JP2004072178A

JP2004072178A - Apparatus for determining permissible packet loss rate and program therefor

Info

Publication number: JP2004072178A
Application number: JP2002224952A
Authority: JP
Inventors: Takaaki Kurita; 栗田　孝昭
Original assignee: Nippon Telegraph and Telephone Corp
Current assignee: Nippon Telegraph and Telephone Corp
Priority date: 2002-08-01
Filing date: 2002-08-01
Publication date: 2004-03-04
Anticipated expiration: 2022-08-01
Also published as: JP3860093B2

Abstract

<P>PROBLEM TO BE SOLVED: To provide an apparatus for determining a permissible packet loss rate that provides a deteriorated video image permissible rate A, and a packet loss rate AL for a long time viewing with respect to the permissible rate A without the need for a subject to actually evaluate a deteriorated video image for a long period of time. <P>SOLUTION: In the apparatus, deteriorated quality Q of a video image for a unit time (10 seconds) with respect to various packet loss rates L is subjectively evaluated to obtain a relational expression Q=C0+C1×ln(L) in advance, a deterioration occurrence time interval is obtained at which the deteriorated video image with each quality per unit time caused by viewing for long time is evaluated to be permissible with respect to various permissible rate A (percentage of approving subjects to all the subjects) to obtain a relative equation of T=D0+D1×Q in advance, the Q is then obtained (10) from the received rate L, and the T is obtained (20) from the Q and the received rate A, and AL=(L/T) ×U is calculated (30). <P>COPYRIGHT: (C)2004,JPO

Description

【０００１】
【発明の属する技術分野】
この発明は、映像信号及び／又は音声や音楽などの音響信号を伝送するサービス・装置の開発時の品質設計において、長時間映像及び／又は音響を視聴する場合の劣化映像及び／又は音響の許容率に対し許容されるパケット損失率を決定する装置及びそのプログラムに関する。
【０００２】
【従来の技術】
従来、映像信号、音声信号、音楽信号、映像信号と音声信号又は音楽信号（以下映像及び／又は音響信号と記す）を伝送するシステムでは、伝送路や伝送装置・端末における符号誤りやパケット到着時間揺らぎによってパケット損失が生じ、映像及び／又は音響品質が劣化する場合がある。従来、許容される映像及び／又は音響品質となるパケット損失率を求める手段として、例えば映像信号にパケット損失を生じさせた１０秒間の劣化映像を、被験者が主観的に評価する方法がある。また例えばＩＴＵ−Ｒ　ＢＴ．５００−１０勧告に基づいて映像の主観品質評価を行なう方法がある。この方法によれば、少なくとも４種類以上の映像シーンを用いて各種損失率の劣化映像を、最小１５名以上の被験者が評価することが勧告されており、評価条件数が膨大となる。そのため、上記勧告で定められた１０秒間を単位時間とする映像提示にならざるを得ず、テレビジョンやインターネット配信の数分から数時間の映像及び／又は音響に時間離散的にパケット損失が生じた場合の劣化映像及び／又は音響を、被験者に提示して品質評価を行なうことは著しく困難である。さらに、数分から数時間の長時間の劣化映像及び／又は音響の品質評価は、被験者に多大の労力をかけるため、著しく困難である。上記従来技術の許容パケット損失率決定方法は、１０秒という単位時間に生じる劣化映像及び／又は音響を評価しており、数分から数時間に劣化が時間離散的に生じた場合に視聴者が許容する劣化発生頻度の要因を考慮していないため、長時間視聴時の劣化映像及び／又は音響の許容率及び許容されるパケット損失率を決定することができないという問題点がある。
【０００３】
【発明が解決しようとする課題】
この発明の目的は、被験者が長時間の劣化映像及び／又は音響を実際に評価せずに、長時間視聴時の劣化映像及び／又は音響に対する許容率と対応する許容されるパケット損失率を決定する装置及びそのプログラムを提供することにある。
【０００４】
【課題を解決するための手段】
この発明の一面によれば、上記目的を達するために、単位時間長の劣化映像及び／又は音響品質とパケット損失率の関係、及び単位時間長の劣化映像及び／又は音響品質と許容される劣化発生時間間隔の関係を用い、単位時間のパケット損失率と長時間視聴時の劣化映像及び／又は音響に対する許容率を与えることで長時間視聴時に許容されるパケット損失率を決定する。
この発明の他面によれば単位時間のパケット損失率と劣化区間の発生時間間隔を与えることで長時間視聴時の劣化映像及び／又は音響に対する許容率と許容されるパケット損失率を決定する。
【０００５】
【発明の実施の形態】
以下にこの発明の実施形態を、実施例により説明する。またこの発明は、映像信号のみならず、音声信号、音楽信号、更に映像信号と音声信号又は／及び音楽信号の組み合せを長時間配信するシステムの設計に必要とする劣化許容率とパケット損失率の決定に適用することができるが、以下の説明では映像信号を代表して説明する。
実施例１
この発明の実施例１は、単位時間のパケット損失率Ｌと、視聴者の長時間視聴時の劣化映像に対する許容率Ａを指定することにより、許容されるパケット損失率ＡＬを決定する。
【０００６】
この発明の実施例１の機能構成を図１に示す。
劣化品質推定部１０は、端子１１からの単位時間のパケット損失率Ｌを入力することにより、単位時間の劣化映像の品質Ｑを出力する。劣化発生時間間隔推定部２０は劣化品質推定部１０よりの単位時間の劣化映像の品質Ｑと、端子１２からの劣化映像に対する許容率Ａを入力することにより、視聴者が許容できる劣化映像区間の発生時間間隔Ｔを出力する。許容パケット損失率算出部３０は、端子１１からの単位時間のパケット損失率Ｌと、劣化発生時間間隔推定部２０よりの劣化映像区間の発生時間間隔Ｔを入力することにより、許容されるパケット損失率ＡＬを端子３１に出力する。
【０００７】
劣化品質推定部１０は、単位時間内にパケット損失を発生させた場合のパケット損失率Ｌと劣化映像の品質Ｑの関係式Ｆ１を、主観評価と統計的手法により予め求めておき、入力された単位時間パケット損失率Ｌを関係式Ｆ１へ代入することで、単位時間の劣化映像の品質Ｑを推定する。関係式Ｆ１の導出例として、主観評価より求めたパケット損失率と劣化映像の品質の関係を図２に示す。この図２に示したデータは次のようにして求めたものである。実験網の両端にＶＴＲ及び映像パケット化装置を設置し、ＤＶ映像を伝送する系を構成し、パケット損失は網内に設置した品質エミュレータを用いランダム損失（損失個数を変えた５段階）を発生させ、評価用映像には１０秒間の標準動画像を８種類用い、映像品質の評価法は二重刺激連続品質尺度（ＤＳＣＱＳ）法を用い、各劣化映像の指示・評価においては、映像品質評価に加え、許容できる劣化発生頻度を求めるため、１０秒間の劣化映像区間が数十分に１回発生したと想定した時許容できるか否かを問い、その回答を求めた。また許容劣化発生頻度は有料／無料視聴により異なる可能性があるため両ケースを評価した。被験者は一般人２４名である。図２中の評価値群４１，４２，４３，４４，４５はそれぞれ１０秒当りの損失個数が１個、３個、５個、１０個の場合における２４人の評価値の各シーンごとの平均値であり、損失個数が１個、３個、５個については同じ損失率の映像を２パターン（１０秒中における損失個所が異なる各２つのパターン）評価した。同じ損失率でも評価映像の相違により品質劣化の程度が大きく異なる。
【０００８】
ＩＴＵ−Ｒ　ＢＴ．８００−２　勧告の標記法に従い、評価映像シーン中で最低値の品質値を最悪品質とし、７５％のシーン数（評価値群４１，４２，４３は各１２個、評価値群４４，４５は各６個）が満たす品質値を７５％品質とすると、これらは損失率の増加に伴い比例するという統計的性質を利用して一定の傾きの直線４６，４５を表わす一意の関係式で映像品質を、最悪品質あるいは７５％品質として表すことができる。
関係式Ｆ１の例１：映像品質Ｑを最悪品質とした場合、
Ｑ＝Ｃ０１＋Ｃ１１×ｌｎ（Ｌ）、Ｃ０１，Ｃ１１は定数
関係式Ｆ１の例２：映像品質Ｑを７５％品質とした場合、
Ｑ＝Ｃ０２＋Ｃ１２×ｌｎ（Ｌ）、Ｃ０２，Ｃ１２は定数。
この実施例では、Ｃ０１＝１２６．６７，Ｃ１１＝９．０６，
Ｃ０２＝１０５．４６，Ｃ１２＝８．３１である。
例えばパケット損失率Ｌが１０^−４（１Ｅ−４）の場合、劣化映像品質Ｑは３０となる。
【０００９】
以上より劣化品質推定部１０は、映像信号にパケット損失を生じさせた映像音声の主観品質評価と統計的性質により関数Ｑ＝Ｆ１（Ｌ）として予め定式化し、このＱ＝Ｆ１（Ｌ）に単位時間のパケット損失率Ｌを代入して単位時間の劣化映像の品質Ｑを演算して求める構成とされている。つまり定数項としてＣ０１又はＣ０２が、１次係数としてＣ１１又はＣ１２がレジスタに格納され、入力されたパケット損失率Ｌの対数値ｌｎ（Ｌ）を求めて、これを変数として一次関数Ｆ１が演算される。
劣化発生時間間隔推定部２０は、予め主観評価実験により求められた、単位時間の劣化映像の品質Ｑと、許容される劣化発生時間間隔Ｔと、劣化許容率Ａとの三つの相互に関連する主観品質値を蓄える。つまり先に述べた実験では、１０秒間の劣化映像区間が数十分間、例えば１時間に１回発生したと想定した時、その長時間映像が許容できるか否かを、各劣化品質と各種劣化発生時間間隔について被験者に問い合せ、その回答から各種許容率Ａについて、劣化映像品質Ｑと許容できる劣化発生時間間隔Ｔとの関係を求める。
劣化発生時間間隔推定部２０は許容率Ａと劣化品質推定部１０から出力された劣化品質Ｑを入力し、劣化発生時間間隔Ｔを出力する。このため、先に求めた、種々の品質値Ｑの単位時間の劣化映像がＴ時間に１回発生したとして許容できるか否かの主観評価結果から、関係式Ｔ＝Ｆ２（Ｑ，Ａ）を得る。劣化映像に対する許容率は、許容できるとした視聴者（被験者）の全被験者に対する人数比（％）としている。図３は、この実験結果より求めた、劣化映像に対する許容率Ａが９０％の場合における、有料視聴と無料視聴の２種類の場合の、劣化映像品質Ｑと許容できる劣化発生時間間隔Ｔとの関係を示している。
【００１０】
関係式Ｆ２の例１：有料視聴の場合
Ｔ（分）＝Ｄ０１＋Ｄ１１×Ｑ
関係式Ｆ２の例２：無料視聴の場合
Ｔ（分）＝Ｄ０２＋Ｄ１２×Ｑ
許容率９０％のこの実施例では、Ｄ０１＝−７．３３，Ｄ１１＝４．３７、
Ｄ０２＝−１３．２２，Ｄ１２＝３．７８
であり、直線４８と４９にそれぞれ対応する。有料視聴は無料視聴に比べ、許容できる劣化発生頻度が低い、これは視聴者の料金支払い意識が反映されているためと考えられる。この例において例えば劣化映像品質Ｑを３０とすると、無料視聴時は、劣化発生時間間隔Ｔはほぼ１００分となる。
【００１１】
つまり劣化発生時間間隔推定部２０には例えば、関数格納部２１が設けられ、関数格納部２１には各許容率Ａについて、関数Ｆ２の定数（Ｄ０１又はＤ０２）と係数（Ｄ１１又はＤ１２）が格納され、入力された許容率Ａについて、定数と係数が関数格納部２１から読み出され、これら定数と係数の一次関数Ｆ２に、入力された劣化品質Ｑが変数として代入され、演算部２２で劣化発生時間間隔Ｔが演算され、これが出力される。なお、関数格納部２１には有料視聴と無料視聴とを区別して、各許容率に対する定数と係数を格納しておいて、有料視聴と無料視聴とで区別して劣化時間間隔Ｔを出力するようにしてもよい。
許容パケット損失率算出部３０は、端子１１よりの単位時間におけるパケット損失率Ｌ及び劣化発生時間間隔推定部２０よりの劣化区間の発生時間間隔Ｔを用い、許容パケット損失率ＡＬを下記関係式から計算する。
関係式Ｆ４：ＡＬ＝Ｌ／（Ｕ／Ｔ）＝Ｌ×Ｔ／Ｕ
【００１２】
単位時間長Ｕは、劣化品質推定部１０で用いる関係式Ｆ１及び劣化発生時間間隔推定部２０で用いる関係式Ｆ２とＦ３を導出する際の単位時間長を用いる。この実施例では、単位時間長Ｕは１０秒である。損失率Ｌは、単位時間Ｕ（＝１０秒）での送出パケット数Ｎ_ＰＵで、単位時間Ｕ内でのパケット損失個数Ｎ_ＰＬを割った値Ｎ_ＰＬ／Ｎ_ＰＵである。

例えば、劣化発生時間間隔推定部２０より算出された劣化発生時間間隔Ｔが３０分（＝１８００秒）の場合の、許容パケット損失率ＡＬは、下記で計算する。この実施例では、許容パケット損失率ＡＬは、ネットワークのシステム設計において用いられることが多い、時間単位や日単位の長時間で算出する。
ＡＬを１時間単位（＝３６００秒）の観測時間で求めた場合、

上記例にあるようにＡＬの算出は、関係式Ｆ４で一意に求めることが可能であり、長時間視聴時の許容できるパケット損失率ＡＬを求めていることになる。
この図１に示した構成によれば、単位時間パケット損失率Ｌと、ユーザにより要求される劣化許容率Ａを入力して、Ｌから劣化品質Ｑを劣化品質推定部１０により求め、劣化発生時間間隔推定部２０で入力された許容率Ａと対応した関数を取り出し、これにＱを代入することにより、劣化発生時間間隔Ｔを求め、このＴとＬから、許容率Ａの場合にシステムに要求される長時間パケット損失率ＡＬが求められる。
【００１３】
実施例２
この発明の実施例２は、単位時間のパケット損失率Ｌと単位時間の劣化映像の発生時間間隔Ｔを指定することにより、劣化映像に対する許容率Ａと許容されるパケット損失率ＡＬを決定する。
この実施例２の機能構成を図４に示し、図１と対応する部分に同一参照番号を付けてある。単位時間の劣化品質推定部１０は、端子１１よりの単位時間のパケット損失率Ｌを入力することにより、単位時間の劣化映像の品質Ｑを出力するもので、実施例１のそれと同様に構成される。損失率算出部３０は、端子１１よりの単位時間のパケット損失率Ｌと、端子３２よりの劣化映像区間の発生時間間隔Ｔを入力することにより、許容されるパケット損失率ＡＬを出力するもので、その構成も実施例１のそれと同様に構成される。許容率推定部５０は端子３２よりの劣化発生時間間隔Ｔと劣化品質推定部１０よりの単位時間の劣化映像の品質Ｑとを入力して劣化映像に対する許容率Ａを端子５１に出力する。
【００１４】
許容率推定部５０は予め主観評価と統計的手法により求められた、単位時間の劣化映像の品質Ｑと、許容される劣化発生時間間隔Ｔと劣化映像に対する許容率Ａとの三つの相互の関連を示す主観品質値を蓄える。許容率Ａの導出例を示す。例えば実施例１で述べた主観評価実験における、１０秒間の劣化映像区間が数十分に１回発生したと想定した時のこれが許容できるか否かの回答に基づき、図５に示すように各種劣化映像品質Ｑについて、許容劣化発生時間間隔Ｔと、劣化映像に対する許容率Ａとの関係をプロットし、種々の品質値Ｑの単位時間の劣化映像がＴ時間に１回発生したとして許容できるか否かの主観評価からこれらの関係式Ｔ＝Ｆ３（Ｑ，Ｔ）を得る。劣化映像に対する許容率は、許容できると評価した視聴者の、全被試験者に対する人数比（％）としている。図５は、劣化映像の映像品質Ｑが２８の場合における、有料視聴と無料視聴の２種類の場合の、許容率Ａと劣化発生時間間隔Ｔの関係を示している。これら主観評価値からＡとＴの関係を表わす曲線５２（有料視聴時）と曲線５３（無料視聴時）の各式Ｆ３を求めることができる。
関係式Ｆ３の例１：有料視聴の場合
Ａ＝Ｄ０１＋Ｄ１１×Ｔ＋Ｄ２１×Ｔ×Ｔ
関係式Ｆ３の例２：無料視聴の場合
Ａ＝Ｄ０２＋Ｄ１２×Ｔ＋Ｄ２２×Ｔ×Ｔ
この実施例では、
Ｄ０１＝１９．３，Ｄ１１＝０．７５５，Ｄ２１＝−１．４０Ｅ−３
Ｄ０２＝１８．８，Ｄ１２＝１．４２，Ｄ２２＝−６．４０Ｅ−３
である。
【００１５】
関係式Ｆ３は、予め種々の映像品質Ｑに対し定める。Ｑ＝２８の場合、許容できる劣化発生時間間隔Ｔを約６０分とすると、有料視聴の場合、曲線５２の式より劣化映像に対する許容率Ａは約５７％となる。このＬとＴから約５７％の人は劣化映像を許容し、その時の長時間パケット損失率ＡＬは（Ｌ／Ｔ）×Ｕにより求まる。
許容率推定部５０は、関数格納部５４に、各品質Ｑごとに、関係式Ｆ３の定数（Ｄ０１又はＤ０２）、１次係数（Ｄ１１又はＤ１２）、２次係数（Ｄ２１又はＤ２２）が格納され入力されたＱに応じて関数格納部５４から、定数、１次係数、２次係数が取り出され、演算部５５でこれらによりなる２次関数Ｆ３にＴを代入して、その関数値が演算されてその結果Ａが出力される。
【００１６】
なお関数格納部５４には有料視聴に対するもの、無料視聴に対するものの両者を格納しておいてもよい。実施例１又は２により求めた各種劣化映像に対する許容率Ａと、その場合の長時間パケット損失率ＡＬとの関係を例えば図６に示すように作っておくとよい。図６では、実線５６，５８は無料視聴を想定したもの、破線５７，５９は有料視聴を想定したものであり、最悪品質に対する関係式を線５６，５７で、７５％品質に対するものを表わす。このグラフからシステム設計時に、ユーザの許容率Ａが予測されると、それと対応するパケット損失率ＡＬを推定することができる。
【００１７】
図１、図４に示した装置はコンピュータに機能させることもできる。この場合は、例えば図７に示すように各部はバス６０により接続され、入力部６１に単位時間パケット損失率Ｌと、許容率Ａが入力設定されると、ワークメモリとして動作する記憶部６２に必要に応じて格納され、劣化品質関数格納部６３に格納されている定数Ｃ０１又はＣ０２と係数（Ｃ１１又はＣ１２）が取り出され、関数Ｆ１にＬを代入した演算がなされて、劣化品質Ｑが求められ、更に入力されたＡにより劣化発生時間間隔関数格納部６４から、対応する定数と係数が取り出され、関数Ｆ２にＱを代入した演算がなされて、劣化発生時間間隔Ｔが計算され、更にＴとＬから（Ｌ／Ｔ）×Ｕが計算され、その結果ＡＬが出力部６７により表示され、記録され又は他の処理装置へ出力される。これらの処理がＣＰＵ（マイクロプロセッサ）６６によりプログラムメモリ６５に格納されている許容パケット損失率決定プログラムを実行することにより行われる。
【００１８】
許容パケット損失率決定プログラムはメモリ６５に予め、ＣＤ−ＲＯＭ、磁気ディスクなどの記録媒体からインストールし、又は通信回線を介してダウンロードしておく。実施例２の場合は入力部６１から、単位時間パケット損失率Ｌと劣化発生時間間隔Ｔが入力され、図７中に括弧書きで示すように図４中に示した関数格納部５４と同様構成の許容率関数格納部６４が用いられ、又は実施例２の説明で行った処理が実行されることになる。
【００１９】
【発明の効果】
上述のようにこの発明によれば、被験者が長時間の劣化映像及び／又は音響を実際に評価することなく、劣化映像及び／又は音響に対する許容率及びこれと対応した長時間視聴時に許容されるパケット損失率が得られる。
【図面の簡単な説明】
【図１】この発明の実施例１の機能構成例を示す図。
【図２】実施例において、劣化品質推定部１０で用いる関係式Ｆ１の導出を説明するためのグラフ。
【図３】実施例において、劣化発生時間間隔推定部２０で用いる関係式Ｆ２の導出例を説明するためのグラフ。
【図４】この発明の実施例２の機能構成例を示す図。
【図５】実施例２において、許容率推定部５０で用いる関係式Ｆ３の導出例を説明するためのグラフ。
【図６】求めた長時間視聴に許容されるパケット損失率と、劣化映像に対する許容率との関係例を示すグラフ。
【図７】この発明装置をコンピュータに機能させるための構成例を示す図。[0001]
TECHNICAL FIELD OF THE INVENTION
According to the present invention, in the quality design at the time of development of a service / device for transmitting a video signal and / or an audio signal such as audio or music, a deterioration video and / or audio allowance when viewing a video and / or audio for a long time is considered. The present invention relates to an apparatus and a program for determining an allowable packet loss rate for a rate.
[0002]
[Prior art]
2. Description of the Related Art Conventionally, in a system for transmitting a video signal, an audio signal, a music signal, a video signal and an audio signal or a music signal (hereinafter, referred to as a video and / or audio signal), a code error or a packet arrival time in a transmission path, a transmission device, or a terminal. The fluctuations may cause packet loss and degrade video and / or audio quality. 2. Description of the Related Art Conventionally, as a means for obtaining a packet loss rate that achieves acceptable video and / or audio quality, there is a method in which a subject subjectively evaluates, for example, a 10-second deteriorated video that causes packet loss in a video signal. Also, for example, ITU-R BT. There is a method of performing subjective quality evaluation of video based on the 500-10 recommendation. According to this method, it is recommended that at least 15 or more subjects evaluate a deteriorated video having various loss rates using at least four or more types of video scenes, and the number of evaluation conditions becomes enormous. For this reason, video presentation must be performed with a unit time of 10 seconds as defined in the above recommendation, and packet loss occurs discretely in video and / or audio for several minutes to several hours of television or Internet distribution. It is extremely difficult to present the deteriorated video and / or sound in such a case to a subject to perform quality evaluation. In addition, long-term degraded video and / or sound quality evaluations of minutes to hours are extremely difficult because they require a great deal of effort on the subject. The above-described prior art allowable packet loss rate determination method evaluates the deteriorated video and / or sound that occurs in a unit time of 10 seconds, and allows the viewer to accept the deterioration when the deterioration occurs discretely from several minutes to several hours. Since the factor of the frequency of occurrence of deterioration is not taken into consideration, there is a problem that it is not possible to determine the allowable rate of the deteriorated video and / or sound and the allowable packet loss rate during long viewing.
[0003]
[Problems to be solved by the invention]
SUMMARY OF THE INVENTION It is an object of the present invention to determine a permissible packet loss rate corresponding to a permissible rate of degraded video and / or audio during long viewing without a subject actually evaluating long-term degraded video and / or audio. And a program therefor.
[0004]
[Means for Solving the Problems]
According to one aspect of the present invention, in order to achieve the above object, a relationship between a deteriorated video and / or sound quality of a unit time length and a packet loss rate, and a deteriorated video and / or sound quality of a unit time length and an allowable deterioration. The packet loss rate permissible during long viewing is determined by giving the packet loss rate per unit time and the permissible rate for degraded video and / or audio during long viewing using the relationship between the occurrence time intervals.
According to another aspect of the present invention, the permissible rate and permissible packet loss rate for deteriorated video and / or audio during long viewing are determined by giving the packet loss rate per unit time and the occurrence time interval of the degraded section.
[0005]
BEST MODE FOR CARRYING OUT THE INVENTION
Hereinafter, embodiments of the present invention will be described with reference to examples. In addition, the present invention provides a system which distributes not only a video signal but also an audio signal, a music signal, and a combination of a video signal and an audio signal and / or a music signal for a long time. Although it can be applied to the determination, in the following description, the video signal will be described as a representative.
Example 1
In the first embodiment of the present invention, an allowable packet loss rate AL is determined by designating a packet loss rate L per unit time and an allowable rate A for a deteriorated image when a viewer views for a long time.
[0006]
FIG. 1 shows a functional configuration of the first embodiment of the present invention.
The degradation quality estimation unit 10 outputs the quality Q of the degradation video per unit time by inputting the packet loss rate L per unit time from the terminal 11. The degradation occurrence time interval estimation unit 20 inputs the quality Q of the degradation video per unit time from the degradation quality estimation unit 10 and the allowable ratio A for the degradation video from the terminal 12, and thereby the degradation video section that the viewer can tolerate. Output the occurrence time interval T. The permissible packet loss rate calculation unit 30 inputs the packet loss rate L per unit time from the terminal 11 and the occurrence time interval T of the deteriorated video section from the deterioration occurrence time interval estimation unit 20, thereby allowing the permissible packet loss. The rate AL is output to the terminal 31.
[0007]
The deterioration quality estimating unit 10 previously obtains a relational expression F1 between the packet loss rate L and the quality Q of the deteriorated video when a packet loss occurs within a unit time by a subjective evaluation and a statistical method, and is input. By substituting the unit time packet loss rate L into the relational expression F1, the quality Q of the deteriorated video per unit time is estimated. As an example of deriving the relational expression F1, FIG. 2 shows the relationship between the packet loss rate obtained from the subjective evaluation and the quality of the deteriorated video. The data shown in FIG. 2 is obtained as follows. VTRs and video packetizers are installed at both ends of the experimental network, and a system for transmitting DV video is constructed. Packet loss generates random losses (5 stages with different numbers of losses) using a quality emulator installed in the network Eight kinds of standard moving images of 10 seconds are used for the evaluation video, and the image quality is evaluated using the double stimulus continuous quality scale (DSCQS) method. In addition, in order to obtain an allowable deterioration occurrence frequency, it was asked whether or not it was allowable when it was assumed that the deterioration video section of 10 seconds occurred once several tens of minutes, and the answer was obtained. In addition, both cases were evaluated because the allowable deterioration occurrence frequency may differ depending on paid / free viewing. The subjects were 24 ordinary people. The

evaluation value groups

41, 42, 43, 44, and 45 in FIG. 2 are the averages of the evaluation values of 24 scenes for each scene when the number of losses per 10 seconds is 1, 3, 5, and 10, respectively. With respect to the loss numbers of 1, 3, and 5, two images having the same loss rate were evaluated (two patterns having different loss locations in 10 seconds). Even at the same loss rate, the degree of quality deterioration greatly differs due to the difference in the evaluation video.
[0008]
ITU-R BT. According to the notation method of the 800-2 recommendation, the lowest quality value in the evaluation video scene is regarded as the worst quality, and the number of scenes of 75% (the

evaluation value groups

41, 42, and 43 are 12 each, and the

evaluation value groups

44 and 45 are the Assuming that the quality values satisfied by each of the six (6) are 75% quality, the video quality is represented by a unique relational expression expressing

straight lines

46 and 45 having a constant slope by utilizing the statistical property that these are proportional to the increase of the loss rate. Can be expressed as worst quality or 75% quality.
Example 1 of relational expression F1: When the video quality Q is the worst quality,
Q = C01 + C11 × ln (L), where C01 and C11 are constant relational expressions F1: Example 2: When the video quality Q is 75% quality,
Q = C02 + C12 × ln (L), where C02 and C12 are constants.
In this embodiment, C01 = 126.67, C11 = 9.06,
C02 = 105.46 and C12 = 8.31.
For example, when the packet loss rate L is 10 ⁻⁴ (1E−4), the degraded video quality Q is 30.
[0009]
As described above, the deterioration quality estimating unit 10 formulates a function Q = F1 (L) in advance based on the subjective quality evaluation and statistical properties of the video and audio that caused the packet loss in the video signal, and the unit is expressed as Q = F1 (L). The configuration is such that the quality Q of the deteriorated video per unit time is calculated by substituting the packet loss rate L per unit time. That is, C01 or C02 is stored in a register as a constant term, and C11 or C12 is stored in a register as a primary coefficient. A logarithmic value ln (L) of the input packet loss rate L is obtained, and a linear function F1 is calculated using this as a variable. You.
The deterioration occurrence time interval estimating unit 20 relates three interrelationships of the quality Q of the deterioration image per unit time, the allowable deterioration occurrence time interval T, and the deterioration allowable rate A, which are obtained in advance by a subjective evaluation experiment. Store subjective quality values. In other words, in the experiment described above, when it is assumed that a 10-second deteriorated video section occurs for several tens of minutes, for example, once an hour, it is determined whether or not the long-time video is permissible. The subject is inquired about the deterioration occurrence time interval, and from the answer, the relationship between the deterioration image quality Q and the allowable deterioration occurrence time interval T is obtained for various allowable rates A.
The deterioration occurrence time interval estimating unit 20 receives the allowable rate A and the deterioration quality Q output from the deterioration quality estimating unit 10, and outputs a deterioration occurrence time interval T. For this reason, the relational expression T = F2 (Q, A) is obtained from the subjective evaluation result obtained as to whether or not the previously determined deteriorated image of various quality values Q per unit time occurs once per T time. obtain. The permissible rate for the deteriorated video is defined as the ratio (%) of the number of viewers (subjects) to all the test subjects that is considered to be permissible. FIG. 3 shows the relationship between the degraded video quality Q and the allowable degradation occurrence time interval T in the two cases of pay viewing and free viewing when the permissible rate A for the degraded video obtained from the experimental results is 90%. Shows the relationship.
[0010]
Example of relational expression F2 1: In the case of pay viewing, T (minute) = D01 + D11 × Q
Example 2 of relational expression F2: For free viewing T (minutes) = D02 + D12 × Q
In this embodiment with a 90% tolerance, D01 = -7.33, D11 = 4.37,
D02 = -13.22, D12 = 3.78
And correspond to the

straight lines

48 and 49, respectively. Paid viewing has a lower permissible deterioration frequency than free viewing, which is considered to reflect the viewer's awareness of paying for the fee. In this example, assuming that the degraded video quality Q is 30, for example, during free viewing, the degradation occurrence time interval T is approximately 100 minutes.
[0011]
That is, the deterioration occurrence time interval estimating unit 20 is provided with, for example, a function storage unit 21, and the function storage unit 21 stores a constant (D01 or D02) and a coefficient (D11 or D12) of the function F2 for each allowable rate A. Then, for the input allowable rate A, constants and coefficients are read from the function storage unit 21, and the input deterioration quality Q is substituted as a variable into a linear function F2 of these constants and coefficients, and the deterioration An occurrence time interval T is calculated and output. It should be noted that the function storage unit 21 stores constants and coefficients for each permissible rate for distinguishing between pay viewing and free viewing, and outputs the deterioration time interval T separately for pay viewing and free viewing. You may.
The allowable packet loss rate calculation unit 30 uses the packet loss rate L per unit time from the terminal 11 and the occurrence time interval T of the degradation section from the degradation occurrence time interval estimation unit 20 to calculate the allowable packet loss rate AL from the following relational expression. calculate.
Relational expression F4: AL = L / (U / T) = L × T / U
[0012]
As the unit time length U, a unit time length for deriving the relational expression F1 used in the deterioration quality estimation unit 10 and the relational expressions F2 and F3 used in the deterioration occurrence time interval estimation unit 20 is used. In this embodiment, the unit time length U is 10 seconds. The loss rate L is a value N _PL / N _PU obtained by dividing the number N _PU of transmitted packets in the unit time U (= 10 seconds) by the number N _PL of lost packets in the unit time U.

For example, when the deterioration occurrence time interval T calculated by the deterioration occurrence time interval estimation unit 20 is 30 minutes (= 1800 seconds), the allowable packet loss rate AL is calculated as follows. In this embodiment, the allowable packet loss rate AL is calculated over a long period of time, such as hours or days, often used in network system design.
When AL is obtained in observation time of one hour unit (= 3600 seconds),

As in the above example, the calculation of AL can be uniquely obtained by the relational expression F4, which means that the allowable packet loss rate AL during long viewing is obtained.
According to the configuration shown in FIG. 1, a unit time packet loss rate L and a degradation tolerance rate A required by a user are input, and a degradation quality Q is obtained from L by the degradation quality estimating unit 10, and the degradation occurrence time is calculated. A function corresponding to the allowable rate A input by the interval estimating unit 20 is extracted, and Q is substituted into the function to determine a deterioration occurrence time interval T. From the T and L, a request is issued to the system in the case of the allowable rate A. The long-term packet loss rate AL is calculated.
[0013]
Example 2
In the second embodiment of the present invention, the permissible rate A and the permissible packet loss rate AL for the deteriorated video are determined by specifying the packet loss rate L per unit time and the generation time interval T of the deteriorated video per unit time.
FIG. 4 shows a functional configuration of the second embodiment, and portions corresponding to those in FIG. 1 are denoted by the same reference numerals. The unit-time deterioration quality estimating unit 10 outputs the unit-time deteriorated video quality Q by inputting the unit-time packet loss rate L from the terminal 11, and has the same configuration as that of the first embodiment. You. The loss rate calculation unit 30 outputs an allowable packet loss rate AL by inputting a packet loss rate L per unit time from the terminal 11 and a generation time interval T of a deteriorated video section from the terminal 32. The configuration is the same as that of the first embodiment. The allowable rate estimating unit 50 receives the deterioration occurrence time interval T from the terminal 32 and the quality Q of the deteriorated image per unit time from the deterioration quality estimating unit 10 and outputs the allowable rate A for the deteriorated image to the terminal 51.
[0014]
The permissible rate estimating unit 50 has three mutual relations between the quality Q of the degraded video per unit time, the permissible degradation occurrence time interval T, and the permissible rate A for the degraded video, which are obtained in advance by a subjective evaluation and a statistical method. Is stored. An example of deriving the allowable ratio A will be described. For example, based on the subjective evaluation experiment described in the first embodiment, based on the answer as to whether or not this is permissible when it is assumed that the degraded video section of 10 seconds occurs once several ten minutes, as shown in FIG. For the deteriorated video quality Q, the relationship between the allowable deterioration occurrence time interval T and the permissible rate A for the deteriorated video is plotted, and it is acceptable if the deteriorated video of various quality values Q in the unit time occurs once in the T time. These relational expressions T = F3 (Q, T) are obtained from the subjective evaluation of whether or not to do so. The permissible rate for the deteriorated video is the ratio (%) of the number of viewers who are evaluated as permissible to all the test subjects. FIG. 5 shows the relationship between the allowable ratio A and the degradation occurrence time interval T in the case of two types of viewing, pay viewing and free viewing, when the video quality Q of the deteriorated video is 28. From these subjective evaluation values, the formulas F3 of the curve 52 (at the time of pay viewing) and the curve 53 (at the time of free viewing) representing the relationship between A and T can be obtained.
Example of relational expression F3 1: In the case of paid viewing A = D01 + D11 × T + D21 × T × T
Example 2 of relational expression F3: For free viewing A = D02 + D12 × T + D22 × T × T
In this example,
D01 = 19.3, D11 = 0.755, D21 = -1.40E-3
D02 = 18.8, D12 = 1.42, D22 = -6.40E-3
It is.
[0015]
The relational expression F3 is determined in advance for various video qualities Q. In the case of Q = 28, assuming that the permissible deterioration occurrence time interval T is about 60 minutes, the permissible rate A for the deteriorated video is about 57% from the equation of the curve 52 for pay viewing. Approximately 57% of the persons from L and T allow a deteriorated video, and the long-term packet loss rate AL at that time is obtained by (L / T) × U.
The allowance estimating unit 50 stores the constant (D01 or D02), the first order coefficient (D11 or D12), and the second order coefficient (D21 or D22) of the relational expression F3 for each quality Q in the function storage unit 54. Constants, first-order coefficients, and second-order coefficients are extracted from the function storage unit 54 according to the input Q, and T is substituted into the quadratic function F3 formed by the arithmetic unit 55 to calculate the function value. As a result, A is output.
[0016]
Note that the function storage unit 54 may store both for paid viewing and free viewing. The relationship between the permissible ratio A for various types of deteriorated video obtained in the first or second embodiment and the long-term packet loss ratio AL in that case may be made, for example, as shown in FIG. In FIG. 6, the

solid lines

56 and 58 are for free viewing, and the

broken lines

57 and 59 are for pay viewing. The relational expressions for the worst quality are shown by

lines

56 and 57, and for 75% quality. When the user's allowable rate A is predicted from the graph at the time of system design, the corresponding packet loss rate AL can be estimated.
[0017]
The devices shown in FIGS. 1 and 4 can be caused to function by a computer. In this case, for example, as shown in FIG. 7, each unit is connected by a bus 60, and when a unit time packet loss rate L and an allowable rate A are input to an input unit 61, the storage unit 62 that operates as a work memory is input to the storage unit 62. The constant C01 or C02 and the coefficient (C11 or C12) stored as necessary and stored in the deterioration quality function storage unit 63 are extracted, and an operation in which L is substituted for the function F1 is performed to obtain the deterioration quality Q. Then, the corresponding constants and coefficients are extracted from the deterioration occurrence time interval function storage unit 64 by the input A, and the operation is performed by substituting Q for the function F2, and the deterioration occurrence time interval T is calculated. Then, (L / T) × U is calculated from L and L. As a result, the AL is displayed on the output unit 67, recorded, or output to another processing device. These processes are performed by the CPU (microprocessor) 66 executing the allowable packet loss rate determination program stored in the program memory 65.
[0018]
The allowable packet loss rate determination program is installed in the memory 65 in advance from a recording medium such as a CD-ROM or a magnetic disk, or is downloaded via a communication line. In the case of the second embodiment, the unit time packet loss rate L and the deterioration occurrence time interval T are input from the input unit 61, and have the same configuration as the function storage unit 54 shown in FIG. 4 as shown in parentheses in FIG. Or the processing described in the second embodiment is executed.
[0019]
【The invention's effect】
As described above, according to the present invention, the test subject does not actually evaluate the degraded video and / or sound for a long time, and the permissible rate for the degraded video and / or the sound and the permissible rate corresponding to the permissible long time viewing are allowed. The packet loss rate is obtained.
[Brief description of the drawings]
FIG. 1 is a diagram showing an example of a functional configuration according to a first embodiment of the present invention.
FIG. 2 is a graph for explaining the derivation of a relational expression F1 used in the deterioration quality estimation unit 10 in the embodiment.
FIG. 3 is a graph for explaining an example of deriving a relational expression F2 used in a deterioration occurrence time interval estimating unit 20 in the embodiment.
FIG. 4 is a diagram showing an example of a functional configuration according to a second embodiment of the present invention.
FIG. 5 is a graph for explaining an example of deriving a relational expression F3 used in an allowable ratio estimating unit 50 in the second embodiment.
FIG. 6 is a graph showing an example of the relationship between the obtained packet loss rate for long-time viewing and the allowable rate for degraded video.
FIG. 7 is a diagram showing an example of a configuration for causing a computer to function according to the present invention.

Claims

Deterioration quality estimation means for estimating the quality of the deteriorated video and / or sound of the unit time length from the packet loss rate of the unit time;
Degradation occurrence time interval estimating means for estimating an occurrence time interval of a degradation section allowed from the estimated quality of the degradation video and / or sound of the unit time length and an allowable rate for the degradation video and / or sound;
Loss rate calculating means for determining a packet loss rate allowed for the deteriorated video and / or sound from the packet loss rate per unit time and the estimated time interval of occurrence of the allowable deterioration section,
A permissible packet loss rate determining apparatus comprising:

Deterioration quality estimation means for estimating the quality of the deteriorated video and / or sound of the unit time length from the packet loss rate of the unit time;
A permissible rate estimating means for estimating a permissible rate for the degraded video and / or sound from the estimated quality of the degraded video and / or sound of the unit time length and the occurrence time interval of the degraded section;
A permissible packet loss rate determination device comprising: a permissible packet loss rate determining unit that determines a permissible packet loss rate for degraded video and / or audio from the per unit time packet loss rate and the occurrence time interval of the degraded section.

The device according to claim 1 or 2,
The deterioration quality estimating means includes a unit having a unit of a packet loss rate as a variable, which is obtained by a test for subjectively evaluating the quality of video and / or audio in which packet loss has occurred in a video and / or audio signal. A permissible packet loss rate determining apparatus, which is a means for calculating the quality of a deteriorated video and / or sound per unit time by substituting a time packet loss rate into a function.

The apparatus according to claim 1,
The deterioration occurrence time interval estimating means obtains, by a subjective evaluation test, a generation time interval of a deterioration image and / or sound section which is permissible to a viewer when viewing a deterioration image and / or sound having a unit time length. A function storage unit that stores a function that uses the deterioration rate as a parameter with the allowable ratio for the deteriorated image and / or sound as a parameter, and stores a function corresponding to the estimated ratio for the estimated deteriorated image and / or sound. Means for calculating a permissible degradation interval occurrence time interval by taking out the degradation quality of the unit time length and substituting the pertinent degradation quality for the unit time length.

The apparatus according to claim 2,
The permissible rate estimating means uses, as a parameter, a deterioration quality obtained by a test for subjectively evaluating a permissible rate for the deteriorated video and / or sound when the viewer views a deteriorated video and / or sound having a unit time length, and A function storage unit for storing a function using the occurrence time interval as a variable, and substituting the occurrence time interval of the deterioration section into a function extracted in accordance with the estimated deterioration quality to calculate the deterioration image and / or sound. Calculating means for calculating an allowable packet loss rate.

A program for causing a computer to function as the allowable packet loss rate determining device according to claim 1.