JP3784361B2 - Auditory filter shape estimation method and apparatus - Google Patents

Description

【００３４】
ここで、ｐはフィルタのバンド幅（傾斜角度）を表す係数、ｒはフィルタのダイナミックレンジを表す係数、ｇは聴覚フィルタの中心周波数からの距離を正規化した値でノッチ幅（＝Δｆ／ｆ_C）である。聴覚フィルタを式（１）でモデル化するためには、ノッチノイズマスキングによって得られるノッチ雑音マスキングデータＰｓ(ｇ)が用いられる。これは、次に示す式（２）で定義される。
【００３５】
【数２】

【００３６】
ここで、ｇ_maxはノッチ幅ｇの上限値、Ｋは各個人の感度、ｆ_Cはノッチノイズの中心周波数、Ｎ_Xは内耳に到達した時のマスカーＭのレベルである。
【００３７】
聴覚フィルタ表示部１０は、聴覚フィルタ算出部９によって得られたフィルタ係数ｐ，ｒから聴覚フィルタの形状を表示する。
ノッチ幅表示部１１は、ノッチ幅変更部５が出力するマスカーのノッチ幅ｇを表示する。
【００３８】
応答部１２は、被験者の操作によって検査音提示部７により提示される検査音から信号音Ｓ’を知覚できた場合の応答信号と知覚できない場合の応答信号を出力する。そして、信号音Ｓ’を知覚できない場合の応答信号が、ノッチ幅変更部５へ出力される。
【００３９】
なお、信号音Ｓ’を知覚できた場合の応答信号のみを使用し、所定時間知覚できた場合の応答信号が出力されない時には、知覚できなかったものと判断し、次のステップに進むようにしてもよい。
【００４０】
以上のように構成した聴覚フィルタの形状推定装置の動作について説明する。信号音Ｓ’に、内耳に到達した時に１Ｈｚ当たりの音圧レベルが一様になるように個々の被験者の外耳と中耳の周波数特性を考慮したマスカーＭをノッチノイズ重畳部６において重畳し、検査音を作成する。作成された検査音は、検査音提示部７から被験者に提示される。
被験者は、検査音を聴取し、信号音Ｓ’を知覚できたか否かを応答部１２により応答する。
【００４１】
ここで、図２に示すように、被験者に対して最初の検査音を提示する際には、検査音提示の開始時に所定時間ｔだけ信号音レベル変更部２が出力する信号音Ｓ’のレベルを０にする（信号音Ｓ’の提示をマスカーＭよりも所定時間ｔだけ遅らせる）ことができる。
【００４２】
ノッチ幅変更部５では、その応答に応じて新たなマスカーＭを作成し、再びノッチノイズ重畳部６に出力する。ここで、マスカーＭのノッチ幅ｇの変更は、被験者の応答に応じて自動的に行うようにしてもよいし（ＣＰＵ等で構成し、専用プログラムを用意する）、測定者がその都度手動で指示してもよい。
【００４３】
ノッチ幅ｇを増加させた後に、被験者に対して新たな検査音を提示する際にも、新たな検査音提示の開始時に所定時間ｔだけ信号音レベル変更部２が出力する信号音Ｓ’のレベルを０にする（信号音Ｓ’の提示をマスカーＭよりも所定時間ｔだけ遅らせる）ことができる。
【００４４】
測定中のノッチ幅ｇは、ノッチ幅表示部１１に表示される。そして、ノッチ幅変更部５は、ノッチ幅ｇを徐々に大きくしていき、被験者の応答により信号音Ｓ’が知覚できる最小ノッチ幅（限界ノッチ幅）ｇ_X-aを測定する。最小ノッチ幅ｇ_X-aは、図３に示すように、ノッチ雑音マスキングデータ特性において、頂点からａ［ｄＢ］減衰した点のバンド幅と等価である。
【００４５】
上限値算出部８では、ノッチ幅変更部５が測定した最小ノッチ幅ｇ_X-aを用いて上限値ｇ_maxを算出する。
聴覚フィルタ算出部９では、ノッチ幅変更部５が測定した最小ノッチ幅ｇ_X-aと上限値算出部８が算出した上限値ｇ_maxを用いてroex（ｐ，ｒ）フィルタのフィルタ係数ｐを、次のようにして算出する。ここでは、式（２）においてＰｓ(０)とＰｓ(ｇ_X-a)の差分ａを、次に示す式（３）とする。
【００４６】
【数３】

【００４７】
更に、式（２）、（３）より差分ａは、次に示す式（４）となる。
【００４８】
【数４】

【００４９】
ここで、ｇ_X-aの値は測定値であり、ｒはダイナミックレンジを定める係数ｘを用いて１０ｌｏｇ₁₀ｒ＝-ｘとし、ａは測定条件によって決まる任意の定数であるので、ｇ_maxの値が決まれば、フィルタ係数ｐの値を式（４）から算出することができる。
【００５０】
また、ｇ_maxの値の決定法としては、以下のような方法が考えられる。信号検出閾値と最小可聴閾値が一致するノッチ幅ｇの値（以下ｇ_tと記す）は、健聴者で１Ｈｚ当たりの音圧レベルが４０ｄＢＳＰＬになるように作成されたマスカーを使用して測定を行った場合、概ねｇ_t＝０．６〜０．７であり、ｇ_max＞０．８の区間の積分はノッチ雑音マスキングデータＰｓ(ｇ)の値に殆ど影響を与えない。
【００５１】
従って、聴覚フィルタの過去の測定データを参照すると、ｇ_maxの値とｇ_tの値の関係は、次に示す式（５）となる。
【００５２】
【数５】

【００５３】
ここで、Ｃの値は、概ね１．１〜１．３となっている。一方、本発明の簡易測定法では、測定値がノッチ幅ｇ_X-aのみであるため、信号検出閾値と最小可聴閾値が一致するノッチ幅ｇ_tの値は未知である。加えて、信号検出閾値と最小可聴閾値が一致するノッチ幅ｇ_tの値は、測定結果（ノッチ幅ｇと信号検出閾値［ｄＢＳＰＬ］との関係）からダイナミックレンジ（値ｘに相当）によっても変化することが分かっている。よって、ｇ_maxの値はノッチ幅ｇ_X-aからｘの値ごとに推定する必要があり、ｇ_maxの値とｇ_X-aの値の関係は、次に示す式（６）となる。
【００５４】
【数６】

【００５５】
ここで、Ａ_xはｘの値に応じて変化する係数である。このＡ_xの値は、式（５）でＣ＝１．２と仮定し、２名の健聴被験者の測定結果（ｇ_tの値）を用いてｇ_maxの値を算出し、このｇ_maxの値とｇ_X-aの値から実験的に決定すると、次に示す表１のようになる。
【００５６】
【表１】

【００５７】
なお、表１では、ｘ＝１０,１５,２０,３０［ｄＢ］におけるｇ_tの値を測定し、Ａ_xの値を算出しているが、例えばｘ＝４０,５０［ｄＢ］におけるＡ_xの値についても予め用意しておくことができる。
【００５８】
従って、聴覚フィルタ算出部９では、上限値ｇ_maxと最小ノッチ幅ｇ_X-aと係数ｒに相当する値ｘと値ａを式（４）に代入してroex（ｐ，ｒ）フィルタのフィルタ係数ｐを算出する。更に、フィルタ係数ｐ，ｒから聴覚フィルタの形状を推定する。
そして、聴覚フィルタ表示部１０では、フィルタ係数ｐとフィルタ係数ｒに相当する値ｘからフィルタ形状を表示する。
【００５９】
このようにして、フィルタ係数ｒに相当する値ｘを、例えば１０，２０，３０，４０，５０［ｄＢ］として、それぞれについてフィルタ係数ｐの推定値を求め、図５に示すように、値ｘをパラメータとした聴覚フィルタの形状（横軸がノッチ幅ｇ、縦軸が減衰量［ｄＢ］）を表示する。
【００６０】
次に、本発明に係る聴覚フィルタの形状推定方法について、図４に示すフローチャートより説明する。
先ず、ステップＳＰ１で、ノイズを付加しない状態で周波数ｆを含む気導、骨導のオージオグラムを測定する。ここで、周波数ｆにおける純音最小可聴閾値をＴ［ｄＢＳＰＬ／Ｈｚ］とする。ステップＳＰ２で、測定する周波数（信号音周波数）ｆを設定する。
【００６１】
次いで、ステップＳＰ３で、ノッチ雑音マスキングデータのダイナミックレンジを定める係数ｘの値を設定する。ここでは、ｘ＝１０としている。更に、ステップＳＰ４において、周波数ｆ［Ｈｚ］、レベルＴ＋ｘ［ｄＢＳＰＬ］の純音（信号音）Ｓを作成する。
【００６２】
次いで、ステップＳＰ５で、信号音Ｓにレベルが十分小さく内耳に到達した時、１Ｈｚ当たりの音圧レベルが一様になるよう補正されたノイズやステップＳＰ１で測定した気骨導のオージオグラムの差から考慮したノイズを重畳し、このノイズのレベルを徐々に上昇させながら、被験者が信号音Ｓを知覚できなくなる最小のレベル（限界マスキングレベル）Ｎ_X’を測定する。
【００６３】
ステップＳＰ６では、レベルＮ_X’のノイズのノッチ幅ｇを所定の値に設定する。ここでは、ｇ＝０．０５としている。また、ステップＳＰ７では、信号音ＳのレベルＴ＋ｘ［ｄＢＳＰＬ］から減じる値ａ［ｄＢ］を所定の値（ｘ＞ａ）に設定する。ここでは、ａ＝５としている。
【００６４】
次いで、ステップＳＰ８で、周波数ｆ［Ｈｚ］、レベルＴ＋ｘ−ａ［ｄＢＳＰＬ］の純音（信号音）Ｓ’を作成する。
更に、ステップＳＰ９で、ノッチの中心周波数ｆ_C（＝ｆ）、ノッチ幅ｇ、レベルＮ_X’のマスカーＭを作成する。
【００６５】
次いで、ステップＳＰ１０で、信号音Ｓ’にマスカーＭを重畳し、検査音として被験者に提示する。更に、ステップＳＰ１１で、被験者に検査音から信号音Ｓ’を知覚できたか否か判断させ、信号音Ｓ’を知覚できたという応答がない場合には、ステップＳＰ１２へ進みノッチ幅ｇの値を徐々に増加させ（ここでは、増加幅を０．０５としている）、ステップＳＰ９〜ステップＳＰ１１を信号音Ｓ’を知覚できたという応答があるまで繰り返す。
【００６６】
ここで、図２に示すように、被験者に対して最初に検査音を提示する際には、検査音提示の開始時に所定時間ｔだけ信号音レベル変更部２が出力する信号音Ｓ’のレベルを０にする（信号音Ｓ’の提示をマスカーＭよりも所定時間ｔだけ遅らせる）ことができる。同様に、ステップＳＰ１２でノッチ幅ｇを増加させた後に、被験者に対して新たな検査音を提示する際にも、新たな検査音提示の開始時に所定時間ｔだけ信号音レベル変更部２が出力する信号音Ｓ’のレベルを０にすることができる。
【００６７】
そして、信号音Ｓ’の提示開始のタイミングと応答のタイミングが著しく異なる場合には、再度同条件で測定を行い、安定した応答が得られるまで測定を繰り返してもよい。また、所定時間ｔは一定値である必要はなく、測定回毎にランダムに変更してもよい。
【００６８】
このように、最初の検査音提示の際に信号音Ｓ’をマスカーＭよりも所定時間ｔだけ遅らせて提示したり、マスカーＭのノッチ幅ｇの値を変化させた時に信号音Ｓ’のみを一旦停止し、所定時間ｔが経過した後に信号音Ｓ’を被験者に提示したりして、その時の被験者の応答のタイミングを観察すれば、被験者の応答の正当性を判断することも可能であると考えられるからである。
【００６９】
図２に示すように、信号音Ｓ’が被験者に提示された後に、信号音Ｓ’を知覚したとの応答が有り、信号音Ｓ’の停止と同時に知覚したとの応答が無くなれば、被験者の応答は正常であると判断できる。一方、信号音Ｓ’が被験者に提示される前に、信号音Ｓ’を知覚したとの応答が有り、しかも信号音Ｓ’の停止後も知覚したとの応答が維持される場合などは、被験者の応答は正常でないと判断できるので、再測定を行うか、または信号音Ｓ’のレベル及びマスカーＭのレベルの設定を変更すればよい。
【００７０】
ステップＳＰ１１で、信号音Ｓ’を知覚できたという応答がされると、ステップＳＰ１３へ進む。
ステップＳＰ１３では、その時のノッチ幅ｇを、その被験者の最小ノッチ幅ｇ_X-aと定める。更に、設定した値ｘ（＝１０［ｄＢ］）におけるＡ_xの値を、表１を用いて決定する（Ａ_x＝３．１５）。
【００７１】
次いで、ステップＳＰ１４で、式（６）にＡ_xの値（３．１５）及び測定した最小ノッチ幅ｇ_X-aの値を代入して被験者に適したノッチ幅ｇの上限値ｇ_maxの値を決定する。
【００７２】
ステップＳＰ１５では、式（４）に測定した最小ノッチ幅ｇ_X-aと、この最小ノッチ幅ｇ_X-aから推定したノッチ幅の上限値ｇ_maxと、予め設定した係数ｒに相当する値ｘ及び値ａを代入して、roex（ｐ，ｒ）フィルタのフィルタ係数ｐ_X（ｘ＝１０）の推定値を算出する。
【００７３】
次いで、ステップＳＰ１６で、ｘ＝５０であるか否かを判断し、ｘ＝５０でなければ、ステップＳＰ１７に進んでｘ＝ｘ＋１０とし、ｘ＝５０になるまでステップＳＰ４〜ステップＳＰ１７を繰り返し、新たなｘ（ｘ＝２０，３０，４０，５０［ｄＢ］）についてのフィルタ係数ｐ_Xの推定値を式（４）より算出する。ここでは、ｘの増加幅を１０とし、ｘ＝５０までの測定を行っている。
【００７４】
ステップＳＰ１６で、ｘ＝５０になったと判断されると、ステップＳＰ１８において、図５に示すように、値ｘをパラメータ（ｘ＝１０，２０，３０，４０，５０［ｄＢ］）とした聴覚フィルタの形状（横軸がノッチ幅ｇ、縦軸が減衰量［ｄＢ］）を表示する。
【００７５】
なお、本発明の実施の形態では、図６（ａ）に示すように、マスカーＭを被験者の外耳と中耳の周波数特性を考慮したノイズにノッチを付加する方法で作成しているが、このマスカーＭは、図６（ｂ）に示すように、２つのバンドノイズ（高周波数側と低周波数側）で構成してもよい。
【００７６】
また、上述した本発明の実施の形態においては、信号音レベル変更部２が出力する信号音信号の周波数ｆと、ノッチ幅変更部５が出力するマスカーＭの中心周波数ｆ_Cを一致させた（ｆ＝ｆ_C）場合について説明したが、必ずしも、一致させる必要はなく、ノッチ幅内に信号音信号が含まれていればよい。
【００７７】
更に、上述した本発明の実施の形態においては、被験者が知覚できる最小ノッチ幅（限界ノッチ幅）ｇ_X-aを求めるのに際し、はじめにマスカーＭのノッチ幅ｇを狭く設定（ｇ＝０．０５）して検査音を提示し、被験者が検査音から信号音Ｓ’を知覚できない場合に、徐々にマスカーＭのノッチ幅ｇを広くして信号音Ｓ’を知覚できるまで再度検査音を提示する手順を説明した。
【００７８】
しかし、被験者が知覚できる最小ノッチ幅ｇ_X-aが求まればよいのであるから、はじめにマスカーＭのノッチ幅ｇを広く設定して検査音を提示し、被験者が検査音から信号音Ｓ’を知覚できる状態にし、徐々にマスカーＭのノッチ幅ｇを狭くして信号音Ｓ’を知覚できなくなるまで再度検査音を提示する手順としてもよい。
【００７９】
また、上述した本発明の実施の形態においては、信号音のレベルや限界マスキングレベルＮ_X’の設定に際し、いわゆる極限法を用いた場合について説明したが、信号音を加算した検査音と信号音を加算していない検査音のいずれかから選択させる二者強制選択法（Two alternative forced choice）等の心理物理測定法に基づいて設定してもよい。
【００８０】
更に、上述した本発明の実施の形態においては、被験者が知覚できる最小ノッチ幅ｇ_X-aを求めるに際し、極限法を用いた場合について説明したが、二者強制選択法等の心理物理測定法に基づいて求めてもよい。
【００８１】
また、上述した本発明の実施の形態においては、聴覚フィルタのモデルをroex（ｐ，ｒ）フィルタとしてフィルタ形状の推定を行ったが、このモデル関数は必ずしもroex（ｐ，ｒ）フィルタである必要はなく、聴覚フィルタのモデルとして適正であれば、他の関数（例えば、roex（ｐ，ｒ）フィルタにおいて、１０ｌｏｇ₁₀ｒ＝０としたroex（ｐ）など）であっても差し支えない。
【００８２】
更に、上述した本発明の実施の形態においては、最小ノッチ幅ｇ_X-a測定時のステップＳＰ１０における検査音提示の開始時に、信号音Ｓ’の提示をマスカーＭよりも所定時間ｔだけ遅らせて被験者の応答の正当性を判断しているが、ステップＳＰ５で限界マスキングレベルＮ_X’測定時にマスカーのレベルを変更する際にも、同様の時間遅れを信号音Ｓに与えて、限界マスキングレベルＮ_X測定における正当性も併せて判断してもよい。
【００８３】
【発明の効果】
以上説明したように請求項１に係る発明によれば、被験者の外耳と中耳の周波数特性を考慮した周波数ｆをノッチに含む、ノッチ幅ｇ、レベルＮ_X’のマスカーを生成することにより、難聴の診断、難聴者の聴覚特性の把握、補聴器フィッティング等を効率的かつ正確に行うために有効な聴覚フィルタの形状を正確に推定することができる。
【００８４】
請求項２に係る発明によれば、入力信号レベルに対応した聴覚フィルタの形状を推定することができ、入力信号レベルと聴覚フィルタとの関係を知ることができる。
【００８５】
請求項３に係る発明によれば、周波数ｆにおける被験者の最小可聴閾値Ｔに任意の値ｘを加算した信号音Ｓに基づいて限界マスキングレベルＮ_X’を決定することによって、より正確に被験者の外耳と中耳における周波数特性を考慮した聴覚フィルタの形状を推定することができる。
【００８６】
請求項４に係る発明によれば、最小ノッチ幅ｇ_X-aから被験者に適したノッチ幅ｇの上限値ｇ_maxを推定することによって、より正確に被験者の外耳と中耳における周波数特性を考慮した聴覚フィルタの形状を推定することができる。
【００８７】
請求項５に係る発明によれば、信号音提示のタイミングと被験者の応答のタイミングから被験者の応答の正当性を容易に判断することができる。
【００８８】
請求項６に係る発明によれば、被験者の外耳と中耳の周波数特性を考慮したノイズを生成するノイズ生成部を備えることにより、難聴の診断、難聴者の聴覚特性の把握、補聴器フィッティング等を効率的かつ正確に行うために有効な聴覚フィルタの形状を正確に推定することができる。
【００８９】
請求項７に係る発明によれば、信号音生成部で生成された信号音を所定レベルに増幅・減衰する信号音レベル変更部を備えることによって、より正確に聴覚フィルタの形状を推定することができる。
【００９０】
請求項８に係る発明によれば、被験者が検査音を知覚できた時のノッチ幅に基づいて被験者に適したノッチ幅の上限値を算出する上限値算出部を備えることによって、より正確に聴覚フィルタの形状を推定することができる。
【００９１】
請求項９に係る発明によれば、信号音提示のタイミングと被験者の応答のタイミングから被験者の応答の正当性を容易に判断することができる。
【図面の簡単な説明】
【図１】本発明に係る聴覚フィルタの形状推定装置の構成図
【図２】検査音の提示と被験者の応答の関係を示すタイミングチャート
【図３】本発明に係る聴覚フィルタの形状推定方法の説明図
【図４】本発明に係る聴覚フィルタの形状推定方法の手順を示すフローチャート
【図５】本発明で求めた聴覚フィルタの形状を示す図
【図６】マスカーの構成図
【符号の説明】
１…信号音生成部、２…信号音レベル変更部、３…ノイズ生成部、４…ノイズレベル変更部、５…ノッチ幅変更部、６…ノッチノイズ重畳部、７…検査音提示部、８…上限値算出部、９…聴覚フィルタ算出部、１０…聴覚フィルタ表示部、１１…ノッチ幅表示部、１２…応答部。[0001]
BACKGROUND OF THE INVENTION
The present invention relates to a method and apparatus for estimating the shape of a human auditory filter.
[0002]
[Prior art]
At present, the hearing characteristics of hearing-impaired people that are most frequently used are hearing tests (audiogram measurement) and speech intelligibility tests. In the hearing test, the frequency characteristics of the minimum hearing threshold of the deaf person can be known, and in the speech intelligibility test, the hearing ability of the deaf person can be known.
However, since the hearing characteristics of the hearing-impaired person vary from person to person, it is considered that only these two methods can grasp only one part of the complicated hearing characteristics.
[0003]
In general, it is said that a hearing-impaired person has a reduced frequency resolution in addition to a decrease in hearing ability and a ability to hear words. Here, the frequency resolution is the ability to distinguish two sounds having different frequencies. For example, a normal hearing person can distinguish between two sounds whose frequencies are close to each other such as 1 kHz and 1.2 kHz, but a hearing-impaired person whose frequency resolution is low cannot distinguish between these two sounds.
[0004]
If the degree of decrease in the frequency resolution increases, the ability to hear words decreases, and the ability to discriminate speech under noise decreases. Knowing the degree of the reduction in frequency resolution is very useful in diagnosis of hearing loss, grasping of hearing characteristics of a hearing impaired person, hearing aid fitting, and the like.
[0005]
In recent years, auditory filters have been proposed as models for expressing the mechanism of frequency analysis of human hearing. This is a concept of expressing the frequency analysis mechanism of the human inner ear with a plurality of band filter banks. The shape of the individual filters (auditory filters) in this filter bank is usually measured using notch noise masking. So far, a simple method for measuring frequency resolution for a hearing impaired person using this auditory filter theory has been disclosed in Japanese Patent Laid-Open No. 6-327654. It is known that the shape of a human auditory filter can be modeled by a roex (p, r) filter.
[0006]
[Problems to be solved by the invention]
However, it is said that auditory filter measurement by notch noise masking can measure the shape of an individual subject's auditory filter with high accuracy. However, the time required for measurement is extremely long, and it is actually an otolaryngology clinical site. It was virtually impossible to measure individual hearing loss at any time in the field of hearing aid fittings. Japanese Patent Laid-Open No. 6-327654 proposes a method for measuring whether or not the frequency resolution has deteriorated in a short time. This method can determine whether or not the frequency resolution has deteriorated. It was not possible to measure the degree of deterioration, and thus the shape of the auditory filter itself.
[0007]
Furthermore, the shape of the auditory filter is said to change according to the level of the input signal. It is said that a masker (notch noise) level of about 40 dB / Hz is optimal for auditory filter measurement of a normal hearing person based on various past knowledge. The shape is said to be broad. However, not all hearing-impaired people have the same change in the shape of the auditory filter as that of normal hearing, depending on these levels. To date, there is no method for measuring the shape change characteristics of an auditory filter according to the level of an individual deaf person.
[0008]
The present invention has been made in view of such problems of the prior art, and its purpose is to efficiently and accurately perform diagnosis of hearing loss, grasp of hearing characteristics of a hearing impaired person, fitting of hearing aids, and the like. Therefore, an object of the present invention is to propose a method and apparatus for accurately estimating the shape of an auditory filter effective in a short time.
[0009]
[Means for Solving the Problems]
In order to solve the above-mentioned problem, the invention according to claim 1 is a method for estimating the shape of an auditory filter by obtaining a coefficient p of a roex (p, r) filter that models the shape of the auditory filter, and at a frequency f. A signal sound S ′ obtained by subtracting an arbitrary value a from a signal sound S obtained by adding an arbitrary value x to the minimum audible threshold T of the subject is generated, and then the frequency f considering the frequency characteristics of the outer ear and the middle ear of the subject is generated. A masker having a notch width g and a level N _X ′ included in the notch is generated, and then a test sound in which the masker is superimposed on the signal sound S ′ is presented to the subject. The notch width g _Xa is measured, and then, from the minimum notch width g _Xa, an upper limit value g _max of the notch width g suitable for the subject, the minimum notch width g _Xa , the value x corresponding to the coefficient r, and the value a calculate p, then calculate And estimates the shape of the auditory filter from said value x corresponding to the coefficient p and the coefficient r.
[0010]
According to a second aspect of the present invention, in the auditory filter shape estimation method according to the first aspect, the shape of the auditory filter is estimated using the value x as a parameter.
[0011]
The invention according to claim 3 is the method for estimating the shape of an auditory filter according to claim 1 or 2, wherein the limit masking level is based on the signal sound S obtained by adding an arbitrary value x to the minimum audible threshold value T of the subject at the frequency f. N _X 'is determined.
[0012]
According to a fourth aspect of the present invention, in the auditory filter shape estimation method according to the first, second, or third aspect, the upper limit value g _max of the notch width g suitable for the subject is estimated from the minimum notch width g _Xa. .
[0013]
According to a fifth aspect of the present invention, in the method for estimating the shape of an auditory filter according to the first, second, third, or fourth aspect, when measuring the limit masking level N _X ′ and / or the minimum notch width g _Xa , After the start of presentation, the signal sound presentation is started after a predetermined time interval.
[0014]
The invention according to claim 6 is an apparatus for estimating the shape of an auditory filter by obtaining a coefficient p of a roex (p, r) filter that models the shape of the auditory filter, and a signal that generates a signal sound of a predetermined frequency. Sound generator, noise generator that generates noise considering the frequency characteristics of the outer and middle ears of subjects without notches, and noise level change that amplifies and attenuates the noise generated by this noise generator to a predetermined level A notch width changing unit that gives the noise a notch including the frequency of the signal sound, and a notch noise superimposing unit that superimposes the notch noise output by the notch width changing unit on the signal sound output by the signal sound level changing unit And a test sound presenting unit that presents the test sound output by the notch noise superimposing unit to the subject, and a roex (p based on the notch width when the test subject can perceive the test sound and its upper limit value. To calculate the coefficient of r) filter p, those having coefficients obtained p, and auditory filter calculating section for estimating the filter shape from r, the auditory filter display unit for displaying the estimated filter shape.
[0015]
The invention according to claim 7 is the auditory filter shape estimation device according to claim 6, further comprising a signal sound level changing unit that amplifies and attenuates the signal sound generated by the signal sound generating unit to a predetermined level. is there.
[0016]
The invention according to claim 8 is the auditory filter shape estimation apparatus according to claim 6 or 7, wherein the upper limit value of the notch width suitable for the subject is calculated based on the notch width when the subject can perceive the test sound. The upper limit calculation part to perform is provided.
[0017]
The invention according to claim 9 is the auditory filter shape estimation device according to claim 6, 7 or 8, wherein the masker is presented to the subject when the limit masking level N _X 'and / or the minimum notch width g _Xa is measured. After starting, the presentation of the signal sound is started after a predetermined time interval.
[0018]
DETAILED DESCRIPTION OF THE INVENTION
Embodiments of the present invention will be described below with reference to the accompanying drawings. Here, FIG. 1 is a block diagram of the auditory filter shape estimation apparatus according to the present invention, FIG. 2 is a timing chart showing the relationship between the presentation of the test sound and the response of the subject, and FIG. 3 is the auditory filter shape estimation according to the present invention. FIG. 4 is a flowchart showing the procedure, FIG. 5 is a diagram showing the shape of the auditory filter obtained by the present invention, and FIG. 6 is a block diagram of the masker. Note that the masker in FIG. 3 has a uniform sound pressure level per 1 Hz, but in reality, it is a level that takes into account the frequency characteristics of the outer ear and the middle ear.
[0019]
As shown in FIG. 1, an auditory filter shape estimation apparatus according to the present invention includes a signal sound generator 1, a signal sound level changer 2, a noise generator 3, a noise level changer 4, a notch width changer 5, a notch A noise superimposing unit 6, an examination sound presenting unit 7, an upper limit calculating unit 8, an auditory filter calculating unit 9, an auditory filter display unit 10, a notch width display unit 11, a response unit 12 and the like are provided.
[0020]
The signal sound generation unit 1 outputs a sine wave signal having a predetermined frequency f as a signal sound (pure sound). The value of the frequency f can be set arbitrarily. The signal sound generation unit 1 may be configured by a CPU and generate a signal sound by a predetermined program, or may be configured by a memory and store a signal sound signal in advance.
[0021]
The signal sound level changing unit 2 amplifies and attenuates the signal sound generated by the signal sound generating unit 1 to a predetermined level. The signal sound level changing unit 2 is a signal sound having a minimum audible threshold value T [dBSPL] of a subject, a signal sound S having a level T + x [dBSPL] obtained by adding an arbitrary value x [dB] to the minimum audible threshold value T [dBSPL]. A signal sound S ′ of level T + x−a [dBSPL] obtained by subtracting an arbitrary value a [dB] from level T + x [dBSPL] is output. Note that x> a.
[0022]
The noise generation unit 3 generates noise having no notch considering the frequency characteristics of the outer ear and the middle ear of each subject so that the sound pressure level per 1 Hz becomes uniform when reaching the inner ear. In this case, for example, a uniform exciting noise (UEN) that is a masker that corrects the frequency characteristics of the outer ear and the middle ear of the subject may be generated. The noise generation unit 3 may be configured by a CPU and generate noise by a predetermined program, or may be configured by a memory and store a noise signal in advance.
[0023]
Here, in a normal hearing person, since the individual difference of the frequency characteristic in an outer ear and a middle ear is small, it is not necessary to change a correction value separately.
For normal hearing, correction values are determined using minimum audible pressure (MAP), minimum audible field (MAF), and equal loudness contour (ELC). It is also possible to generate a masker.
[0024]
There are two methods for measuring the absolute threshold of sound, which indicates the minimum level of sound that can be detected without any other disturbing sound. The first method is to measure the relationship between the frequency and the minimum audible sound pressure near the entrance of the ear canal or inside the ear canal, and the measured threshold is called the minimum audible sound pressure (MAP). Another method is to measure the level of sound presented from the speaker in a large anechoic room (a room surrounded by walls with high sound absorption) at the center of the listener's head. The threshold value is called the minimum audible field (MAF).
[0025]
In addition, there is a method in which 1000Hz pure sound with fixed sound level and inspection sound are presented alternately, and the intensity of the inspection sound is changed to match the volume with the pure sound. What is obtained is called an isosensitive curve (ELC) of loudness.
[0026]
On the other hand, the hearing impaired may have disabilities in the outer and middle ears, which can be predicted to be different from the frequency characteristics of the outer ear and middle ear of a normal hearing person. It is necessary to generate a masker by performing correction considering the above.
[0027]
In the case of a hearing-impaired person with a disability in the outer ear or the middle ear, the audiogram values of the bone conduction and the air conduction differ, so that the difference can be used to generate a masker by taking correction into account. Further, the correction of the frequency characteristics of the outer ear can be determined from the frequency characteristics near the entrance of the ear canal and the result of measuring the frequency characteristics by providing a probe microphone near the eardrum.
[0028]
The noise level changing unit 4 amplifies and attenuates the noise generated by the noise generating unit 3 to a predetermined level. The noise level changing unit 4 outputs noise of level N _X ′ that can mask the signal sound S of level T + x [dBSPL].
[0029]
The notch width changing unit 5 gives the noise output from the noise level changing unit 4 a notch whose center frequency f _C is the same as the frequency f of the signal sound (f _C = f). The notch width g of this notch is changed at any time according to the response of the subject. The notch width changing unit 5 may be configured as a filter that realizes a desired notch, or a plurality of notch noises (maskers) are stored in a memory in advance, and noise having various notches is selected as needed. May be used.
[0030]
The notch noise superimposing unit 6 superimposes the notch noise of the center frequency f _C , the notch width g, and the level N _X ′ output from the notch width changing unit 5 on the signal sound S ′ output from the signal sound level changing unit 2. Use inspection sound.
The inspection sound presenting unit 7 presents the inspection sound output by the notch noise superimposing unit 6 to the subject. The subject listens to the test sound and responds whether or not the signal sound S ′ can be perceived.
[0031]
The upper limit calculator 8 calculates the upper limit g _max of the notch width suitable for the subject based on the notch width g when the subject can perceive the test sound.
The auditory filter calculation unit 9 calculates the filter coefficient p of the roex (p, r) filter based on the notch width g and the upper limit g _max when the subject can perceive the signal sound S ′, and the obtained filter The shape of the auditory filter is estimated from the coefficients p and r.
[0032]
The roex (p, r) filter is defined by the following equation (1).
[0033]
[Expression 1]

[0034]
Here, p is a coefficient representing the bandwidth (tilt angle) of the filter, r is a coefficient representing the dynamic range of the filter, g is a value obtained by normalizing the distance from the center frequency of the auditory filter, and the notch width (= Δf / f _C ). In order to model the auditory filter by the equation (1), notch noise masking data Ps (g) obtained by notch noise masking is used. This is defined by the following equation (2).
[0035]
[Expression 2]

[0036]
Here, g _max is the upper limit value of the notch width g, K is the sensitivity of each individual, f _C is the center frequency of the notch noise, and N _X is the level of the masker M when it reaches the inner ear.
[0037]
The auditory filter display unit 10 displays the shape of the auditory filter from the filter coefficients p and r obtained by the auditory filter calculation unit 9.
The notch width display unit 11 displays the notch width g of the masker output from the notch width changing unit 5.
[0038]
The response unit 12 outputs a response signal when the signal sound S ′ can be perceived from the test sound presented by the test sound presenting unit 7 by the operation of the subject and a response signal when the signal sound S ′ cannot be perceived. Then, a response signal when the signal sound S ′ cannot be perceived is output to the notch width changing unit 5.
[0039]
It should be noted that only the response signal when the signal sound S ′ can be perceived is used, and when the response signal when the signal sound S ′ can be perceived for a predetermined time is not output, it is determined that the signal sound S ′ cannot be perceived and the process proceeds to the next step. .
[0040]
The operation of the auditory filter shape estimation apparatus configured as described above will be described. The notch noise superimposing unit 6 superimposes the masker M in consideration of the frequency characteristics of the outer ear and the middle ear of each subject so that the sound pressure level per 1 Hz is uniform when the signal sound S ′ reaches the inner ear. Create inspection sound. The created inspection sound is presented to the subject from the inspection sound presentation unit 7.
The subject listens to the test sound and responds by the response unit 12 as to whether or not the signal sound S ′ has been perceived.
[0041]
Here, as shown in FIG. 2, when the first test sound is presented to the subject, the level of the signal sound S ′ output by the signal sound level changing unit 2 for a predetermined time t at the start of the test sound presentation. Can be set to 0 (presentation of the signal sound S ′ is delayed from the masker M by a predetermined time t).
[0042]
The notch width changing unit 5 creates a new masker M according to the response and outputs it again to the notch noise superimposing unit 6. Here, the change of the notch width g of the masker M may be automatically performed according to the response of the subject (configured by a CPU or the like, and a dedicated program is prepared), or manually by the measurer each time. You may instruct.
[0043]
Even when a new examination sound is presented to the subject after increasing the notch width g, the signal sound S ′ output by the signal sound level changing unit 2 for a predetermined time t at the start of the presentation of the new examination sound. The level can be set to 0 (presentation of the signal sound S ′ can be delayed by a predetermined time t from the masker M).
[0044]
The notch width g being measured is displayed on the notch width display section 11. Then, the notch width changing unit 5 gradually increases the notch width g, and measures the minimum notch width (limit notch width) g _{Xa at} which the signal sound S ′ can be perceived by the response of the subject. As shown in FIG. 3, the minimum notch width g _Xa is equivalent to the bandwidth of a point attenuated by a [dB] from the apex in the notch noise masking data characteristic.
[0045]
The upper limit calculation unit 8 calculates the upper limit g _max using the minimum notch width g _Xa measured by the notch width changing unit 5.
The auditory filter calculation unit 9 uses the minimum notch width g _Xa measured by the notch width changing unit 5 and the upper limit value g _max calculated by the upper limit value calculation unit 8 to calculate the filter coefficient p of the roex (p, r) filter as follows: Calculate as follows. Here, the difference a between Ps (0) and Ps (g _Xa ) in the equation (2) is defined as the following equation (3).
[0046]
[Equation 3]

[0047]
Furthermore, the difference a is expressed by the following expression (4) from the expressions (2) and (3).
[0048]
[Expression 4]

[0049]
Here, the value of g _Xa is a measure, r is a 10 log ₁₀ r = -x using coefficients x defining the dynamic range, since a is an optional constant determined by the measurement conditions, the value of g _max is If determined, the value of the filter coefficient p can be calculated from the equation (4).
[0050]
Further, as a method for determining the value of g _max , the following method can be considered. Signal detection threshold and minimum audible threshold (hereinafter referred to as g _t) value matching notch width g is subjected to measurement using a masker sound pressure level per 1Hz is created so as to 40dBSPL in hearing person In this case, g _t = 0.6 to 0.7, and integration in the interval of g _max > 0.8 has little effect on the value of the notch noise masking data Ps (g).
[0051]
Therefore, referring to past measurement data of the auditory filter, the relationship between the value of g _{max and} the value of g _t is expressed by the following equation (5).
[0052]
[Equation 5]

[0053]
Here, the value of C is approximately 1.1 to 1.3. On the other hand, in the simple measurement method of the present invention, since the measured value is only the notch width g _Xa , the value of the notch width g _t at which the signal detection threshold matches the minimum audible threshold is unknown. In addition, the value of the notch width g _t to the signal detection threshold and minimum audible threshold matches the measurement result changes from (relationship between notch width g and the signal detection threshold [dBSPL]) by the dynamic range (corresponding to a value x) I know you will. Therefore, the value of g _max needs to be estimated for each value of x from the notch width g _Xa, and the relationship between the value of g _{max and} the value of g _Xa is expressed by the following equation (6).
[0054]
[Formula 6]

[0055]
Here, A _x is a coefficient that changes according to the value of x. The value of this A _x assumes that C = 1.2 in equation (5), calculates the value of g _max using 2 people hearing subjects the measurement result (the value of g _t), the g _max When experimentally determined from the value and the value of g _Xa, the following Table 1 is obtained.
[0056]
[Table 1]

[0057]
In Table 1, x = 10, 15, 20, 30 to measure the value of g _t in [dB], but calculates the value of A _x, for example x = 40, 50 A in [dB] _x The value of can also be prepared in advance.
[0058]
Therefore, the auditory filter calculation unit 9 substitutes the upper limit value g _max , the minimum notch width g _Xa , the value x and the value a corresponding to the coefficient r into the equation (4), and the filter coefficient p of the roex (p, r) filter. Is calculated. Further, the shape of the auditory filter is estimated from the filter coefficients p and r.
The auditory filter display unit 10 displays the filter shape from the value x corresponding to the filter coefficient p and the filter coefficient r.
[0059]
In this way, the value x corresponding to the filter coefficient r is set to 10, 20, 30, 40, 50 [dB], for example, and an estimated value of the filter coefficient p is obtained for each, and as shown in FIG. The shape of the auditory filter with the parameter as a parameter (notch width g on the horizontal axis and attenuation [dB] on the vertical axis) is displayed.
[0060]
Next, the auditory filter shape estimation method according to the present invention will be described with reference to the flowchart shown in FIG.
First, in step SP1, the air conduction and bone conduction audiograms including the frequency f are measured without adding noise. Here, the pure tone minimum audible threshold value at the frequency f is T [dBSPL / Hz]. In step SP2, a frequency (signal sound frequency) f to be measured is set.
[0061]
Next, in step SP3, the value of the coefficient x that determines the dynamic range of the notch noise masking data is set. Here, x = 10. Further, in step SP4, a pure sound (signal sound) S having a frequency f [Hz] and a level T + x [dBSPL] is created.
[0062]
Next, in step SP5, when the level of the signal sound S is sufficiently small and reaches the inner ear, the noise corrected to make the sound pressure level per 1 Hz uniform and the difference in the air bone conduction audiogram measured in step SP1. The minimum level (limit masking level) N _X ′ at which the subject cannot perceive the signal sound S is measured while superimposing the considered noise and gradually increasing the noise level.
[0063]
In step SP6, the noise notch width g at level N _X 'is set to a predetermined value. Here, g = 0.05. In step SP7, the value a [dB] subtracted from the level T + x [dBSPL] of the signal sound S is set to a predetermined value (x> a). Here, a = 5.
[0064]
Next, in step SP8, a pure tone (signal tone) S ′ having a frequency f [Hz] and a level T + x−a [dBSPL] is created.
Further, in step SP9, a masker M having a notch center frequency f _C (= f), a notch width g, and a level N _X ′ is created.
[0065]
Next, in step SP10, the masker M is superimposed on the signal sound S ′ and presented as a test sound to the subject. Further, in step SP11, the subject is made to judge whether or not the signal sound S ′ can be perceived from the inspection sound. If there is no response that the signal sound S ′ can be perceived, the process proceeds to step SP12 and the value of the notch width g is set. It is gradually increased (in this case, the increase range is set to 0.05), and steps SP9 to SP11 are repeated until there is a response that the signal sound S ′ can be perceived.
[0066]
Here, as shown in FIG. 2, when the test sound is first presented to the subject, the level of the signal sound S ′ output by the signal sound level changing unit 2 for a predetermined time t at the start of the test sound presentation. Can be set to 0 (presentation of the signal sound S ′ is delayed from the masker M by a predetermined time t). Similarly, when the new test sound is presented to the subject after the notch width g is increased in step SP12, the signal sound level changing unit 2 outputs only the predetermined time t at the start of the new test sound presentation. The level of the signal sound S ′ to be performed can be reduced to zero.
[0067]
If the timing of starting the presentation of the signal sound S ′ and the response timing are significantly different, the measurement may be performed again under the same conditions, and the measurement may be repeated until a stable response is obtained. Further, the predetermined time t does not need to be a constant value, and may be changed randomly every measurement.
[0068]
In this way, when the first inspection sound is presented, the signal sound S ′ is presented after being delayed by a predetermined time t from the masker M, or only the signal sound S ′ is displayed when the value of the notch width g of the masker M is changed. It is also possible to judge the validity of the response of the subject by temporarily stopping and presenting the signal sound S ′ to the subject after the predetermined time t has elapsed and observing the response timing of the subject at that time. Because it is considered.
[0069]
As shown in FIG. 2, after the signal sound S ′ is presented to the subject, there is a response that the signal sound S ′ is perceived, and if there is no response that the signal sound S ′ is perceived simultaneously with the stop of the signal sound S ′, Can be determined to be normal. On the other hand, when there is a response that the signal sound S ′ is perceived before the signal sound S ′ is presented to the subject, and the response that the signal sound S ′ is perceived after the stop of the signal sound S ′ is maintained, etc. Since it can be determined that the response of the subject is not normal, remeasurement may be performed, or the settings of the level of the signal sound S ′ and the level of the masker M may be changed.
[0070]
If a response that the signal sound S ′ can be perceived in step SP11, the process proceeds to step SP13.
In step SP13, the notch width g at that time is determined as the minimum notch width g _Xa of the subject. Further, the value of A _{x at} the set value x (= 10 [dB]) is determined using Table 1 (A _x = 3.15).
[0071]
Next, at step SP14, the value of A _x (3.15) and the value of the measured minimum notch width g _Xa are substituted into Equation (6) to determine the upper limit value g _max of the notch width g suitable for the subject. To do.
[0072]
In step SP15, the minimum notch width g _Xa measured in the equation (4), the upper limit value g _max of the notch width estimated from the minimum notch width g _Xa , the value x and the value a corresponding to the preset coefficient r are obtained. By substituting, an estimated value of the filter coefficient p _X (x = 10) of the roex (p, r) filter is calculated.
[0073]
Next, in step SP16, it is determined whether or not x = 50. If x = 50 is not satisfied, the process proceeds to step SP17 to set x = x + 10, and steps SP4 to SP17 are repeated until x = 50. An estimated value of the filter coefficient p _X for x (x = 20, 30, 40, 50 [dB]) is calculated from the equation (4). Here, the increase width of x is set to 10, and measurement up to x = 50 is performed.
[0074]
If it is determined in step SP16 that x = 50, as shown in FIG. 5, an auditory filter using the value x as a parameter (x = 10, 20, 30, 40, 50 [dB]) in step SP18. (The horizontal axis is the notch width g and the vertical axis is the attenuation [dB]).
[0075]
In the embodiment of the present invention, as shown in FIG. 6A, the masker M is created by a method of adding notches to noise considering the frequency characteristics of the outer ear and middle ear of the subject. The masker M may be composed of two band noises (high frequency side and low frequency side) as shown in FIG.
[0076]
In the above-described embodiment of the present invention, the frequency f of the signal sound signal output from the signal sound level changing unit 2 and the center frequency f _C of the masker M output from the notch width changing unit 5 are matched ( The case of f = f _C ) has been described, but it is not always necessary to match, and it is only necessary that the signal sound signal is included within the notch width.
[0077]
Further, in the above-described embodiment of the present invention, when determining the minimum notch width (limit notch width) g _Xa that can be perceived by the subject, first, the notch width g of the masker M is set narrow (g = 0.05). The test sound is presented, and when the subject cannot perceive the signal sound S ′ from the test sound, the procedure of presenting the test sound again until the signal sound S ′ can be perceived by gradually increasing the notch width g of the masker M. explained.
[0078]
However, since it is only necessary to obtain the minimum notch width g _Xa that can be perceived by the subject, first, the test sound is presented by setting the notch width g of the masker M wide, and the subject can perceive the signal sound S ′ from the test sound. The procedure may be such that the test sound is presented again until the notch width g of the masker M is gradually narrowed until the signal sound S ′ cannot be perceived.
[0079]
In the above-described embodiment of the present invention, the case where the so-called limit method is used for setting the signal sound level and the limit masking level N _X ′ has been described. You may set based on psychophysical measurement methods, such as the two alternative forced choice method (Two alternative forced choice) which makes it select from either of the test | inspection sound which has not added.
[0080]
Furthermore, in the above-described embodiment of the present invention, the case where the limit method is used to determine the minimum notch width g _Xa that can be perceived by the subject has been described. However, based on a psychophysical measurement method such as a two-party forced selection method. You may ask.
[0081]
In the above-described embodiment of the present invention, the filter shape is estimated using the model of the auditory filter as the roex (p, r) filter. However, the model function is not necessarily a roex (p, r) filter. However, as long as it is appropriate as a model of an auditory filter, other functions (for example, roex (p) with ₁₀ log ₁₀ r = 0 in the roex (p, r) filter) may be used.
[0082]
Further, in the above-described embodiment of the present invention, the presentation of the signal sound S ′ is delayed from the masker M by a predetermined time t at the start of the inspection sound presentation in step SP10 when measuring the minimum notch width g _Xa . Although the legitimacy of the response is judged, when the masker level is changed at the time of measuring the limit masking level N _X ′ at step SP5, a similar time delay is given to the signal sound S to measure the limit masking level N _X. You may also judge the legitimacy.
[0083]
【The invention's effect】
As described above, according to the first aspect of the invention, by generating a masker with notch width g and level N _X ′ including frequency f in consideration of frequency characteristics of the subject's outer ear and middle ear, It is possible to accurately estimate the shape of an auditory filter effective for efficiently and accurately performing hearing loss diagnosis, hearing characteristics of a hearing impaired person, hearing aid fitting, and the like.
[0084]
According to the second aspect of the present invention, the shape of the auditory filter corresponding to the input signal level can be estimated, and the relationship between the input signal level and the auditory filter can be known.
[0085]
According to the invention of claim 3, by determining the limit masking level N _X ′ based on the signal sound S obtained by adding an arbitrary value x to the minimum audible threshold T of the subject at the frequency f, It is possible to estimate the shape of the auditory filter in consideration of frequency characteristics in the outer ear and the middle ear.
[0086]
According to the fourth aspect of the invention, the upper limit value g _max of the notch width g suitable for the subject is estimated from the minimum notch width g _Xa, thereby more accurately considering the frequency characteristics in the outer ear and the middle ear of the subject. The shape of the filter can be estimated.
[0087]
According to the invention which concerns on Claim 5, the validity of a test subject's response can be judged easily from the timing of signal sound presentation and the test subject's response timing.
[0088]
According to the invention of claim 6, by providing a noise generation unit that generates noise in consideration of the frequency characteristics of the outer ear and middle ear of the subject, diagnosis of hearing loss, grasp of hearing characteristics of the deaf person, hearing aid fitting, etc. It is possible to accurately estimate the shape of an effective auditory filter for efficient and accurate execution.
[0089]
According to the seventh aspect of the invention, the shape of the auditory filter can be estimated more accurately by including the signal sound level changing unit that amplifies and attenuates the signal sound generated by the signal sound generating unit to a predetermined level. it can.
[0090]
According to the eighth aspect of the present invention, by providing the upper limit value calculation unit that calculates the upper limit value of the notch width suitable for the subject based on the notch width when the subject can perceive the test sound, the auditory sense can be more accurately heard. The shape of the filter can be estimated.
[0091]
According to the invention which concerns on Claim 9, the validity of a test subject's response can be judged easily from the timing of signal sound presentation and the test subject's response timing.
[Brief description of the drawings]
FIG. 1 is a configuration diagram of an auditory filter shape estimating apparatus according to the present invention. FIG. 2 is a timing chart showing the relationship between the presentation of a test sound and the response of a subject. FIG. 4 is a flowchart showing the procedure of the method for estimating the shape of the auditory filter according to the present invention. FIG. 5 is a diagram showing the shape of the auditory filter obtained by the present invention.
DESCRIPTION OF SYMBOLS 1 ... Signal sound production | generation part, 2 ... Signal sound level change part, 3 ... Noise production | generation part, 4 ... Noise level change part, 5 ... Notch width change part, 6 ... Notch noise superimposition part, 7 ... Test sound presentation part, 8 ... upper limit calculation unit, 9 ... auditory filter calculation unit, 10 ... auditory filter display unit, 11 ... notch width display unit, 12 ... response unit.

Claims (9)

A method for estimating the shape of an auditory filter by obtaining a coefficient p of a roex (p, r) filter that models the shape of an auditory filter, wherein an arbitrary value x is added to the minimum audible threshold T of a subject at a frequency f. A signal sound S ′ obtained by subtracting an arbitrary value a from the signal sound S, and then a masker having a notch width g and a level N _X ′ including the frequency f considering the frequency characteristics of the outer ear and the middle ear of the subject. Next, the test sound in which the masker is superimposed on the signal sound S ′ is presented to the subject, and the minimum notch width g _Xa of the subject is measured while changing the notch width g, and then the minimum notch width g _Xa The coefficient p is calculated from the upper limit value g _max of the notch width g suitable for the subject, the minimum notch width g _Xa , the value x corresponding to the coefficient r, and the value a, and then the calculated coefficient p and coefficient r Listen from the corresponding value x Shape estimation method of auditory filter and estimates the shape of the filter. 2. The method for estimating an auditory filter shape according to claim 1, wherein the shape of the auditory filter is estimated using the value x as a parameter. 3. The method of estimating a shape of an auditory filter according to claim 1, wherein the limit masking level N _X ′ is determined based on a signal sound S obtained by adding an arbitrary value x to the minimum audible threshold value T of the subject at the frequency f. A method for estimating the shape of a characteristic auditory filter. 4. The method for estimating the shape of an auditory filter according to claim 1, wherein the upper limit value g _max of the notch width g suitable for the subject is estimated from the minimum notch width g _Xa. . 5. The shape estimation method for an auditory filter according to claim 1, 2, 3 or 4, wherein when the limit masking level N _X ′ and / or the minimum notch width g _Xa is measured, the masker is presented to the subject after starting. A method for estimating the shape of an auditory filter, wherein the presentation of a signal sound is started after a time interval of. A device for estimating a shape of an auditory filter by obtaining a coefficient p of a roex (p, r) filter that models the shape of an auditory filter, a signal sound generating unit that generates a signal sound of a predetermined frequency, and an outer ear of a subject A noise generation unit that generates noise not having a notch width in consideration of the frequency characteristics of the middle ear, a noise level change unit that amplifies and attenuates the noise generated by the noise generation unit to a predetermined level, and the signal to noise A notch width changing unit for providing a notch including the frequency of the sound, a notch noise superimposing unit for superimposing the notch noise output by the notch width changing unit on the signal sound output by the signal sound level changing unit, and the notch noise superimposing unit A test sound presenting unit that presents the test sound output by the subject to the test subject, and a coefficient of the roex (p, r) filter based on the notch width and the upper limit value when the test subject can perceive the test sound Calculates the coefficients resulting p, and auditory filter calculating section for estimating the filter shape from r, the shape estimation apparatus of the auditory filter characterized in that it comprises an auditory filter display unit for displaying the estimated filter shape. The shape estimation device for an auditory filter according to claim 6, further comprising a signal sound level changing unit that amplifies and attenuates the signal sound generated by the signal sound generation unit to a predetermined level. apparatus. The shape estimation device for an auditory filter according to claim 6 or 7, further comprising an upper limit value calculation unit that calculates an upper limit value of a notch width suitable for the subject based on the notch width when the subject can perceive the test sound. An auditory filter shape estimation device characterized by the above. 9. The shape estimation apparatus for an auditory filter according to claim 6, 7 or 8, wherein when a limit masking level N _X ′ and / or a minimum notch width g _Xa is measured, a predetermined time has elapsed after the masker is presented to the subject. An auditory filter shape estimation apparatus, wherein presentation of a signal sound is started after an interval.

Images