JP4189042B2

JP4189042B2 - Loudspeaker

Info

Publication number: JP4189042B2
Application number: JP06143497A
Authority: JP
Inventors: 実福島; 博昭竹山; 香子田中; 章寺澤
Original assignee: Panasonic Corp; Matsushita Electric Works Ltd
Current assignee: Panasonic Corp; Panasonic Electric Works Co Ltd
Priority date: 1997-03-14
Filing date: 1997-03-14
Publication date: 2008-12-03
Anticipated expiration: 2017-03-14
Also published as: JPH10257159A

Description

【０００１】
【発明の属する技術分野】
本発明は、家庭内、ビルディング、工場等で用いられる拡声通話機に関するものである。
【０００２】
【従来の技術】
従来より、インターホンや電話機あるいはＰＨＳ等の拡声通話機においては、スピーカからマイクロホンへの音響フィードバックおよびハイブリッド回路（２−４線変換回路）におけるインピーダンスの不整合により閉ループが形成され、増幅器の利得が大きすぎる等の理由により上記閉ループの利得が１倍以上になるとハウリングが生じるため、通話品質を確保する上でハウリングの抑圧が必要不可欠な課題となっていた。
【０００３】
そこで従来は、送受話信号のレべルに応じて受話信号または送話信号に所定量の損失を挿入することで閉ループ利得を抑圧するというハウリング抑圧方式が用いられてきた。また、別の方式としてエコーキャンセラを用いるものもあるが、エコーキャンセラにおける適応フィルタの係数が収束していない過渡状態や系の変動によりエコー経路が急激に変化した場合等において不安定化しやすいため、挿入損失と併用する場合が多い。
【０００４】
上記何れの場合においても、挿入損失量を大きくしすぎた場合には通話中に切断感を生じる（音声が途切れる）等の通話品質の劣化を招くため、挿入損失量を必要最小限とすることが望ましい。一方、拡声通話機の前面に手や顔を近づけたときのように音響的反射係数が急激に増加した場合に閉ループ利得が１倍以上となり、その結果、ハウリングを生じることがある。このような問題に対して、従来、図１２に示すように閉ループ内にハウリング検出器２４を設け、ハウリングを検出した場合に送話側並びに受話側の信号経路に設けた減衰器６₁，６₂を制御器７により制御して挿入損失量を通常よりも大きい値に設定する制御方式が用いられていた。なお、同図中、１はマイクロホン、３はスピーカ、２及び４は増幅器、５はハイブリッド（２−４線変換）回路である。
【０００５】
【発明が解決しようとする課題】
ところが上記従来例では、ハウリング検出器２４でハウリングを検出してから挿入損失によりハウリングを抑圧する処理を行うため、ハウリング発生前後において挿入損失量が大きく変化し、これによって却って通話品質の劣化を招くという問題があった。また、ハウリング検出時に挿入する損失量は、原理的に必要最小限の値に制御することができずにかなり大きな値となるため、双方向同時通話性能の実現が困難となっていた。
【０００６】
本発明は、上記問題点に鑑みて為されたものであり、その目的とするところは、ハウリングの抑圧によって安定した通話が実現できるとともに、挿入損失量を必要最小限とすることにより双方向同時通話の実現可能性を高めることができる拡声通話機を提供することにある。
【０００７】
【課題を解決するための手段】
請求項１の発明は、上記目的を達成するために、集音した音を送話側の音声信号として出力するマイクロホンと、マイクロホンからの音声信号を増幅する第１の増幅手段と、受話側の音声信号に応じて鳴動するスピーカと、スピーカへ出力される音声信号を増幅する第２の増幅手段と、送話側及び受話側と外部の通話回線との間で２−４線変換を行う２−４線変換手段と、送話側及び受話側の少なくとも一方の信号経路に所定量の損失を挿入する損失挿入手段と、損失挿入手段から挿入される損失量を可変制御する制御手段と、信号経路に送出したサンプル信号に対する応答信号に応じて、マイクロホン及びスピーカを通じて形成される閉ループにおける利得余裕を推定するとともに制御手段を介して推定結果に基づく損失量の調整を行う挿入損失量調整手段とを備え、挿入損失量調整手段が、信号経路にサンプル信号を送出するサンプル信号発生器と、音声信号にサンプル信号を加算する加算器と、サンプル信号に対する応答信号の包絡線を検波する包絡線検波器と、包絡線検波器の出力信号に基づいて閉ループにおける利得余裕を推定する閉ループ利得余裕推定部と、閉ループ利得余裕推定部にて利得余裕の推定処理が行われる際に信号経路への挿入損失量を、所望の利得余裕値を得るために必要な値よりも大きな値に設定するとともに推定処理終了後に推定結果から求められる調整量を設定前の元の値に加えた値へ切り換える挿入損失量切換部とを具備して成り、閉ループ利得余裕推定部が、包絡線検波器の出力信号の微小時間における変位を求める微分器を有し、微分器で求めた包絡線検波器の出力信号の微小時間における変位に基づいて閉ループにおける利得余裕を推定して成ることを特徴とし、ハウリングが発生するまでの閉ループ系の利得余裕を所望の値に維持することができ、ハウリングを生じることなく安定した通話品質を維持することができる。また、必要以上に挿入損失量を大きくすることがないため、双方向同時通話性能の実現可能性が高まる。さらに、閉ループ系にサンプル信号を入力し、その応答信号の包絡線成分から閉ループ系の時定数に関する情報を抽出し、抽出された情報から閉ループ系の発振に対する利得余裕値を推定し、推定結果に基づいて送話側及び受話側への挿入損失量を制御することにより、随時閉ループ系の利得余裕値を所定値以上に維持することができる。しかも、応答信号の包絡線成分の絶対値ではなく、時間に対する傾きから推定処理を行うため、サンプル信号のレべル、信号送出時間や包絡線検波器の応答特性等の影響を除去することができる。さらに、利得余裕値の推定処理を行っている間に閉ループ系の変動などがあった場合でもハウリングが生じるのを防ぐことができる。
【０００８】
請求項２の発明は、請求項１の発明において、サンプル信号発生器をインパルス信号発生器として成ることを特徴とし、利得余裕値と密接な関係にあるインパルス応答特性から閉ループ系の利得余裕値を推定することができる。
請求項３の発明は、請求項１の発明において、サンプル信号発生器をバーストノイズ発生器として成ることを特徴とし、バーストノイズを用いた場合にも、信号送出時間が閉ループ系の遅延時間に対して十分に短ければ、その応答波形の包絡線成分より閉ループ系の時定数に関する情報を抽出することができる。しかもバーストノイズの信号送出時間を適切な値とすることにより通話中における違和感をなくし、サンプル信号の送出による通話品質の劣化を抑えることができる。
【０００９】
請求項４の発明は、請求項１〜３の何れかの発明において、挿入損失量調整手段が通話中の無音区間を検出する無音検出器を具備し、無音検出器において無音区間が検出されたときにサンプル信号発生器からサンプル信号を発生させて成ることを特徴とし、観測される応答信号が音声信号に重畳されることがなく、精度良く推定処理を行うことができる。
【００１０】
請求項５の発明は、請求項１〜４の何れかの発明において、通話の開始及び終了を検出する検出手段を備え、挿入損失量調整手段が、検出手段の検出結果に応じて非通話時に閉ループにおける利得余裕を推定するとともに通話開始前に推定結果に基づく所要の損失量を信号経路に挿入させて成ることを特徴とし、非通話時においても閉ループ系の利得余裕値を所望の値とすることができるため、通話開始直後から安定した通話を実現することができる。
【００１１】
【発明の実施の形態】
図１は本発明の参考例を示すブロック図であり、集音した音を送話側の音声信号（以下、送話信号と呼ぶ）として出力するマイクロホン１と、マイクロホン１からの送話信号を増幅する第１の増幅器２と、受話側の音声信号（以下、受話信号と呼ぶ）に応じて鳴動するスピーカ３と、スピーカ３へ出力される受話信号を増幅する第２の増幅器４と、送話側及び受話側と外部の通話回線との間で２−４線変換を行う２−４線変換手段たるハイブリッド回路５と、送話側及び受話側の信号経路に所定量の損失を挿入する損失挿入手段たる減衰器（アッテネータ）６₁，６₂と、各減衰器６₁，６₂から挿入される損失量を可変制御する制御器７と、信号経路に送出したサンプル信号に対する応答信号に応じて、マイクロホン１及びスピーカ３を通じて形成される閉ループ系における利得余裕を推定するとともに制御器７を介して推定結果に基づく損失量の調整を行う挿入損失量調整手段８とを備えている。なお、本参考例では挿入損失量調整手段８を受話側の信号経路上に設けているが、送話側の信号経路上に設けることも勿論可能である。
【００１２】
一方、図２は上記挿入損失量調整手段８の具体的な構成を示すブロック図であり、受話側の信号経路にサンプル信号たるインパルス信号を送出するインパルス信号発生器９と、受話信号にインパルス信号を加算する加算器１０と、インパルス信号に対する応答信号の包絡線を検波する包絡線検波器１１と、包絡線検波器１１の出力信号に基づいて閉ループ系における利得余裕を推定する閉ループ利得余裕推定部１２とを備えている。ここで、閉ループ利得余裕推定部１２は、後述するように包絡線検波器１１の出力信号レベルを予め求めた閾値レベルと比較する比較器１３と、比較結果に基づいて閉ループ系の利得余裕値を推定する判定部１４とを具備している。なお、サンプル信号として用いるインパルス信号は、パルス幅が充分に短い単一パルス信号であってもよい。また包絡線検波器１１は、整流回路とローパスフィルタ回路の合成回路や、巡回型ローパスフィルタやリーク積分器等のデジタル回路、あるいはＤＳＰ（ＤigitalＳignalＰrocessor）などの信号処理手段によって構成することができる。
【００１３】
図３に示すように、本参考例においては第２の増幅器４→スピーカ３→マイクロホン１→第１の増幅器２→減衰器６₁→ハイブリッド回路５→減衰器６₂→挿入損失量調整手段８→第２の増幅器４により閉ループが形成されている。ここで、各部における伝達関数を以下のように定義する。
Ｓ：スピーカ３の電気機械変換特性
Ｇ：スピーカ３からマイクロホン１への音響伝達特性
Ｍ：マイクロホン１の音響電気変換特性
Ｋｒ：第２の増幅器４の増幅特性
Ｋｘ：第１の増幅器２の増幅特性
Ａｒ：受話側の減衰器６₂の減衰特性
Ａｘ：送話側の減衰器６₁の減衰特性
Γ：ハイブリッド回路５における反射伝達関数
また、インパルス信号発生器９から出力されるインパルス信号をＰ、外部の通話回線から伝送されてくる遠端話者音声入力信号をＹ、マイクロホン１の集音する近端話者音声信号と周囲雑音との和をＸとすると、挿入損失量調整手段８の構成要素の一つである加算器１０の出力信号（応答信号）Ｑは下記式で表される。
【００１４】
【式１】

【００１５】
なお、上記式のＬ(s)は上記閉ループ系における一巡伝達関数、ｓはラプラス変数をそれぞれ表す。
ここで、閉ループ系の安定性は上記一巡伝達関数Ｌ(s)により判別することができる。すなわち、極座標系における一巡伝達関数Ｌ(s)のθ成分（＝∠Ｌ(s)）が∠Ｌ(s)＝２ｎπ（ｎは整数）となる全ての周波数において、ｒ成分（＝｜Ｌ(s)｜）が｜Ｌ(s)｜＜１ならば閉ループ系は安定、｜Ｌ(s)｜≧１となる周波数が存在すれば閉ループ系は不安定となり、その周波数において発振してハウリングが生じる。また、閉ループ系が安定である場合に、∠Ｌ(s)＝２ｎπとなる全ての周波数における利得の最大値をＬ_MAXとすれば、閉ループ系の利得余裕値は１／Ｌ_MAXで表される。よって、閉ループ系の安定性の尺度は閉ループ利得余裕値により評価することができる。
【００１６】
一方、閉ループ利得余裕値は、閉ループ系のインパルス応答特性と密接な関係があり、閉ループ利得余裕値が大きいほどインパルス応答信号Ｑの振幅が時間とともに急激に減衰し、閉ループ利得余裕値が小さいほど減衰が緩やかになる。そこで本発明は、閉ループ利得余裕推定部１２において閉ループ利得余裕値を推定するためのサンプル信号（インパルス信号）Ｐを上記閉ループ系に与えたときの応答信号Ｑを観測し、その応答信号Ｑの包絡線成分から閉ループ系の時定数に関する情報を抽出して閉ループ利得余裕値の推定を行うとともに、挿入損失量調整手段８において推定結果に基づき、閉ループ利得余裕値をハウリングが生じない所望の値となるように閉ループ系への挿入損失量を調整する。
【００１７】
すなわち、上述のように包絡線検波器１１で得られる応答信号Ｑの包絡線成分の時間特性が閉ループ利得余裕値が大きいほど減衰が早く且つ小さいほど減衰が緩やかになるという性質を有することから、閉ループ利得余裕推定部１２において事前に学習された種々の利得余裕値に対する包絡線検波器１１の出力データから閾値レベルを求めておき、比較器１３において観測される包絡線検波器１１の出力信号レベルを上記閾値レベルと比較することにより、その比較結果に基づいて判定部１４にて閉ループ利得余裕値が推定できる。そして、その推定結果から、閉ループ利得余裕値を設計仕様で定めた値とするために必要な損失量を挿入するべく、制御器７に信号を伝送して制御器７によって減衰器６₁，６₂の減衰量を調節している。
【００１８】
上述のように本参考例によれば、インパルス信号に対する応答信号から、閉ループ系でハウリングが生じない利得余裕値を推定し、その推定結果に基づいて信号経路に挿入する損失量を調整する挿入損失量調整手段８を備えているので、通話中においても閉ループ系の安定度に応じて挿入損失量を制御し、常に利得余裕値を仕様で定めた値に維持することができ、ハウリングを生じることなく安定した通話品質を維持することができる。また、従来例に比較して必要以上に挿入損失量を大きくする必要がないため、双方向同時通話性能の実現可能性が高まるという利点もある。
【００１９】
なお、本参考例ではサンプル信号発生器にインパルス信号発生器９を用いたが、代わりにバーストノイズ発生器を用いてバーストノイズをサンプル信号に用いてもよい。この場合には、信号を送出している状態から送出を停止した瞬間からの閉ループ系の過渡応答を観測し、その包絡線成分から閉ループ系の時定数に関する情報を抽出する。而してサンプル信号にバーストノイズを用いた場合にも、信号送出時間が閉ループ系の遅延時間に対して十分に短ければ、その応答波形の包絡線成分より閉ループ系の時定数に関する情報を抽出することができ、バーストノイズの信号送出時間を適切な値とすることにより通話中における違和感をなくし、サンプル信号の送出による通話品質の劣化を抑えることができるという利点がある。但し、種々の利得余裕値に対する閾値レべルを求めておく必要があることはインパルス信号の場合と同様である。
【００２０】
（実施形態１）
図４は本発明の実施形態１における挿入損失量調整手段８を示すブロック図である。すなわち、本実施形態は挿入損失量調整手段８を除く他の構成が参考例と共通であるので、共通する部分については同一の番号を付して説明及び一部図示は省略する。
【００２１】
本実施形態における挿入損失量調整手段８においては、閉ループ利得余裕推定部１２に包絡線検波器１１の出力信号の時間微分を求めるための微分器１５が設けてある。つまり、包絡線検波器１１から出力される応答信号の包絡線成分の減衰曲線の傾きが利得余裕値が大きいほど急となり、利得余裕値が小さいほど緩やかとなる性質を利用して、微分器１５によって上記減衰曲線の傾きを求め、事前に求めておいた種々の利得余裕値に対する閾値レべルと比較器１３において比較し、その比較結果に基づいて判定部１４にて閉ループ利得余裕値が推定できるものである。後は参考例と同様に推定結果に基づいて挿入損失量が調整される。
【００２２】
上述のように本実施形態によれば、応答信号の時間微分を求める微分器１５を閉ループ利得余裕推定部１２に設けたので、参考例のように応答信号の包絡線成分の絶対値ではなく、時間に対する傾き（時間微分値）から推定処理を行うため、サンプル信号（インパルス信号あるいはバーストノイズなど）のレべル、信号送出時間や包絡線検波器１１の応答特性等の影響を除去することができるという利点がある。
【００２３】
（実施形態２）
図５は本発明の実施形態２における挿入損失量調整手段８を示すブロック図である。すなわち、本実施形態は挿入損失量調整手段８を除く他の構成が実施形態１と共通であるので、共通する部分については同一の番号を付して説明及び一部図示は省略する。
【００２４】
本実施形態における挿入損失量調整手段８においては、受話信号のレベルから（あるいは送話信号のレベルであってもよい）、通話中の無音区間を検出する無音検出器１６と、サンプル信号発生器（インパルス信号発生器９あるいはバーストノイズ発生器）と加算器１０の間に挿入され、無音検出器１６にて無音区間が検出された場合にのみオンされるスイッチ１７とが設けてあって、通話中の無音区間にのみ閉ループ系にサンプル信号（インパルス信号又はバーストノイズ）を入力し、閉ループ利得余裕推定部１２にて利得余裕値の推定処理を行っている。
【００２５】
例えば、マイクロホン１で集音した受話信号にインパルス信号等のサンプル信号が重畳された場合について考えてみる。上述の式１を簡略化して表すと下記式のようになる。
Ｑ(s)＝Ｈ₁(s)Ｐ(s)＋Ｈ₂(s)Ｙ(s)＋Ｈ₃(s)Ｘ(s)…式２
ここでＨ₁(s)〜Ｈ₃(s)の分母が共通であるため、これらのうちの一つが安定であることが確認できれば、全て安定な伝達関数となる。さらに上記式を時間領域で表すと下記式のようになる。
【００２６】
ｑ(t)＝ｈ₁(t)＊ｐ(t)＋ｈ₂(t)＊ｙ(t)＋ｈ₃(t)＊ｘ(t)…式３
但し、Ｈ₁(s)＝Ｌ〔ｈ₁(t)〕，Ｈ₂(s)＝Ｌ〔ｈ₂(t)〕，Ｈ₃(s)＝Ｌ〔Ｌ〔ｈ₃(t)〕であり、Ｌ〔〕はラプラス変換、＊は畳み込み演算を各々表す。特にサンプル信号ｐ(t)がインパルス信号δ(t)の場合には、ｈ₁(t)＊δ(t)＝ｈ₁(t)であるから上記式は下記式のように表される。
【００２７】
ｑ(t)＝ｈ₁(t)＋ｈ₂(t)＊ｙ(t)＋ｈ₃(t)＊ｘ(t)…式４
従って、遠端話者音声信号ｙ(t)及び近端話者音声信号ｘ(t)がともにゼロである場合に、図３に示すように加算器１０の出力信号からインパルス応答ｈ₁(t)を直接観測することができる。しかし、これらの遠端話者音声信号ｙ(t)及び近端話者音声信号ｘ(t)がインパルス信号δ(t)に重畳されて閉ループ中を伝搬する場合、式４右辺の第２項及び第３項がノイズ成分となり、推定処理が困難になる場合がある。
【００２８】
そこで、本実施形態では無音検出器１６にて通話中の無音区間を検出し、その無音区間に推定用のサンプル信号（インパルス信号又はバーストノイズ）をサンプル信号発生器（インパルス信号発生器９あるいはバーストノイズ発生器）から発生させることにより、観測される応答信号が音声信号に重畳されることがなく、精度良く推定処理を行うことができる。
【００２９】
（実施形態３）
図６は本発明の実施形態３における挿入損失量調整手段８を示すブロック図である。すなわち、本実施形態は挿入損失量調整手段８を除く他の構成が実施形態１と共通であるので、共通する部分については同一の番号を付して説明及び一部図示は省略する。
【００３０】
本実施形態における挿入損失量調整手段８には、閉ループ利得余裕推定部１２にて利得余裕の推定処理が行われる際に信号経路への挿入損失量を比較的に大きな値に設定するとともに推定処理終了後に推定結果から求められる調整量を設定前の元の値に加えた値へ切り換える挿入損失量切換部１８が設けてある。また閉ループ利得余裕推定部１２では、推定処理中に挿入損失量切換部１８により追加して挿入された損失量を考慮し、この追加された損失量を差し引いて閉ループ系の利得余裕値を推定する。そして、推定処理の完了後は、挿入損失量切換部１８にて所望の利得余裕値を得るために必要な値に挿入損失量を調整する。
【００３１】
例えば、図７に示すようにマイクロホン１やスピーカ３の近傍に人が近づいた場合にはスピーカ３からマイクロホン１への音響伝達特性Ｇが変動し、あるいは図８に示すように閉ループ系に従来周知の構成を有するエコーキャンセラ１９が閉ループ系に挿入された場合には、それまでに収束していた適応フィルタ１９ａの係数がサンプル信号（インパルス信号あるいはバーストノイズ）の入力によって攪乱され、結果的に閉ループ系の特性が変動し、閉ループ系への挿入損失量が小さいときに上記変動が急激に生じると閉ループ系が不安定化してハウリングが生じるおそれがある。しかしながら、本実施形態おいては、挿入損失量切換部１８が閉ループ利得余裕推定部１２にて利得余裕の推定処理が行われる際に信号経路への挿入損失量を比較的に大きな値に設定するので、上記のようなハウリングの発生を防止することができる。
【００３２】
また図９及び図１０に示すようなＡＧＣ（自動利得制御）回路２０が閉ループ系に挿入された構成においては、上記推定処理中にレベルの大きな遠端話者音声信号Ｙが伝送されてきた場合に、ＡＧＣ回路２０における利得調整機能によって閉ループ系の利得が変化してしまい、利得余裕値が正しく推定できなくなるおそれがある。しかしながら、本実施形態では推定処理が行われる際に充分に大きな損失量が挿入されているため、遠端話者音声信号Ｙのレベルを充分に低減して推定処理中にＡＧＣ回路の利得制御がかかるのを防止することができる。
【００３３】
上述のように本実施形態によれば、利得余裕の推定処理が行われる際に信号経路への挿入損失量を比較的に大きな値に設定するとともに推定処理終了後に推定結果から求められる調整量を設定前の元の値に加えた値へ切り換える挿入損失量切換部１８を挿入損失量調整手段８に備えたので、利得余裕値の推定処理を行っている間に上記のような閉ループ系の変動などがあった場合でもハウリングが生じるのを防ぐことができるという利点がある。
【００３４】
（実施形態４）
図１１は本発明の実施形態４を示すブロック図であり、基本的な構成は実施形態１〜３と共通であるので共通する部分については同一の符号を付して説明は省略し、本実施形態の特徴となる部分についてのみ説明する。
本実施形態は、通話の開始及び終了を検出する通話／非通話検出部２１と、送話信号を増幅する第１の増幅器２と減衰器６₁との間の信号経路に挿入されて通話／非通話検出部２１によってオン・オフされる通話回路スイッチ２２とを備え、挿入損失量調整手段８が、上記通話／非通話検出部２１の検出結果に応じて非通話時に閉ループにおける利得余裕を推定するとともに、通話開始前に推定結果に基づく所要の損失量を信号経路に挿入させる点に特徴がある。
【００３５】
通話／非通話検出部２１は、拡声通話機の函体に設けられた通話スイッチや送話信号中の音声を検出して反応する音声検出スイッチ等のスイッチ手段２３により通話状態と非通話状態の判別を行うとともに、非通話状態において通話回路スイッチ２２を閉成し、閉ループを形成した状態で挿入損失量調整手段８により閉ループ系の利得余裕値が所望の値となるように挿入損失量の調整を行わせる。
【００３６】
すなわち、通常は非通話状態で通話回路スイッチ２２が開成されるのであるが、本実施形態においては非通話時に通話回路スイッチ２２を閉成して挿入損失量の調整を行うようにしたため、非通話時においても閉ループ系の利得余裕値を所望の値とすることができ、その結果、通話開始直後から安定した通話を実現することができる。
【００３７】
【発明の効果】
請求項１の発明は、集音した音を送話側の音声信号として出力するマイクロホンと、マイクロホンからの音声信号を増幅する第１の増幅手段と、受話側の音声信号に応じて鳴動するスピーカと、スピーカへ出力される音声信号を増幅する第２の増幅手段と、送話側及び受話側と外部の通話回線との間で２−４線変換を行う２−４線変換手段と、送話側及び受話側の少なくとも一方の信号経路に所定量の損失を挿入する損失挿入手段と、損失挿入手段から挿入される損失量を可変制御する制御手段と、信号経路に送出したサンプル信号に対する応答信号に応じて、マイクロホン及びスピーカを通じて形成される閉ループにおける利得余裕を推定するとともに制御手段を介して推定結果に基づく損失量の調整を行う挿入損失量調整手段とを備え、挿入損失量調整手段が、信号経路にサンプル信号を送出するサンプル信号発生器と、音声信号にサンプル信号を加算する加算器と、サンプル信号に対する応答信号の包絡線を検波する包絡線検波器と、包絡線検波器の出力信号に基づいて閉ループにおける利得余裕を推定する閉ループ利得余裕推定部と、閉ループ利得余裕推定部にて利得余裕の推定処理が行われる際に信号経路への挿入損失量を、所望の利得余裕値を得るために必要な値よりも大きな値に設定するとともに推定処理終了後に推定結果から求められる調整量を設定前の元の値に加えた値へ切り換える挿入損失量切換部とを具備して成り、閉ループ利得余裕推定部が、包絡線検波器の出力信号の微小時間における変位を求める微分器を有し、微分器で求めた包絡線検波器の出力信号の微小時間における変位に基づいて閉ループにおける利得余裕を推定して成るので、ハウリングが発生するまでの閉ループ系の利得余裕を所望の値に維持することができ、ハウリングを生じることなく安定した通話品質を維持することができるという効果がある。また、必要以上に挿入損失量を大きくすることがないため、双方向同時通話性能の実現可能性が高まるという効果がある。さらに、閉ループ系にサンプル信号を入力し、その応答信号の包絡線成分から閉ループ系の時定数に関する情報を抽出し、抽出された情報から閉ループ系の発振に対する利得余裕値を推定し、推定結果に基づいて送話側及び受話側への挿入損失量を制御することにより、随時閉ループ系の利得余裕値を所定値以上に維持することができ、しかも、応答信号の包絡線成分の絶対値ではなく、時間に対する傾きから推定処理を行うため、サンプル信号のレべル、信号送出時間や包絡線検波器の応答特性等の影響を除去することができるという効果がある。さらに、利得余裕値の推定処理を行っている間に閉ループ系の変動などがあった場合でもハウリングが生じるのを防ぐことができるという効果がある。
【００３８】
請求項２の発明は、サンプル信号発生器をインパルス信号発生器として成るので、利得余裕値と密接な関係にあるインパルス応答特性から閉ループ系の利得余裕値を推定することができるという効果がある。
請求項３の発明は、サンプル信号発生器をバーストノイズ発生器として成ることを特徴とし、バーストノイズを用いた場合にも、信号送出時間が閉ループ系の遅延時間に対して十分に短ければ、その応答波形の包絡線成分より閉ループ系の時定数に関する情報を抽出することができる。しかもバーストノイズの信号送出時間を適切な値とすることにより通話中における違和感をなくし、サンプル信号の送出による通話品質の劣化を抑えることができる。
【００３９】
請求項４の発明は、挿入損失量調整手段が通話中の無音区間を検出する無音検出器を具備し、無音検出器において無音区間が検出されたときにサンプル信号発生器からサンプル信号を発生させて成るので、観測される応答信号が音声信号に重畳されることがなく、精度良く推定処理を行うことができるという効果がある。
【００４０】
請求項５の発明は、通話の開始及び終了を検出する検出手段を備え、挿入損失量調整手段が、検出手段の検出結果に応じて非通話時に閉ループにおける利得余裕を推定するとともに通話開始前に推定結果に基づく所要の損失量を信号経路に挿入させて成るので、非通話時においても閉ループ系の利得余裕値を所望の値とすることができ、その結果、通話開始直後から安定した通話を実現することができるという効果がある。
【図面の簡単な説明】
【図１】本発明の参考例を示すブロック図である。
【図２】同上における挿入損失量調整手段を示すブロック図である。
【図３】同上の動作を説明するための説明図である。
【図４】実施形態１における挿入損失量調整手段を示すブロック図である。
【図５】実施形態２における挿入損失量調整手段を示すブロック図である。
【図６】実施形態３における挿入損失量調整手段を示すブロック図である。
【図７】同上を説明するための説明図である。
【図８】同上を説明するための説明図である。
【図９】同上を説明するための説明図である。
【図１０】同上を説明するための説明図である。
【図１１】実施形態４を示すブロック図である。
【図１２】従来例を示すブロック図である。
【符号の説明】
１マイクロホン
２第１の増幅器
３スピーカ
４第２の増幅器
５ハイブリッド回路
６₁，６₂減衰器
７制御器
８挿入損失量調整手段[0001]
BACKGROUND OF THE INVENTION
  The present invention relates to a loudspeaker used in homes, buildings, factories, and the like.
[0002]
[Prior art]
  Conventionally, in a loudspeaker such as an interphone, a telephone, or a PHS, a closed loop is formed by acoustic feedback from a speaker to a microphone and impedance mismatch in a hybrid circuit (2-4 wire conversion circuit), and the gain of the amplifier is large. Since howling occurs when the closed-loop gain becomes 1 or more for reasons such as too much, suppression of howling has become an indispensable issue in securing call quality.
[0003]
  Therefore, conventionally, a howling suppression method has been used in which a closed loop gain is suppressed by inserting a predetermined amount of loss into a received signal or a transmitted signal in accordance with the level of the transmitted / received signal. In addition, there is another method that uses an echo canceller, but it tends to be unstable when the echo path changes suddenly due to a transient state where the coefficient of the adaptive filter in the echo canceller has not converged or system fluctuations. Often used in conjunction with insertion loss.
[0004]
  In any of the above cases, if the amount of insertion loss is increased too much, the quality of the call will be degraded during a call (sound will be interrupted). Is desirable. On the other hand, when the acoustic reflection coefficient increases abruptly, such as when a hand or face is brought close to the front of the loudspeaker, the closed loop gain becomes 1 or more, and as a result, howling may occur. Conventionally, with respect to such a problem, a howling detector 24 is provided in a closed loop as shown in FIG. 12, and when howling is detected, an attenuator 6 provided in the signal path on the transmitting side and the receiving side.₁, 6₂Is controlled by the controller 7 to set the insertion loss amount to a larger value than usual. In the figure, 1 is a microphone, 3 is a speaker, 2 and 4 are amplifiers, and 5 is a hybrid (2-4 wire conversion) circuit.
[0005]
[Problems to be solved by the invention]
  However, in the above conventional example, since howling is detected by the howling detector 24 and then processing is performed to suppress howling due to insertion loss, the amount of insertion loss largely changes before and after howling occurs, which in turn causes deterioration in speech quality. There was a problem. In addition, since the amount of loss to be inserted at the time of howling detection cannot be controlled to a necessary minimum value in principle and becomes a considerably large value, it has been difficult to realize bidirectional simultaneous call performance.
[0006]
  The present invention has been made in view of the above-described problems, and an object of the present invention is to realize a stable telephone call by suppressing howling and simultaneously reduce bidirectional insertion by minimizing the amount of insertion loss. An object of the present invention is to provide a loudspeaker capable of increasing the feasibility of a call.
[0007]
[Means for Solving the Problems]
  In order to achieve the above object, the first aspect of the present invention provides a microphone that outputs the collected sound as an audio signal on the transmitting side, a first amplifying means that amplifies the audio signal from the microphone, A speaker that rings according to an audio signal, a second amplifying unit that amplifies the audio signal output to the speaker, and 2-4 line conversion between the transmission side, the reception side, and an external communication line 2 -Wire conversion means, loss insertion means for inserting a predetermined amount of loss into at least one signal path on the transmission side and reception side, control means for variably controlling the amount of loss inserted from the loss insertion means, and signal Insertion that estimates the gain margin in the closed loop formed through the microphone and the speaker according to the response signal to the sample signal sent to the path and adjusts the loss amount based on the estimation result via the control means. Loss amount adjusting means, and the insertion loss amount adjusting means includes a sample signal generator for sending a sample signal to the signal path, an adder for adding the sample signal to the audio signal, and an envelope of a response signal for the sample signal. An envelope detector for detection, and a closed loop gain margin estimation unit for estimating a gain margin in the closed loop based on an output signal of the envelope detector;When the gain margin estimation process is performed in the closed loop gain margin estimation unit, the amount of insertion loss to the signal path is set to a value larger than a value necessary for obtaining a desired gain margin value, and the estimation process ends. An insertion loss amount switching unit that switches the adjustment amount obtained from the estimation result to a value added to the original value before setting;The closed-loop gain margin estimation unit has a differentiator that obtains a displacement in a minute time of the output signal of the envelope detector, and the displacement in a minute time of the output signal of the envelope detector obtained by the differentiator The gain margin in the closed loop system can be maintained at a desired value until howling occurs, and stable call quality is maintained without generating howling. can do. In addition, since the amount of insertion loss is not increased more than necessary, the possibility of realizing bidirectional simultaneous call performance is increased. Furthermore, a sample signal is input to the closed loop system, information on the time constant of the closed loop system is extracted from the envelope component of the response signal, a gain margin value for the oscillation of the closed loop system is estimated from the extracted information, and the estimation result is Based on this, by controlling the insertion loss amounts on the transmitting side and the receiving side, the gain margin value of the closed loop system can be maintained at a predetermined value or more as needed. Moreover, since the estimation process is performed from the gradient with respect to time rather than the absolute value of the envelope component of the response signal, it is possible to eliminate the influence of the level of the sample signal, the signal transmission time, the response characteristic of the envelope detector, and the like. it can.Further, howling can be prevented even when there is a change in the closed loop system during the process of estimating the gain margin value.
[0008]
  The invention of claim 2 is characterized in that, in the invention of claim 1, the sample signal generator is an impulse signal generator, and the gain margin value of the closed loop system is determined from the impulse response characteristic closely related to the gain margin value. Can be estimated.
  The invention of claim 3 is characterized in that, in the invention of claim 1, the sample signal generator is a burst noise generator, and even when burst noise is used, the signal transmission time is less than the delay time of the closed loop system. If it is sufficiently short, information on the time constant of the closed loop system can be extracted from the envelope component of the response waveform. In addition, by setting the burst noise signal transmission time to an appropriate value, it is possible to eliminate a sense of incongruity during a call and to suppress deterioration in call quality due to the transmission of sample signals.
[0009]
  The invention according to claim 4 is the invention according to any one of claims 1 to 3, wherein the insertion loss amount adjusting means includes a silence detector that detects a silence interval during a call, and the silence detector detects a silence interval. It is characterized in that the sample signal is sometimes generated from the sample signal generator, and the observed response signal is not superimposed on the audio signal, and the estimation process can be performed with high accuracy..
[0010]
  Claim5The invention of claim 1 to claim 14In any of the inventions, a detecting means for detecting the start and end of a call is provided, and the insertion loss amount adjusting means estimates the gain margin in the closed loop during a non-call according to the detection result of the detecting means and before the call starts. The required loss amount based on the estimation result is inserted into the signal path, and the gain margin value of the closed loop system can be set to a desired value even during non-calling. Can be realized.
[0011]
DETAILED DESCRIPTION OF THE INVENTION
  Figure1 of the present inventionReference exampleA microphone 1 that outputs the collected sound as a voice signal on the transmission side (hereinafter referred to as a transmission signal), and a first amplifier 2 that amplifies the transmission signal from the microphone 1. , A speaker 3 that rings in response to an audio signal on the receiving side (hereinafter referred to as a receiving signal), a second amplifier 4 that amplifies the receiving signal output to the speaker 3, and a transmitting side, a receiving side, and an external Hybrid circuit 5 as 2-4 line conversion means for performing 2-4 line conversion with a speech line, and an attenuator (attenuator) as loss insertion means for inserting a predetermined amount of loss into the signal paths on the transmission side and

reception side

6₁, 6₂And each

attenuator

6₁, 6₂The controller 7 that variably controls the amount of loss inserted from the signal and the gain margin in the closed loop system formed through the microphone 1 and the speaker 3 according to the response signal to the sample signal sent to the signal path and the controller 7 And an insertion loss amount adjusting means 8 for adjusting the loss amount based on the estimation result. BookReference exampleThen, the insertion loss amount adjusting means 8 is provided on the signal path on the receiving side, but it is of course possible to provide it on the signal path on the transmitting side.
[0012]
  On the other hand, FIG. 2 is a block diagram showing a specific structure of the insertion loss amount adjusting means 8. The impulse signal generator 9 for sending an impulse signal as a sample signal to the signal path on the reception side, and the impulse signal as the reception signal. , An envelope detector 11 for detecting the envelope of the response signal with respect to the impulse signal, and a closed loop gain margin estimator for estimating the gain margin in the closed loop system based on the output signal of the envelope detector 11 12. Here, the closed loop gain margin estimation unit 12 compares the output signal level of the envelope detector 11 with a previously obtained threshold level as described later, and calculates the gain margin value of the closed loop system based on the comparison result. And a determination unit 14 for estimation. The impulse signal used as the sample signal may be a single pulse signal with a sufficiently short pulse width. The envelope detector 11 can be composed of a synthesis circuit of a rectifier circuit and a low-pass filter circuit, a digital circuit such as a cyclic low-pass filter and a leak integrator, or a signal processing means such as a DSP (Digital Signal Processor).
[0013]
  As shown in FIG.Reference exampleIn the second amplifier 4 → speaker 3 → microphone 1 → first amplifier 2 → attenuator 6₁→ Hybrid circuit 5 → Attenuator 6₂→ Insertion loss adjustment means 8 → The second amplifier 4 forms a closed loop. Here, the transfer function in each part is defined as follows.
  S: Electromechanical conversion characteristics of speaker 3
  G: Sound transmission characteristics from the speaker 3 to the microphone 1
  M: Acoustoelectric conversion characteristics of microphone 1
  Kr: amplification characteristic of the second amplifier 4
  Kx: amplification characteristic of the first amplifier 2
  Ar: Attenuator 6 on the receiver side₂Damping characteristics
  Ax: Transmitter side attenuator 6₁Damping characteristics
  Γ: reflection transfer function in the hybrid circuit 5
Also, the impulse signal output from the impulse signal generator 9 is P, the far-end talker voice input signal transmitted from the external speech line is Y, the near-end talker voice signal collected by the microphone 1 and ambient noise. And X is the output signal (response signal) Q of the adder 10, which is one of the components of the insertion loss amount adjusting means 8, is represented by the following equation.
[0014]
[Formula 1]

[0015]
  In the above equation, L (s) represents a one-round transfer function in the closed loop system, and s represents a Laplace variable.
  Here, the stability of the closed-loop system can be discriminated by the one-round transfer function L (s). That is, the r component (= | L () at all frequencies where the θ component (= ∠L (s)) of the circular transfer function L (s) in the polar coordinate system is ∠L (s) = 2nπ (n is an integer). If s) |) is | L (s) | <1, the closed loop system is stable, and if there is a frequency where | L (s) | ≧ 1, the closed loop system becomes unstable and oscillates at that frequency and howling occurs. Arise. Further, when the closed loop system is stable, the maximum value of the gain at all frequencies where ∠L (s) = 2nπ is expressed as L_MAXThen, the gain margin value of the closed loop system is 1 / L_MAXIt is represented by Therefore, a measure of the stability of the closed loop system can be evaluated by the closed loop gain margin value.
[0016]
  On the other hand, the closed-loop gain margin value is closely related to the impulse response characteristics of the closed-loop system. The larger the closed-loop gain margin value, the more the amplitude of the impulse response signal Q attenuates with time, and the smaller the closed-loop gain margin value, the more attenuated. Becomes moderate. Accordingly, the present invention observes the response signal Q when the closed loop gain margin estimation unit 12 applies the sample signal (impulse signal) P for estimating the closed loop gain margin value to the closed loop system, and envelopes the response signal Q. Information on the time constant of the closed-loop system is extracted from the line component to estimate the closed-loop gain margin value, and the closed-loop gain margin value becomes a desired value that does not cause howling based on the estimation result in the insertion loss adjustment unit 8. Adjust the amount of insertion loss to the closed loop system.
[0017]
  That is, since the time characteristic of the envelope component of the response signal Q obtained by the envelope detector 11 as described above has the property that the larger the closed loop gain margin value, the faster the attenuation and the smaller the attenuation, the slower the attenuation. A threshold level is obtained from the output data of the envelope detector 11 for various gain margin values learned in advance by the closed loop gain margin estimation unit 12, and the output signal level of the envelope detector 11 observed by the comparator 13. Is compared with the threshold level, the determination unit 14 can estimate the closed-loop gain margin based on the comparison result. Then, from the estimation result, a signal is transmitted to the controller 7 and the attenuator 6 is transmitted by the controller 7 in order to insert a loss amount necessary to make the closed loop gain margin value a value determined by the design specification.₁, 6₂The amount of attenuation is adjusted.
[0018]
  Book as aboveReference exampleAccording to the present invention, there is provided the insertion loss amount adjusting means 8 for estimating a gain margin value that does not cause howling in a closed loop system from the response signal to the impulse signal and adjusting the loss amount to be inserted into the signal path based on the estimation result. Therefore, even during a call, the amount of insertion loss can be controlled according to the stability of the closed-loop system, and the gain margin value can always be maintained at the value specified in the specification, maintaining stable call quality without causing howling. can do. Moreover, since it is not necessary to increase the amount of insertion loss more than necessary as compared with the conventional example, there is an advantage that the possibility of realizing bidirectional simultaneous call performance is increased.
[0019]
  BookReference exampleIn this example, the impulse signal generator 9 is used as the sample signal generator, but burst noise may be used as the sample signal by using a burst noise generator instead. In this case, the transient response of the closed loop system from the moment when the transmission is stopped from the state in which the signal is transmitted is observed, and information on the time constant of the closed loop system is extracted from the envelope component. Thus, even when burst noise is used for the sample signal, if the signal transmission time is sufficiently shorter than the delay time of the closed loop system, information on the time constant of the closed loop system is extracted from the envelope component of the response waveform. In addition, there is an advantage that by making the burst noise signal transmission time an appropriate value, it is possible to eliminate a sense of incongruity during a call and to suppress deterioration in call quality due to the transmission of a sample signal. However, as in the case of the impulse signal, it is necessary to obtain threshold levels for various gain margin values.
[0020]
  (Embodiment1)
  FIG. 4 shows an embodiment of the present invention.1It is a block diagram which shows the insertion loss amount adjustment means 8 in FIG. That is, the present embodiment has other configurations excluding the insertion loss amount adjusting means 8.Reference exampleTherefore, common parts are denoted by the same reference numerals, and description and partial illustration thereof are omitted.
[0021]
  In the insertion loss amount adjusting means 8 in the present embodiment, a differentiator 15 for obtaining a time derivative of the output signal of the envelope detector 11 is provided in the closed loop gain margin estimation unit 12. That is, using the property that the slope of the attenuation curve of the envelope component of the response signal output from the envelope detector 11 becomes steeper as the gain margin value is larger, and becomes gentler as the gain margin value is smaller, the differentiator 15 is used. The slope of the attenuation curve is obtained by the above, the threshold level for various gain margin values obtained in advance is compared with the comparator 13, and the closed loop gain margin value is estimated by the determination unit 14 based on the comparison result. It can be done. AfterReference exampleThe insertion loss amount is adjusted based on the estimation result in the same manner as described above.
[0022]
  As described above, according to the present embodiment, since the differentiator 15 for obtaining the time derivative of the response signal is provided in the closed loop gain margin estimation unit 12,Reference exampleSince the estimation process is performed from the gradient (time differential value) with respect to time, not the absolute value of the envelope component of the response signal, the level of the sample signal (impulse signal or burst noise, etc.), signal transmission time and envelope There is an advantage that the influence of the response characteristic of the line detector 11 can be removed.
[0023]
  (Embodiment2)
  FIG. 5 shows an embodiment of the present invention.2It is a block diagram which shows the insertion loss amount adjustment means 8 in FIG. That is, the present embodiment is different from the insertion loss amount adjusting means 8 in the other embodiments.1Therefore, common parts are denoted by the same reference numerals, and description and partial illustration thereof are omitted.
[0024]
  In the insertion loss amount adjusting means 8 in the present embodiment, a silence detector 16 that detects a silent section during a call from a received signal level (or may be a transmitted signal level), and a sample signal generator A switch 17 is provided between the (impulse signal generator 9 or burst noise generator) and the adder 10 and is turned on only when a silence interval is detected by the silence detector 16. A sample signal (impulse signal or burst noise) is input to the closed loop system only in the middle silent section, and the gain margin value estimation process is performed by the closed loop gain margin estimation unit 12.
[0025]
  For example, consider a case where a sample signal such as an impulse signal is superimposed on the reception signal collected by the microphone 1. The above formula 1 is simplified and expressed as the following formula.
  Q (s) = H₁(s) P (s) + H₂(s) Y (s) + H_Three(s) X (s) ... Equation 2
Where H₁(s) -H_ThreeSince the denominator of (s) is common, if one of these can be confirmed to be stable, all transfer functions are stable. Furthermore, when the above equation is expressed in the time domain, the following equation is obtained.
[0026]
  q (t) = h₁(t) * p (t) + h₂(t) * y (t) + h_Three(t) * x (t) ... Equation 3
However, H₁(s) = L [h₁(t)], H₂(s) = L [h₂(t)], H_Three(s) = L [L [h_Three(t)], L [] represents Laplace transform, and * represents convolution. Especially when the sample signal p (t) is an impulse signal δ (t), h₁(t) * δ (t) = h₁Since (t), the above formula is expressed as the following formula.
[0027]
  q (t) = h₁(t) + h₂(t) * y (t) + h_Three(t) * x (t) ... Equation 4
Accordingly, when both the far-end speaker voice signal y (t) and the near-end talker voice signal x (t) are zero, the impulse response h is obtained from the output signal of the adder 10 as shown in FIG.₁(t) can be observed directly. However, when these far-end speaker audio signal y (t) and near-end speaker audio signal x (t) are superimposed on the impulse signal δ (t) and propagated in the closed loop, the second term on the right side of Equation 4 is used. And the third term becomes a noise component, and the estimation process may be difficult.
[0028]
  Therefore, in the present embodiment, the silence detector 16 detects a silence interval during a call, and an estimation sample signal (impulse signal or burst noise) is detected in the silence interval as a sample signal generator (impulse signal generator 9 or burst). By generating from the noise generator, the observed response signal is not superimposed on the audio signal, and the estimation process can be performed with high accuracy.
[0029]
  (Embodiment3)
  FIG. 6 shows an embodiment of the present invention.3It is a block diagram which shows the insertion loss amount adjustment means 8 in FIG. That is, the present embodiment is different from the insertion loss amount adjusting means 8 in the other embodiments.1Therefore, common parts are denoted by the same reference numerals, and description and partial illustration thereof are omitted.
[0030]
  The insertion loss amount adjusting means 8 in the present embodiment sets the insertion loss amount to the signal path to a relatively large value and performs estimation processing when the gain margin estimation processing is performed by the closed loop gain margin estimation unit 12. An insertion loss amount switching unit 18 is provided for switching the adjustment amount obtained from the estimation result after completion to a value obtained by adding the original value before setting. Further, the closed loop gain margin estimation unit 12 considers the loss amount added and inserted by the insertion loss amount switching unit 18 during the estimation process, and subtracts the added loss amount to estimate the gain margin value of the closed loop system. . Then, after the estimation process is completed, the insertion loss amount switching unit 18 adjusts the insertion loss amount to a value necessary for obtaining a desired gain margin value.
[0031]
  For example, when a person approaches the vicinity of the microphone 1 or the speaker 3 as shown in FIG. 7, the acoustic transfer characteristic G from the speaker 3 to the microphone 1 fluctuates or is conventionally known in a closed loop system as shown in FIG. When the echo canceller 19 having the following structure is inserted into the closed loop system, the coefficient of the adaptive filter 19a that has converged until then is disturbed by the input of the sample signal (impulse signal or burst noise), and as a result, the closed loop When the characteristics of the system fluctuate and the amount of insertion loss into the closed loop system is small, the above fluctuation abruptly causes the closed loop system to become unstable and howling may occur. However, in the present embodiment, the insertion loss amount switching unit 18 sets the insertion loss amount to the signal path to a relatively large value when the gain margin estimation process is performed by the closed loop gain margin estimation unit 12. Therefore, occurrence of the above howling can be prevented.
[0032]
  In the configuration in which the AGC (automatic gain control) circuit 20 as shown in FIGS. 9 and 10 is inserted in the closed loop system, a far-end speaker voice signal Y having a large level is transmitted during the estimation process. In addition, the gain adjustment function in the AGC circuit 20 may change the gain of the closed loop system, and the gain margin value may not be estimated correctly. However, in the present embodiment, a sufficiently large loss amount is inserted when the estimation process is performed. Therefore, the level of the far-end speaker voice signal Y is sufficiently reduced, and the gain control of the AGC circuit is performed during the estimation process. This can be prevented.
[0033]
  As described above, according to the present embodiment, when the gain margin estimation process is performed, the amount of insertion loss to the signal path is set to a relatively large value, and the adjustment amount obtained from the estimation result after the estimation process is completed. Since the insertion loss amount switching unit 18 for switching to a value added to the original value before the setting is provided in the insertion loss amount adjusting means 8, the closed loop system fluctuation as described above is performed during the process of estimating the gain margin value. There is an advantage that it is possible to prevent howling even if there is a problem.
[0034]
  (Embodiment4)
  FIG. 11 shows an embodiment of the present invention.4Is a block diagram showing the basic configuration of the first to first embodiments3Therefore, common parts are denoted by the same reference numerals and description thereof is omitted, and only the characteristic parts of the present embodiment will be described.
  In the present embodiment, a call / non-call detector 21 that detects the start and end of a call, a first amplifier 2 that amplifies a transmission signal, and an attenuator 6₁And a call circuit switch 22 inserted in a signal path between the call and non-call detecting unit 21 and turned on / off by the call / non-call detecting unit 21. Accordingly, there is a feature in that the gain margin in the closed loop is estimated during non-calling, and a required loss amount based on the estimation result is inserted into the signal path before the call starts.
[0035]
  The call / non-call detection unit 21 is switched between a call state and a non-call state by a switch 23 such as a call switch provided in a box of a loudspeaker or a voice detection switch that detects and reacts to voice in a transmission signal. In addition, the communication circuit switch 22 is closed in a non-calling state, and the insertion loss amount is adjusted by the insertion loss amount adjusting means 8 so that the gain margin value of the closed loop system becomes a desired value in a state where a closed loop is formed. To do.
[0036]
  That is, the call circuit switch 22 is normally opened in a non-call state, but in the present embodiment, the call loss is adjusted by closing the call circuit switch 22 during a non-call. Even at the time, the gain margin value of the closed loop system can be set to a desired value, and as a result, a stable call can be realized immediately after the start of the call.
[0037]
【The invention's effect】
  According to a first aspect of the present invention, there is provided a microphone that outputs the collected sound as an audio signal on the transmitting side, a first amplifying means that amplifies the audio signal from the microphone, and a speaker that rings according to the audio signal on the receiving side. A second amplifying means for amplifying the audio signal output to the speaker, a 2-4 line converting means for performing 2-4 line conversion between the transmitting side and the receiving side and an external communication line, Loss insertion means for inserting a predetermined amount of loss into at least one of the signal paths on the talk side and the reception side, control means for variably controlling the amount of loss inserted from the loss insertion means, and a response to the sample signal sent to the signal path An insertion loss amount adjusting means for estimating a gain margin in a closed loop formed through a microphone and a speaker according to a signal and adjusting a loss amount based on an estimation result via a control means; The input loss amount adjusting means includes a sample signal generator that sends a sample signal to the signal path, an adder that adds the sample signal to the audio signal, an envelope detector that detects an envelope of a response signal to the sample signal, A closed loop gain margin estimator for estimating the gain margin in the closed loop based on the output signal of the envelope detector;When the gain margin estimation process is performed in the closed loop gain margin estimation unit, the amount of insertion loss to the signal path is set to a value larger than a value necessary for obtaining a desired gain margin value, and the estimation process ends. An insertion loss amount switching unit that switches the adjustment amount obtained from the estimation result to a value added to the original value before setting;The closed-loop gain margin estimation unit has a differentiator that obtains a displacement in a minute time of the output signal of the envelope detector, and the displacement in a minute time of the output signal of the envelope detector obtained by the differentiator Therefore, the gain margin in the closed loop system until howling occurs can be maintained at a desired value, and stable call quality can be maintained without generating howling. There is an effect that can be done. Moreover, since the amount of insertion loss is not increased more than necessary, there is an effect that the possibility of realizing bidirectional simultaneous call performance is increased. Furthermore, a sample signal is input to the closed loop system, information on the time constant of the closed loop system is extracted from the envelope component of the response signal, a gain margin value for the oscillation of the closed loop system is estimated from the extracted information, and the estimation result is By controlling the insertion loss amount on the transmitter side and receiver side based on this, the gain margin value of the closed loop system can be maintained at a predetermined value or more at any time, and not the absolute value of the envelope component of the response signal Since the estimation process is performed from the inclination with respect to time, it is possible to eliminate the influence of the level of the sample signal, the signal transmission time, the response characteristic of the envelope detector, and the like.Furthermore, there is an effect that howling can be prevented even when there is a variation in the closed loop system during the process of estimating the gain margin value.
[0038]
  According to the second aspect of the present invention, since the sample signal generator is an impulse signal generator, the gain margin value of the closed loop system can be estimated from the impulse response characteristic that is closely related to the gain margin value.
  The invention of claim 3 is characterized in that the sample signal generator is a burst noise generator, and even when burst noise is used, if the signal transmission time is sufficiently shorter than the delay time of the closed loop system, Information on the time constant of the closed loop system can be extracted from the envelope component of the response waveform. In addition, by setting the burst noise signal transmission time to an appropriate value, it is possible to eliminate a sense of incongruity during a call and to suppress deterioration in call quality due to the transmission of sample signals.
[0039]
  According to a fourth aspect of the present invention, the insertion loss amount adjusting means includes a silence detector for detecting a silence interval during a call, and generates a sample signal from the sample signal generator when a silence interval is detected by the silence detector. Therefore, the observed response signal is not superimposed on the audio signal, and the estimation process can be performed with high accuracy..
[0040]
  Claim5The invention includes a detecting unit that detects the start and end of a call, and the insertion loss amount adjusting unit estimates a gain margin in a closed loop during a non-call according to a detection result of the detecting unit, and displays an estimation result before the call starts. Since the required loss amount based on the signal path is inserted into the signal path, the gain margin value of the closed loop system can be set to a desired value even during non-calling. As a result, a stable call can be realized immediately after the start of the call. There is an effect that can be.
[Brief description of the drawings]
[Figure 1]Reference example of the present inventionFIG.
FIG. 2 is a block diagram showing insertion loss amount adjusting means in the same as above.
FIG. 3 is an explanatory diagram for explaining the operation described above.
FIG. 4 Embodiment1It is a block diagram which shows the insertion loss amount adjustment means in.
FIG. 52It is a block diagram which shows the insertion loss amount adjustment means in.
FIG. 6 is an embodiment.3It is a block diagram which shows the insertion loss amount adjustment means in.
FIG. 7 is an explanatory diagram for explaining the above.
FIG. 8 is an explanatory diagram for explaining the above.
FIG. 9 is an explanatory diagram for explaining the above.
FIG. 10 is an explanatory diagram for explaining the above.
FIG. 11 Embodiment4FIG.
FIG. 12 is a block diagram showing a conventional example.
[Explanation of symbols]
  1 Microphone
  2 First amplifier
  3 Speaker
  4 Second amplifier
  5 Hybrid circuit
  6₁, 6₂Attenuator
  7 Controller
  8 Insertion loss adjustment means

Claims

A microphone that outputs the collected sound as an audio signal on the transmitting side, a first amplifying means that amplifies the audio signal from the microphone, a speaker that rings according to the audio signal on the receiving side, and an output to the speaker A second amplifying means for amplifying the audio signal; a 2-4 line converting means for performing 2-4 line conversion between the transmitting side and receiving side and an external communication line; and at least a transmitting side and a receiving side A loss insertion means for inserting a predetermined amount of loss into one of the signal paths; a control means for variably controlling the amount of loss inserted from the loss insertion means; and a microphone and a response signal for the sample signal sent to the signal path An insertion loss amount adjusting means for estimating a gain margin in a closed loop formed through a speaker and adjusting a loss amount based on an estimation result via a control means; A sample signal generator for sending the sample signal to the signal path, an adder for adding the sample signal to the audio signal, an envelope detector for detecting the envelope of the response signal to the sample signal, and the output of the envelope detector A closed loop gain margin estimation unit for estimating a gain margin in a closed loop based on a signal, and an insertion loss amount into a signal path when a gain margin estimation process is performed in the closed loop gain margin estimation unit, and a desired gain margin value An insertion loss amount switching unit that sets the value larger than the value necessary to obtain and switches the adjustment amount obtained from the estimation result after the estimation process to the value added to the original value before the setting , The closed-loop gain margin estimation unit has a differentiator that calculates the displacement of the envelope detector output signal in a minute time, and the envelope detector output signal obtained by the differentiator in the minute time of the output signal. Speaker-phone call device characterized by comprising estimating the gain margin in a closed loop based on the displacement.

The loudspeaker as claimed in claim 1, wherein the sample signal generator is an impulse signal generator.

2. The loudspeaker as claimed in claim 1, wherein the sample signal generator is a burst noise generator.

The insertion loss amount adjusting means includes a silence detector for detecting a silence interval during a call, and a sample signal is generated from a sample signal generator when a silence interval is detected by the silence detector. The loudspeaker according to any one of claims 1 to 3.

A detecting means for detecting the start and end of a call is provided, and the insertion loss amount adjusting means estimates a gain margin in a closed loop during non-calling according to a detection result of the detecting means, and a required loss based on the estimation result before the start of the call 5. A loudspeaker as claimed in any one of claims 1 to 4, characterized in that a quantity is inserted into the signal path .