is easy. When processing as 10-bit digital data,
{(RF−min) / (max−min)} × 1024 (2)
Normalized by.
[0122]
The digital data c is detected by the multilevel level determination means 13 and converted to cell level data e corresponding to the multilevel level of each cell 1, and the cell level conversion means 14 converts the cell level data into reproduction information g. Thus, information is reproduced from the recording medium 3 on which multilevel recording is performed.
[0123]
Further, at the time of recording, the input information h is converted into cell level data e by the multi-value level converting means 15, the recording pulse generating means 16 generates the light quantity and timing signal of recording and erasing pulses, and supplies them to the LD driving circuit 17 As a result, the recording mark is recorded on the recording medium 3 by the LD 4.
[0124]
Under such a configuration, in the present embodiment, the first principle is used, and the determination criterion of the cell of interest is switched and changed by the determination criterion changing means at the previous cell level, thereby preventing a level determination error. It is what you do. For this reason, as shown in FIG. 8, the multilevel level determination means 13 includes a calculation means 18 for determining the cell level and a cell level holding means (multilevel level) for holding the cell level data e of the previous cell. Holding means) 19. The calculation means 18 obtains the cell level data e from the table RAM 20 based on the table RAM 20 in which level data for each input digital data (reproduction level) is recorded, and the previous cell level f and the value of the digital data c. The address generating means 21 for reading is configured.
[0125]
In the table RAM 20 functioning as a criterion changing means, cell level data is separated into cell level data of the previous cell (preceding cell) at an address corresponding to the value of digital data as shown in FIG. Saved. Therefore, for example, when the cell level data of the previous cell is “2” and the digital data is “131”, the cell level data 1 is read by reading the address “2 * 1024 + 131”. That is, by reading the address of the digital data to which the offset generated from the cell level data of the previous cell is added, the cell level determination criteria can be changed for each previous cell to generate the cell level. The 3-bit cell level data string reproduced from this one cell is collected by the cell data conversion means 14 to become reproduction information g of the 8-bit data string. As a result, conventionally, 3 cells are included in one reproduction beam, and even a minute cell that causes a reproduction error due to the influence of intersymbol interference can be reproduced without error.
[0126]
The data in the table RAM 20 is recorded in advance in all combinations (8 * 8 * 8 = 512) of 3 cells in an area other than the user data recording area such as the lead-in area of the recording medium 3. This is reproduced to obtain the cell level in each case, and is recorded in the table RAM 20 before reproduction. However, it is also possible to cancel the reflectance fluctuation or the like by preparing a table in advance and adjusting the offset value at 0 level or the like.
[0127]
A second embodiment of the present invention will be described with reference to FIGS. In the present embodiment, as in the example shown in FIG. 1, an example of application to a case in which user data is recorded based on a combination of 8 values−8 values−4 values based on multi-value recording of 8 values. Show. In the present embodiment, the fourth and fifth principles are used for the determination of the multi-level, and the cell level of the previous cell (previous cell) and the cell after it (continuation cell) are used. The level determination error is prevented by switching the cell level determination criteria of the cell of interest with reference to the cell level.
[0128]
A configuration example of the information recording / reproducing apparatus of the present embodiment will be described with reference to FIG. Basically, it is the same as in the case of the information recording / reproducing apparatus shown in FIG. 7, but the reproduction signal a detected by the PD 6 is amplified by the RF amplifier 9 and is processed by a waveform equalization processing circuit (not shown). After the intersymbol interference is suppressed, it is converted into digital data c by the digitizing means 24 constituted by the sampling synchronization detecting means 22 and the A / D converting means 23. Although not described here, the digital data c is subjected to the normalization process as described above so that the value of the reproduced digital data does not fluctuate. Also in this case, it is normalized by the above-described equation (2).
[0129]
The digital data c is temporarily stored by the digital holding means (digital data holding means) 25 so that the sampled cell group signal data e can be selectively extracted by the CPU 26 in units of one cell.
[0130]
Here, in order to reproduce the multilevel data, three reference tables for performing multilevel level determination (target cell determination table, preceding cell reference table, and continuation cell reference table) are created. In 8-level recording, the combination of 3 consecutive cells is 8³= 512 ways exist. Therefore, the user data of the recording medium 3 is not recorded by learning pattern recording means (not shown) using all the combination patterns having three marks (cells) as one group as known multi-level learning pattern information. After being recorded in the innermost peripheral area, it is reproduced by a learning means (not shown), and digital data c sampled at the center cell of three consecutive cells is temporarily saved by the digital holding means 25. The stored data is stored in the temporary storage RAM 27 shown in FIG. 11 in the arrangement as shown in FIG. 12, and the address generation means 28 selectively specifies the address data in the temporary storage RAM 27. Data is ready to be read.
[0131]
First, based on this information stored in the temporary storage RAM 27, the attention cell determination table 29 in which the level of the preceding cell (previous level) and the level of the continuing cell (rear level) are ignored is created. For example, when the cell level of the cell of interest is “0”, there are 64 combinations of the previous level number 8 × the subsequent level number 8 = 64. The reference signal y0 used for determination of each level is an average value of these 64 patterns. By performing the same calculation even when the cell level of the target cell is “1” to “7”, the target cell determination table 29 as shown in FIG. 12 can be created. In the configuration shown in FIG. 10, the learning patterns (512 patterns) temporarily stored in the temporary storage RAM 27 are sequentially read, and the result of calculation processing by the CPU 26 by the above-described method is recorded in the RAM 30. This process is executed by the CPU 26 as a function of the reference table creating means.
[0132]
Note that the cell-of-interest determination tables 32a and 32b relating to quaternary cells are created in the same manner. The attention cell determination table 32a shown in FIG. 13A has an even value corresponding to FIG. 6A, and the attention cell determination table 32b shown in FIG. 13B has an odd value corresponding to FIG. 6B. Correspond to each case.
[0133]
Next, a method for creating the preceding cell reference table 33 will be described. As in the case of the attention cell determination table 29, the cell signal data e obtained from all the combination records of the three consecutive cells and temporarily stored is selectively read out. In the preceding cell reference table 33, a table in which the level of the cell to be continued is ignored is created.
[0134]
The creation method will be described with reference to FIG. For example, when the cell level of the preceding cell is “0” and the cell level of the target cell is “0”, the data is temporarily enclosed in the temporary storage RAM 27 in the same arrangement as in FIG. Eight pieces of data are averaged and a reference signal y1 used for cell level determination is calculated. As a result, the preceding cell reference table 33 having 64 combinations ignoring the level of the continuation cell can be created. By calculating in the same way even if the cell level of the target cell is “1” to “7”, a table as shown in FIG. 14 can be created. In the configuration shown in FIG. 10, the learning patterns (512 patterns) temporarily stored in the temporary storage RAM 27 are sequentially read, and the result of calculation processing by the CPU 26 by the above-described method is recorded in the RAM 30. This process is executed by the CPU 26 as a function of the reference table creating means.
[0135]
The preceding cell reference tables 34a and 34b relating to the quaternary cells are similarly created. When the preceding cell reference table 34a shown in FIG. 15 (a) has an even value corresponding to FIG. 6 (a), the preceding cell reference table 34b shown in FIG. 15 (b) has an odd value corresponding to FIG. 6 (b). Correspond to each case.
[0136]
Further, a method for creating the continuation cell reference table 35 will be described. As in the case of the attention cell determination table 29, the cell signal data e obtained from all the combination records of the three consecutive cells and temporarily stored is selectively read out. In the continuation cell reference table 35, a table in which the level of the preceding cell is ignored is created.
[0137]
The creation method will be described with reference to FIG. For example, when the cell level of the continuation cell is “0” and the cell level of the target cell is “0”, the data is temporarily enclosed in the temporary storage RAM 27 in the same arrangement as in FIG. Eight data are averaged and a reference signal y2 used for level determination is calculated. As a result, the continuation cell reference table 35 having 64 combinations ignoring the cell level of the preceding cell can be created. By calculating in the same way even if the cell level of the target cell is “1” to “7”, a table as shown in FIG. 16 can be created. In the configuration shown in FIG. 10, the learning patterns (512 patterns) temporarily stored in the temporary storage RAM 27 are sequentially read, and the result of calculation processing by the CPU 26 by the above-described method is recorded in the RAM 30. This process is executed by the CPU 26 as a function of the reference table creating means.
[0138]
The continuation cell reference tables 36a and 36b relating to the quaternary cells are created in the same manner. When the continuation cell reference table 36a shown in FIG. 17A has an even value corresponding to FIG. 6A, the continuation cell reference table 36b shown in FIG. 17B has an odd value corresponding to FIG. 6B. Correspond to each case.
[0139]
A combination of the fourth principle and the fifth principle with reference to the RAM 30 in which the learning results (the attention cell determination table 29, the preceding cell reference table 33, the continuation cell reference table 35) obtained in this way are recorded. A method for determining the multi-value level will be described with reference to FIG. A schematic flowchart of FIG. 18 shows a processing example for determining the multi-value level executed by the CPU 26 by a combination of the fourth principle and the fifth principle.
[0140]
The reproduced signal data string of the cell 1 is digitized in units of cells and temporarily stored in the temporary storage RAM 27. Let this digitized signal value be y. Here, reading of the digital data SLn-1 of the preceding cell Cn-1 preceding the target cell Cn (step S1), determination of the cell level CLn-1 (S2), reading of the digital data SLn of the target cell Cn (S3) ), Determination of the cell level CLn (S4), reading of the digital data SLn + 1 of the continuation cell Cn + 1 continuing to the target cell Cn (S5), and determination of the cell level CLn + 1 (S6) are sequentially performed. Here, in step S4, the cell level CLn closest to the y value is determined in the target cell determination table 29, and the process of step S4 is executed as a function of the primary calculation means. Further, the process of step S6 is executed as a function of the continued cell level calculation means. Then, the cell level data CLn and the signal value y are written in the RAM 31 as the determination value holding means by designating an address so that the relation between the data can be understood.
[0141]
Next, three consecutive cell level information (CLn-1, CLn, CLn + 1) and the digitized signal value y of the cell of interest Cn are read from the RAM 31, and the fourth and fifth principles are read out. Based on this, the magnitude relation between the cell levels CLn−1 and CLn + 1 is compared (S7), and a redetermination process regarding the target cell Cn is performed.
[0142]
Here, if CLn-1> CLn + 1 as a result of comparison, the target cell level value having the value y1 closest to the digital data SLn is determined by referring to the preceding cell reference table 33 to determine the level of the target cell Cn. The value DL is determined (S8). If CLn-1 <CLn + 1 as a result of the comparison, by referring to the continuation cell reference table 35, the target cell level value having the value y2 closest to the digital data SLn is determined as the level determination value of the target cell Cn. It is determined as DL (S9).
[0143]
On the other hand, if CLn-1 = CLn + 1 as a result of the comparison, the target cell level value having the value y1 closest to the digital data SLn is once referred to the preceding cell reference table 33, and the level determination value of the target cell Cn. In addition to DL1, the continuation cell reference table 35 is referred to, and the target cell level value having the value y2 closest to the digital data SLn is set as the level determination value DL2 of the target cell Cn (S10). Then, the magnitude relation between these level determination values DL1 and DL2 is compared. If DL1 = DL2, DL1 or DL2 is arbitrarily selected and set as the level determination value DL of the target cell Cn (S11), the cell level CLn of the target cell Cn is fixed to DL, and the determined DL is rewritten to the RAM 31 (S12).
[0144]
Also, in the case of DL1 ≠ DL2, if | y-y1 | <| y-y2 | according to the magnitude relationship between | y−y1 | and | y−y2 | If DL1 is selected and set as the value DL, and | y−y1 |> | y−y2 |, DL2 is selected and set as the level determination value DL of the target cell Cn (S13), and the cell level CLn of the target cell Cn is set. The DL is confirmed and the confirmed DL is rewritten to the RAM 31 (S12).
[0145]
The processes in steps S7 to S13 are executed as functions of the reference determination value selection unit and the re-determination calculation unit.
[0146]
As for the quaternary level cell, as shown in FIG. 19, the digital data SLn of the target cell is read (S21), and the cell level having the y0 value closest to y with reference to the target cell determination table 32a or 32b. The value is set as a determination value DLn (S22), and the determination is made only by the target cell itself.
[0147]
Then, the probability of erroneous determination can be further reduced by sequentially performing multi-value determination of the next and subsequent cells with the multi-value data rewritten in this way.
[0148]
Note that the timings of writing and reading to the RAMs 27, 30, and 31 and the arithmetic processing are synchronized with the sampling timing by the sampling synchronization detecting means 22 shown in FIG.
[0149]
A third embodiment of the present invention will be described with reference to FIG. In the present embodiment, as in the example shown in FIG. 1, an example of application to a case in which user data is recorded based on a combination of 8 values−8 values−4 values based on multi-value recording of 8 values. Show. In this embodiment, the sixth principle is used for the determination of the multi-level, thereby preventing a level determination error. Similarly to the case of the second embodiment, the information recording / reproducing apparatus shown in FIG. 10 is used, and three reference tables for performing multi-level determination (target cell determination table 29, preceding A cell reference table 33 and a continuation cell reference table 35) are created.
[0150]
FIG. 20 is a flowchart illustrating a method for determining a multi-value level based on the sixth principle while referring to the RAM 30 in which learning results (target cell determination table 29, preceding cell reference table 33, continuation cell reference table 35) are recorded. Will be described with reference to FIG. Steps S1 to S6 are performed in the same manner as in FIG.
[0151]
After step S6, as a temporary determination process, the target cell level value having the value y1 closest to the digital data SLn is temporarily set as the level determination value DL1 of the target cell Cn with reference to the preceding cell reference table 33, and the continuation cell. With reference to the reference table 35, the target cell level value having the value y2 closest to the digital data SLn is set as the level determination value DL2 of the target cell Cn (S31). Thereafter, the cell levels CLn−1 and CLn + 1 are compared in size (S32), and redetermination processing regarding the target cell Cn is performed.
[0152]
If | y−y1 | = | y−y2 | as a result of the comparison, the level determination value remains the original CLn (S33). If | y−y1 | <| y−y2 |, DL1 is selected and set as the level determination value DL of the target cell Cn (S34), the cell level CLn of the target cell Cn is determined to be DL, and the determined DL Is rewritten to the RAM 31 (S35). If | y−y1 |> | y−y2 |, DL2 is selected and set as the level determination value DL of the target cell Cn (S36), the cell level CLn of the target cell Cn is determined to be DL, and the determined DL Is rewritten to the RAM 31 (S35).
[0153]
Then, the probability of erroneous determination can be further reduced by sequentially performing multi-value determination of the next and subsequent cells with the multi-value data rewritten in this way.
[0154]
Note that the timings of writing and reading to the RAMs 27, 30, and 31 and the arithmetic processing are synchronized with the sampling timing by the sampling synchronization detecting means 22 shown in FIG.
[0155]
Table 3 shows the result of comparing the degree of the effect of the second and third embodiments with the case of only the level determination using only the target cell. The error rate described in Table 3 is an error rate in cell units. The recording density is determined by the cell length. The shorter the cell length, the higher the recording density.
[0156]
[Table 3]

[0157]
According to the results shown in Table 3, the error rate reduction effect varies depending on the recording density, but the combination of the fourth principle and the fifth principle increases the erroneous determination rate in units of cells by about 0.2 to 0.6 times. It can be seen that the false determination rate can be reduced by about 0.1 to 0.4 times in the sixth principle.
[0158]
As described above, when level determination is performed in association with the cell information of the preceding cell and the continuation cell positioned before and after the target cell, the “combination of the fourth principle and the fifth principle” and the “sixth principle” are used. It has been confirmed that the erroneous determination can be greatly improved.
[0159]
A fourth embodiment of the present invention will be described with reference to FIGS. In the present embodiment, as in the example shown in FIG. 1, an example of application to a case in which user data is recorded based on a combination of 8 values−8 values−4 values based on multi-value recording of 8 values. As shown above, the determination method of the multi-value level is not based on the method of switching the reference table as described above, but is determined based on the result of learning all the patterns of combinations of three consecutive recording marks. . A configuration example of the learning table 37 of the present embodiment is shown in FIG. This learning table 37 is in a state before the above three tables (the attention cell determination table 29, the preceding cell reference table 33, and the continuation cell reference table 35) are processed. Hereinafter, this learning table 37 is referred to as a three-continuous recording mark learning table.
[0160]
A method for determining the multi-value level will be described with reference to the flowchart shown in FIG. 22 while referring to the RAM 30 in which the learning result (three consecutive recording mark learning table 37) is recorded. Steps S1 to S6 are performed in the same manner as described above.
[0161]
After the process of step S6, when re-determining the level of the cell of interest, eight candidate values of CLn are selected from the combinations of the preceding and subsequent cell level data CLn-1 and CLn + 1 as shown in FIG. Then, the ideal values y0 to y7 are read from the learning table 37 (S41), and CLn having a value closest to the actual measurement value y is set as the determination value DL of the target cell Cn (S42). FIG. 21 shows a state where eight reference values of the target cell level are extracted from the learning table 37 when CLn−1 = 1 and CLn + 1 = 5. The DL is determined as the cell level CLn by such re-determination, and the cell level CLn is rewritten to the determined DL in the RAM 31 (S43).
[0162]
Then, the probability of erroneous determination can be further reduced by sequentially performing multi-value determination of the next and subsequent cells with the multi-value data rewritten in this way.
[0163]
Note that the timings of writing and reading to the RAMs 27, 30, and 31 and the arithmetic processing are synchronized with the sampling timing by the sampling synchronization detecting means 22 shown in FIG.
[0164]
As a result of measuring the error rate in units of cells by the method of the present embodiment, it was found that the error rate is improved with respect to the determination by the cell of interest, which is about 4% when the cell length is 0.40 μm.
[0165]
A fifth embodiment of the present invention will be described with reference to FIGS. In each of the above-described embodiments, the preceding cell, the continuation cell, or the relationship between the preceding cell and the continuation cell and the attention cell is learned in advance, and the attention cell is determined by using a learning table indicating the learning result. However, in the present embodiment, such a determination process is performed on a plurality of consecutive cell strings to select a plurality of correct candidate strings, and the combination with the smallest error among them is determined as the determination value of the target cell. A determination method is provided.
[0166]
The concept of this embodiment will be described with reference to FIG. First, a multi-level is known as an initial value, and the prior cell reference table 33 is referred to from the known initial value, and three candidates (cell level values with the smallest error between y and y i, A level value of +1 for the cell level value and a level value of -1 for the cell level value) are selected. Similarly, using these three candidates as known values, the preceding cell reference table 33 is referred to from these known values, and further three candidates are selected. Through the calculations so far, nine candidate combinations can be extracted as combinations of the target cell level and the continuation cell level.
[0167]
For each of the nine candidate combinations, the sum of squares of errors between the learning table and the observed value Σ (σi0²+ Σij²). However, 1 ≦ i and j ≦ 3 (see FIG. 24). A combination that minimizes the sum of squares of the errors is calculated, and the target cell level determination value at this time is set as a correct answer. By repeating this calculation sequentially, the probability of correct answers can be further increased. In this embodiment, candidate selection is performed up to the combination of the target cell and the continuation cell. However, as shown in FIG. 26 and FIG. 27, if it is repeated including the next cell following the continuation cell, the probability of the correct answer is further increased. Needless to say, it goes up.
[0168]
In the present embodiment, the reference table (preceding cell reference table 33) at the preceding cell level is used. However, only the cell of interest is determined first, and based on the determination result, as described above, the preceding cell reference table 33 and the continuation are continued. If the candidate value and the error calculation are performed using the method of selectively switching between the cell reference table 35 or the three continuous recording mark learning table 37, it is possible to increase the accuracy rate of the determination in the same manner.
[0169]
FIG. 25 is a flowchart showing an algorithm for extracting a combination of candidate values of a target cell and a continuation cell and selecting a correct combination using the three continuous recording mark learning table 37. The processing in steps S1 to S6 and S41 is the same as that described above with reference to FIG. After the process of step S41, candidate values X10, X20, and X30 are selected based on the comparison between y and yi (S51), and errors σ10, σ20, and σ30 between y and yi are calculated (S52). Thereafter, assuming that one candidate value X10 is CLn, comparison reference values y0 to y7 at the target cell level are extracted from the relationship between CLn-1 and CLn + 1 (S53). Then, candidate values X11, X12, and X13 are selected from the comparison between y and yi (S54), and errors σ11, σ12, and σ13 between y and yi are calculated (S55).
[0170]
Also for another candidate value X20, comparison reference values y0 to y7 at the cell level of interest are extracted from the relationship between CLn-1 and CLn + 1 as CLn (S56). Then, candidate values X21, X22, and X23 are selected from the comparison between y and yi (S57), and errors σ21, σ22, and σ23 between y and yi are calculated (S58). Similarly, for the candidate value X30, the comparative reference values y0 to y7 at the target cell level are extracted as CLn from the relationship between CLn-1 and CLn + 1 (S59). Then, candidate values X31, X32, and X33 are selected from the comparison between y and yi (S60), and errors σ31, σ32, and σ33 between y and yi are calculated (S61).
[0171]
Based on these calculation results, a combination path that minimizes the sum of squares of errors is determined as a correct answer, and CLn at this time is determined as a determination value DLn (S62).
[0172]
As a method of reducing the amount of calculation, not only the cell of interest but also a continuation cell of this combination may be processed as a correct answer as a combination that minimizes the square sum of the errors. That is, the continuation cell is handled as known in the next process, and the sequential calculation may be started. In this embodiment, the combination of the three candidates “cell level value with the smallest error between y and y i, level value +1 for the cell level value, and level −1 for the cell level value. Although the “level value” is used, the calculation amount can be reduced by using two candidates close to y. In this case, there are four combinations of the target cell and continuation cell candidates, and the amount of calculation can be reduced to half or less.
[0173]
In these embodiments, a system in which a user can record and reproduce information has been described. However, the same applies to a read-only ROM disk as a recording medium. For example, the learning patterns for creating the attention cell determination table 29, the preceding cell reference table 33, and the continuation cell reference table 35 are combined with the ROM information used by the user, and phase pits (pits formed as irregularities on the disk substrate). ). By reading out this learning pattern in the same manner as described above, it is possible to prevent erroneous determination of the level even in a ROM-dedicated disk.
[0174]
In the embodiment described above, the learning area is provided in the innermost area where no user data is recorded. However, the learning area is provided at the head where user data is recorded in block units (for example, 256 cell units = 256 * 3 bits units). It does not matter. In this way, recording and reproduction can be performed while learning, so that the probability of erroneous determination can be further reduced.
[0175]
【The invention's effect】
According to the first aspect of the present invention, even when the output level of the reproduction signal of the attention cell is affected by the marks of the preceding cell and the continuation cell, the attention is made with reference to the multi-value determination values of these preceding cells and continuation cells. By determining the multilevel level of the cell, the possibility of erroneous determination of the multilevel level of the cell of interest can be reduced, and there is no inconvenience such as restrictions on the format of the recording medium.In addition, the influence of the preceding cell and continuation cell marks positioned before and after the target cell on the multilevel determination of the target cell is judged from the magnitude of intersymbol interference, and the multilevel determination value to be used is selectively switched. Since the multi-value level of the cell of interest is re-determined, the possibility of erroneous determination of the multi-value level of the cell of interest can be reduced even when the playback signal is affected by the marks of the cells located before and after, The reliability of multilevel determination can be improved.
[0180]
Claim2According to the described invention, the claims1In the information reproduction method described above, in order to determine the amount of intersymbol interference that the marks of the preceding cell and the continuation cell positioned before and after the cell of interest give to the multilevel determination of the cell of interest, the reproduction signal from each cell Since the determination is made in advance based on the determination result in which the multi-value determination is performed in, the multi-value determination value of the preceding cell, the continuation cell, or the multi-value determination value of the preceding cell and the continuation cell is referred to. It is possible to appropriately select whether or not to perform the multi-value determination.
[0181]
Claim3According to the described invention, the claims2In the described information reproduction method, the degree of intersymbol interference can be effectively used to appropriately select the multi-level determination value of the preceding cell, thereby improving the reliability of the multi-level determination. it can.
[0182]
Claim4According to the described invention, the claims2Or3In the described information reproduction method, the degree of intersymbol interference can be effectively used to appropriately select the multilevel determination value of the continuation cell, and the reliability of the multilevel determination can be improved. it can.
[0183]
Claim5According to the described invention, the claims2,3Or4In the described information reproduction method, it is possible to appropriately perform the determination process when the amount of intersymbol interference is substantially equal by effectively utilizing the degree of intersymbol interference.
[0184]
Claim6According to the described invention, the claims2Or5In the information reproduction method according to any one of the above, a signal value predicted from a learning result and an actual measurement value are calculated from the learning result, and the amount of intersymbol interference that the mark of the preceding cell or the continuation cell gives to the determination value of the mark of the cell of interest. Therefore, the result of pre-determining the magnitude of the intersymbol interference amount is ambiguous with the judgment value obtained by performing the multi-value judgment on the reproduction signal from the mark of each cell. However, the multi-value determination of the mark of the cell of interest can be performed appropriately, and the reliability of the multi-value determination can be improved.
[0185]
Claim7According to the described invention, the claims1In the described information reproducing method, the marks of the preceding cell and the continuation cell positioned before and after the target cell are predicted from the learning result in order to determine the amount of intersymbol interference given to the multilevel determination of the target cell. Since the comparison result of the error between the signal value and the actual measurement value is used, one of the multilevel determination values of the preceding cell, the continuation cell, or the multilevel determination values of the preceding cell and the continuation cell is referred to. Can be appropriately selected, and the reliability of multi-value determination can be improved.
[0186]
Claim8According to the described invention, the claims7In the described information reproduction method, it is possible to appropriately use the degree of intersymbol interference to appropriately determine the multilevel determination value of the cell of interest, and to improve the reliability of multilevel determination.
[0187]
Claim9According to the described invention, the claims7Or8In the described information reproduction method, it is possible to appropriately use the degree of intersymbol interference to appropriately determine the multilevel determination value of the cell of interest, and to improve the reliability of multilevel determination.
[0188]
Claim10According to the described invention, the claims1Or9In the information reproduction method according to any one of the above, since the re-determination result is fed back, the mark between the preceding cell and the continuation cell positioned before and after the target cell is given to the determination value of the target cell mark. The reliability of the result of determining the magnitude of the interference amount can be further improved, and the reliability of multilevel determination can be improved.
[0189]
Claim11According to the described invention, the claims1Or1015. The information reproducing method according to any one of claims 5 to 14, wherein multilevel determination of a target cell is performed using a reference table created based on level value information of a reproduction signal obtained from learning pattern information. This information reproduction method can be easily realized, and a reference table is created based on the result of statistical calculation processing by reproducing an information area in which learning pattern information of known multilevel levels is recorded in advance. It is possible to learn disturbance effects such as fluctuations in optical power and substrate distortion, and to create a reference table that can correct misjudgment due to the disturbance effects.
[0190]
Claim12According to the described invention, the claims11In the described information reproduction method, by using the attention cell determination table, the preceding cell reference table, and the continuation cell reference table as the reference table, it is possible to appropriately determine the amount of intersymbol interference, etc. As a result, the reliability of multi-value determination can be improved.
[0191]
Claim13According to the described invention, the claims12In the information reproducing method described, the reference table can be created with as little arithmetic processing as possible.
[0195]
Claim14According to the information reproducing apparatus of the described invention, the influence of the marks of the preceding cell and the continuation cell positioned before and after the cell of interest on the multilevel determination of the cell of interest is determined based on the magnitude of intersymbol interference and used. Even if the playback signal is affected by the marks of the cells located before and after the multi-level judgment value is selectively switched and the multi-level level of the target cell is redetermined by the redetermination calculation means, the multi-level level of the target cell The possibility of misjudgment is reduced, and the reliability of multi-valued judgment can be increased.
[0196]
Claim15According to the described invention, the claims14In the information reproducing apparatus described above, in order to determine the amount of intersymbol interference that the marks of the preceding cell and the continuation cell positioned before and after the cell of interest give to the multilevel determination of the cell of interest, the reproduction signal from each cell Since the determination is made in advance based on the determination result in which the multi-value determination is performed, any multi-value determination value among the multi-value determination values of the preceding cell, the continuation cell, or the preceding cell and the continuation cell is selected. It is possible to appropriately select whether to refer to, and to improve the reliability of multilevel determination.
[0197]
Claim16According to the described invention, the claims15In the described information reproducing apparatus, it is possible to appropriately utilize the degree of intersymbol interference to appropriately select the reference value for the multilevel determination value of the preceding cell, and to improve the reliability of the multilevel determination. it can.
[0198]
Claim17According to the described invention, the claims15Or16In the described information reproducing apparatus, the degree of intersymbol interference can be effectively used to appropriately select the multi-level determination value of the continuation cell, and the reliability of the multi-level determination can be improved. it can.
[0199]
Claim18According to the described invention, the claims15,16Or17In the described information reproducing apparatus, it is possible to appropriately perform the determination process when the amount of intersymbol interference is substantially equal by effectively utilizing the degree of intersymbol interference.
[0200]
Claim19According to the described invention, the claims15Or18In the information reproducing apparatus according to any one of the above, a signal value predicted from a learning result and an actual measurement value of the amount of intersymbol interference that the mark of the preceding cell or the continuation cell gives to the determination value of the mark of the cell of interest Therefore, even if the result of pre-determining the magnitude of the intersymbol interference amount is ambiguous with the judgment value obtained by performing multi-value judgment on the reproduction signal from the mark of each cell. The multi-value determination of the mark of the cell of interest can be performed appropriately, and the reliability of the multi-value determination can be improved.
[0201]
Claim20According to the described invention, the claims15In the described information reproducing apparatus, the marks of the preceding cell and the continuation cell positioned before and after the cell of interest are predicted from the learning result in order to determine the amount of intersymbol interference given to the multilevel determination of the cell of interest. Since the comparison result of the error between the signal value and the measured value is used, which multi-value judgment value is to be referred to among the multi-value judgment values of the preceding cell, the continuation cell, or the preceding cell and the continuation cell? Can be appropriately selected, and the reliability of multi-value determination can be improved.
[0202]
Claim21According to the described invention, the claims20In the described information reproducing apparatus, the multilevel determination value of the cell of interest can be appropriately determined by effectively utilizing the degree of intersymbol interference, and the reliability of multilevel determination can be improved.
[0203]
Claim22According to the described invention, the claims20Or21In the described information reproducing apparatus, the multilevel determination value of the cell of interest can be appropriately determined by effectively utilizing the degree of intersymbol interference, and the reliability of multilevel determination can be improved.
[0204]
Claim23According to the described invention, the claims15Or22In the information reproducing apparatus according to any one of the above, since the re-determination result is fed back, the mark between the preceding cell and the continuation cell positioned before and after the target cell is given to the determination value of the target cell mark. The reliability of the result of determining the magnitude of the interference amount can be further improved, and the reliability of multilevel determination can be improved.
[0205]
Claim24According to the described invention, the claims15Or23In the information reproducing apparatus according to any one of claims 1 to 4, by performing multilevel determination of a cell of interest using a reference table created based on level value information of a reproduced signal obtained from learning pattern information,15Or23The information reproducing apparatus described can be easily realized, and the reference table is created based on the result of reproducing the information area in which the learning pattern information of known multi-level levels is recorded in advance and performing statistical calculation processing. It is possible to learn disturbance effects such as fluctuations in reproduction light power and substrate distortion, and to create a reference table that can correct erroneous determination due to the disturbance effects.
[0206]
Claim25According to the described invention, the claims24In the described information reproducing apparatus, it is possible to appropriately determine the amount of intersymbol interference by creating an attention cell determination table, a preceding cell reference table, and a continuation cell reference table as reference tables. As a result, the reliability of multilevel determination can be improved.
[0207]
Claim26According to the described invention, the claims25In the described information reproducing apparatus, the reference table can be created with as little arithmetic processing as possible.
[0208]
Claim27According to the information recording / reproducing apparatus of the described invention, the multilevel determination of the cell of interest is performed by using the reference table created based on the level value information of the reproduction signal obtained from the learning pattern information.15Or23It is possible to easily realize an information recording / reproducing apparatus equivalent to the described information reproducing apparatus, and to record a reference table based on a result of actually recording learning pattern information of a known multi-value level in advance on a recording medium, reproducing it, and performing statistical calculation processing. Since it is created, it is possible to learn disturbance effects such as fluctuations in recording / reproducing light power and substrate distortion, and to create a reference table capable of correcting erroneous determination due to this disturbance effect.
[0209]
Claim28According to the described invention, the claims27In the described information recording / reproducing apparatus, it is possible to actually record learning pattern information of a known multi-value level in advance on a recording medium without disturbing the user data area, so that the creation of the reference table can be confused.
[0210]
Claim29According to the described invention, the claims27Or28In the described information recording / reproducing apparatus, the attention cell determination table, the preceding cell reference table, and the continuation cell reference table are created as the reference table, so that it is possible to appropriately determine the amount of intersymbol interference. As a result, the reliability of multilevel determination can be improved.
[0211]
Claim30According to the described invention, the claims29In the described information recording / reproducing apparatus, the reference table can be created with as little arithmetic processing as possible.
[Brief description of the drawings]
FIG. 1 is an explanatory diagram of a mark shape, a reproduction signal, etc., showing an example of an information recording / reproducing system for explaining the principle of an embodiment of the present invention.
FIG. 2 is a characteristic diagram showing a level change of a target cell due to the influence of preceding and succeeding cells.
FIG. 3 is a characteristic diagram showing a level change of a target cell according to a level of a preceding cell.
FIG. 4 is a characteristic diagram showing changes in the level of a target cell according to the level of a continuation cell.
FIG. 5 is a characteristic diagram showing a level change of a target cell according to the levels of a preceding cell and a continuation cell.
FIG. 6 is a characteristic diagram showing a state in the case of four values.
FIG. 7 is a block diagram showing an information recording / reproducing apparatus according to the first embodiment of the present invention.
FIG. 8 is a block diagram showing a configuration example of the multi-value level determination means.
FIG. 9 is an explanatory diagram showing an example of data in a table RAM.
FIG. 10 is a block diagram showing an information recording / reproducing apparatus according to a second embodiment of the present invention.
FIG. 11 is a block diagram illustrating a configuration example of the digital holding unit.
FIG. 12 is an explanatory diagram showing a target cell determination table and a method for creating the target cell determination table.
FIG. 13 is an explanatory diagram showing a cell-of-interest determination table for four values.
FIG. 14 is an explanatory diagram showing a preceding cell reference table and a creation method thereof.
FIG. 15 is an explanatory diagram showing a preceding cell reference table in the case of four values.
FIG. 16 is an explanatory diagram showing a continuation cell reference table and a creation method thereof.
FIG. 17 is an explanatory diagram showing a continuation cell reference table in the case of four values.
FIG. 18 is a schematic flowchart illustrating an example of determination processing.
FIG. 19 is a flowchart illustrating a processing example of only a target cell.
FIG. 20 is a schematic flowchart illustrating an example of determination processing according to the third embodiment of this invention.
FIG. 21 is an explanatory diagram showing a three-continuous recording mark learning table and a data extraction method thereof according to the fourth embodiment of the present invention.
FIG. 22 is a schematic flowchart illustrating an example of determination processing.
FIG. 23 is an explanatory diagram schematically showing a determination method according to the fifth embodiment of this invention.
FIG. 24 is an explanatory diagram showing the error totaling algorithm.
FIG. 25 is a schematic flowchart illustrating an example of determination processing.
FIG. 26 is an explanatory diagram schematically showing a modified example of the determination method.
FIG. 27 is an explanatory diagram showing the error totaling algorithm.
[Explanation of symbols]
1 cell
3 recording media
11 A / D conversion means
18 Calculation means
19 Multi-level data holding means
20 Criteria changing means
23 A / D conversion means
25 Digital data holding means
29 Attention cell judgment table
31 judgment value holding means
33 Preceding cell reference table
35 Continue cell reference table
S4 Primary calculation means
S6 Continue cell level calculation means
S7 to S13 Reference determination value selection means, redetermination calculation means

Claims (30)

An information reproducing method for reproducing the multi-level from a reproduction signal of a recording medium in which a mark representing multi-level information is recorded in a cell,
The preceding cell preceding the target cell of interest, the continuing cell to continue, or the multi-value determination value of the preceding cell and the continuing cell are referred to determine the multi-value level of the cell of interest ,
Further, according to the magnitude of the amount of intersymbol interference from the preceding cell with respect to the reproduction signal of the target cell or the magnitude of the amount of intersymbol interference from the continuation cell with respect to the reproduction signal of the attention cell The multi-value determination value is selected from among the multi-value determination values of the continuation cell or the preceding cell and the continuation cell, and the multi-value level of the target cell is determined with reference to the multi-value determination value. An information reproduction method for re-determination . Information reproducing method according to claim 1, wherein the determination as to the magnitude of the amount of intersymbol interference on the basis of multi-value determination result by the level of the reproduction signal from each cell. When the amount of intersymbol interference from the preceding cell with respect to the reproduction signal of the target cell is larger than the amount of intersymbol interference from the continuation cell with respect to the reproduction signal of the target cell, 3. The information reproducing method according to claim 2, wherein a multi-value level of the target cell is re-determined with reference to a multi-value determination value. When the amount of intersymbol interference from the preceding cell with respect to the reproduction signal of the target cell is smaller than the amount of intersymbol interference from the continuation cell with respect to the reproduction signal of the target cell, claim 2 or 3 information reproducing method according to the to redetermine the multivalue level of the target cell with reference to the multi-level determination value. The amount of intersymbol interference from the preceding cell with respect to the reproduction signal of the attention cell is substantially the same as the amount of intersymbol interference from the continuation cell with respect to the reproduction signal of the attention cell, and the multilevel determination of the preceding cell The multi-value determination value of the target cell re-determined with reference to the value is Y1, and the multi-value determination value of the target cell re-determined with reference to the multi-value determination value of the continuation cell is Y2. The information reproducing method according to claim 2 , 3 or 4 , wherein when the value determination values Y1 and Y2 match, the multi-value determination value of the target cell is determined to be Y1 or Y2. The amount of intersymbol interference from the preceding cell with respect to the reproduction signal of the attention cell is substantially the same as the amount of intersymbol interference from the continuation cell with respect to the reproduction signal of the attention cell, and the multilevel determination of the preceding cell The multi-value determination value of the target cell re-determined with reference to the value is Y1, and the multi-value determination value of the target cell re-determined with reference to the multi-value determination value of the continuation cell is Y2. When the value determination values Y1 and Y2 do not match, the learned level value of the reproduction signal predicted from the multi-value determination value Y1 of the re-determined target cell is y1, and the re-determined target cell of the re-determined target cell When y2 is the learned level value of the reproduction signal predicted from the multi-level determination value Y2, and y is the level value of the reproduction signal actually measured in the cell of interest, | y−y1 | and | y−y2 | Compare the size relationship, and The determination value | y-y1 | <| y-y2 | a long if multilevel decision value Y1, | y-y1 |> | claim which is adapted to determine if the multi-level decision value Y2 | y-y2 The information reproduction method according to any one of 2 to 5 . After performing multilevel determination at the level of the playback signal from each cell, a playback signal predicted from the multilevel determination value of the target cell re-determined with reference to the multilevel determination value of the preceding cell has been learned. The level value, the learned level value of the reproduction signal predicted from the multi-value determination value of the attention cell re-determined with reference to the multi-value determination value of the continuation cell, and the actual reproduction signal of the attention cell by comparison with the level value of the information reproducing method according to claim 1, wherein the determination as to the magnitude of the amount of intersymbol interference. The learned level value of the reproduction signal predicted from the multi-value determination value Y1 of the re-determined cell of interest is y1, and the reproduction signal of learning is predicted from the multi-value determination value Y2 of the re-determined cell of interest. Assuming that the finished level value is y2 and the level value of the measured reproduction signal of the target cell is y, the result of comparison of the magnitude relationship between | y−y1 | and | y−y2 | 8. The information reproducing method according to claim 7, wherein if | y−y2 |, the multilevel determination value of the cell of interest is determined as a multilevel determination value Y1. The learned level value of the reproduction signal predicted from the multi-value determination value Y1 of the re-determined cell of interest is y1, and the reproduction signal of learning is predicted from the multi-value determination value Y2 of the re-determined cell of interest. Assuming that the finished level value is y2 and the level value of the measured reproduction signal of the target cell is y, the result of comparison of the magnitude relationship between | y−y1 | and | y−y2 | | y-y2 | a long if claim 7 or 8 information reproducing method according to to determine the multi-level decision value to multivalued determination value Y2 of the target cell. When said multilevel decision value of the target cell is rewritten by the re-determination, the information of any one described in claims 1 as adapted to use the rewritten multilevel decision value multilevel decision of continuing cells after 9 Playback method. The learning pattern information is reproduced from a recording medium in which all the combination patterns having three marks as one group are recorded in advance as learning pattern information of known multilevel levels, and a reproduction signal obtained from these learning pattern information level value to create the reference table based on the information, the information reproducing method of any one of claims 1 to 10 as adapted to use the reference table in the multi-level determination of the target cell. 12. The information reproducing method according to claim 11 , wherein the reference table includes an attention cell determination table, a preceding cell reference table, and a continuation cell reference table. The target cell determination table is created by a calculation process ignoring the level values of the preceding cell and the continuation cell in the level value information of the reproduction signal obtained from the learning pattern information, and the preceding cell reference table is the learning pattern. Among the level value information of the reproduction signal obtained from the information, the continuation cell reference table is created by ignoring the level value of the continuation cell. 13. The information reproducing method according to claim 12 , wherein the information reproducing method is created by arithmetic processing ignoring the level value of the preceding cell.   An information reproducing apparatus for reproducing the multi-level from a reproduction signal of a recording medium in which marks representing multi-level information are recorded in cells, and an A / D conversion means for converting the reproduction signal into digital data; Digital data holding means for holding the converted digital data, primary calculation means for determining based on a reproduction signal of only a mark of a target cell regarding the multi-value level of the converted digital data, and this primary calculation Determination value holding means for temporarily holding a multi-value determination value by the means; continuation cell level calculation means for determining a multi-value level of a continuation cell continuing to the attention cell prior to calculation of the attention cell; The amount of intersymbol interference from the preceding cell for the reproduction signal, or the amount of intersymbol interference from the continuation cell for the reproduction signal of the target cell In response, the reference determination value selection means for selecting any one of the multi-value determination values of the preceding cell, the continuation cell, or the multi-value determination values of the preceding cell and the continuation cell, and this reference determination value selection An information reproducing apparatus comprising: a re-determination calculation unit that re-determines the multi-level level of the target cell with reference to the multi-level determination value selected by the unit. 15. The information reproduction apparatus according to claim 14, wherein the reference determination value selection unit determines the magnitude of the intersymbol interference amount based on a multilevel determination result based on a level of a reproduction signal from each cell. The reference determination value selection unit is configured such that the amount of intersymbol interference from the preceding cell with respect to the reproduction signal of the target cell is larger than the amount of intersymbol interference from the continuation cell with respect to the reproduction signal of the target cell. The information reproducing apparatus according to claim 15 , wherein, in the case, the multi-value determination value of the preceding cell is selected. The reference determination value selection unit is configured such that the amount of intersymbol interference from the preceding cell with respect to the reproduction signal of the target cell is smaller than the amount of intersymbol interference from the continuation cell with respect to the reproduction signal of the target cell. The information reproducing apparatus according to claim 15 or 16 , wherein in the case, a multi-value determination value of the continuation cell is selected. The reference determination value selection unit is configured such that the amount of intersymbol interference from the preceding cell with respect to the reproduction signal of the target cell is substantially the same as the amount of intersymbol interference from the continuation cell with respect to the reproduction signal of the target cell. The multi-value determination value of the preceding cell and the multi-value determination value of the continuation cell are selected, and the re-determination calculation means re-determines with reference to the multi-value determination value of the preceding cell. When the multi-value determination value of the cell is Y1, the multi-value determination value of the target cell re-determined with reference to the multi-value determination value of the continuing cell is Y2, and these multi-value determination values Y1 and Y2 match. The information reproducing apparatus according to claim 15 , 16 or 17 , wherein the multilevel determination value of the target cell is determined as Y1 or Y2. The reference determination value selection unit is configured such that the amount of intersymbol interference from the preceding cell with respect to the reproduction signal of the target cell is substantially the same as the amount of intersymbol interference from the continuation cell with respect to the reproduction signal of the target cell. The multi-value determination value of the preceding cell and the multi-value determination value of the continuation cell are selected, and the re-determination calculation means re-determines the attention that has been re-determined with reference to the multi-value determination value of the preceding cell When the multi-value determination value of the cell is Y1, the multi-value determination value of the target cell re-determined with reference to the multi-value determination value of the continuation cell is Y2, and these multi-value determination values Y1 and Y2 do not match Is the learned level value of the reproduction signal predicted from the multi-value determination value Y1 of the re-determined target cell, y1, and the reproduction signal predicted from the multi-value determination value Y2 of the re-determined cell of interest The learned level value of y2 When the level value of the actually measured reproduction signal of the eye cell is y, the magnitude relationship between | y−y1 | and | y−y2 | is compared, and the multilevel determination value of the target cell is determined as | y−y1 | The information reproduction according to any one of claims 15 to 18 , wherein <| y-y2 | is determined as a multi-value determination value Y1, and | y-y1 |> | y-y2 | is determined as a multi-value determination value Y2. apparatus. The reference determination value selection means performs a multilevel determination at the level of a reproduction signal from each cell, and then determines the multilevel determination value of the target cell determined again with reference to the multilevel determination value of the preceding cell. A learned level value of the reproduced signal to be predicted, a learned level value of the reproduced signal predicted from the multivalued determination value of the target cell re-determined with reference to the multivalued determination value of the continuation cell, and The information reproducing apparatus according to claim 15 , wherein the magnitude of the intersymbol interference amount is determined by comparison with a level value of an actually reproduced signal of a target cell. The re-determination calculating means uses y1 as the learned level value of the reproduction signal predicted from the multi-value determination value Y1 of the re-determined target cell, and from the multi-value determination value Y2 of the re-determined target cell. The result of comparison of the magnitude relationship between | y−y1 | and | y−y2 |, where y2 is the learned level value of the predicted reproduction signal and y is the level value of the actually measured reproduction signal of the target cell. 21. The information reproducing apparatus according to claim 20, wherein if | y−y1 | <| y−y2 |, the multilevel determination value of the target cell is determined as a multilevel determination value Y1. The re-determination calculating means uses y1 as the learned level value of the reproduction signal predicted from the multi-value determination value Y1 of the re-determined target cell, and from the multi-value determination value Y2 of the re-determined target cell. The result of comparison of the magnitude relationship between | y−y1 | and | y−y2 |, where y2 is the learned level value of the predicted reproduction signal and y is the level value of the actually measured reproduction signal of the target cell. 22. The information reproducing apparatus according to claim 20, wherein if | y−y1 |> | y−y2 |, the multilevel determination value of the target cell is determined as a multilevel determination value Y2. The redetermination calculating means, when said multilevel decision value of the target cell is rewritten by the re-determination, any one of claims 15 to 22 using the rewritten multilevel decision value multilevel decision of continuing cells after An information reproducing apparatus according to one. Learning means for reproducing the learning pattern information from a recording medium in which all combination patterns having three marks as one group are recorded in advance as learning pattern information of known multilevel levels, and learning by the learning means comprising a reference table creation means for creating a reference table based on the level value information of the reproduction signal obtained from the pattern information, and any claims 15 to 23 as adapted to use the reference table in the multi-level determination of the target cell An information reproducing apparatus according to claim 1. 25. The information reproducing apparatus according to claim 24 , wherein the reference table includes an attention cell determination table, a preceding cell reference table, and a continuation cell reference table. The reference table creating means creates the attention cell determination table by calculation processing ignoring the level values of the preceding cell and the continuation cell among the level value information of the reproduction signal obtained from the learning pattern information, and the learning pattern information The preceding cell reference table is created by an arithmetic process ignoring the level value of the continuation cell among the level value information of the reproduction signal obtained from 26. The information reproducing apparatus according to claim 25 , wherein the continuation cell reference table is created by an arithmetic process ignoring the level value of.   An information recording / reproducing apparatus for reproducing the multi-level from a reproduction signal of a writable recording medium in which marks representing multi-level information are recorded in a cell, wherein the A / D converts the reproduction signal into digital data A conversion means; a digital data holding means for holding the converted digital data; and a primary calculation means for judging based on a reproduction signal of only a mark of a target cell as a target with respect to the multi-value level of the converted digital data; A determination value holding means for temporarily holding a multi-value determination value by the primary calculation means; a continuation cell level calculation means for determining a multi-value level of a continuation cell continuing to the attention cell prior to the calculation of the attention cell; The amount of intersymbol interference from the preceding cell for the playback signal of the target cell, or the code from the continuation cell for the playback signal of the target cell Reference determination value selection means for selecting any one of the multilevel determination values of the preceding cell, the continuation cell, or the multilevel determination values of the preceding cell and the continuation cell according to the magnitude of the interference amount; , Re-determination calculation means for re-determining the multi-value level of the cell of interest with reference to the multi-value determination value selected by the reference determination value selection means, and all combination patterns having three marks as one group. Learning pattern recording means for recording on the recording medium as learning pattern information of known multilevel levels, learning means for reproducing the learning pattern information from the recording medium, and reproduction obtained from these learning pattern information by the learning means An information recording / reproducing apparatus comprising: reference table creating means for creating a reference table used for multi-value determination of the cell of interest based on signal level value information. 28. The information recording / reproducing apparatus according to claim 27 , wherein the learning pattern recording means records all combination patterns outside the user data area of the recording medium. Examples criteria table, the subject cell and the determination table, a preceding cell reference table, the information recording and reproducing apparatus according to claim 27 or 28, wherein and a continuing cell reference table. The reference table creating means creates the attention cell determination table by calculation processing ignoring the level values of the preceding cell and the continuation cell among the level value information of the reproduction signal obtained from the learning pattern information, and the learning pattern information The preceding cell reference table is created by the calculation process ignoring the level value of the continuation cell among the level value information of the reproduction signal obtained from the above, and the preceding cell among the level value information of the reproduction signal obtained from the learning pattern information 30. The information recording / reproducing apparatus according to claim 29 , wherein the continuation cell reference table is created by an arithmetic process ignoring the level value of the information.

Images