JP2014178578A - Sound processor - Google Patents

Sound processor Download PDF

Info

Publication number
JP2014178578A
JP2014178578A JP2013053485A JP2013053485A JP2014178578A JP 2014178578 A JP2014178578 A JP 2014178578A JP 2013053485 A JP2013053485 A JP 2013053485A JP 2013053485 A JP2013053485 A JP 2013053485A JP 2014178578 A JP2014178578 A JP 2014178578A
Authority
JP
Japan
Prior art keywords
acoustic
filter
component
phase shift
value
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
JP2013053485A
Other languages
Japanese (ja)
Inventor
Janner Geordi
ジェイナー ジョルディ
Marxer Ricardo
マークサー リカルド
Bonada Jordi
ボナダ ジョルディ
Kazunobu Kondo
多伸 近藤
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Yamaha Corp
Original Assignee
Yamaha Corp
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Yamaha Corp filed Critical Yamaha Corp
Priority to JP2013053485A priority Critical patent/JP2014178578A/en
Publication of JP2014178578A publication Critical patent/JP2014178578A/en
Pending legal-status Critical Current

Links

Images

Landscapes

  • Circuit For Audible Band Transducer (AREA)

Abstract

PROBLEM TO BE SOLVED: To provide a sound processor capable of preventing a residual component out of a sound component to be suppressed from auditorily being perceived.SOLUTION: A separation filter formation part 32 forms a separation filter G (τ) having a coefficient value g (k,τ) of each frequency set so as to suppress a specific component of a sound signal for each unit period overlapping each other on a time base. A phase shift filter formation part 34 forms a phase shift filter P (τ) varying only the amount of phase shift θ (k,τ) corresponding to a coefficient value g (k,τ) of the frequency out of the separation filter G (τ) and a variation n (k,τ) varied for each unit period, for each unit period.

Description

本発明は、複数の音響成分が混合された音響信号から特定の音響成分を分離(抽出または抑圧)する技術に関する。   The present invention relates to a technique for separating (extracting or suppressing) a specific acoustic component from an acoustic signal in which a plurality of acoustic components are mixed.

音響信号の目的音成分と非目的音成分とが混在した音響信号について非目的音成分を抑圧(各成分を分離)するための各種の技術が従来から提案されている。例えば特許文献1には、音響信号の単位期間(フレーム)毎のスペクトルに抑圧係数列(例えばウィナーフィルタ)を作用させることで音響信号のうち目的音成分以外の非目的音成分を抑圧(目的音成分を強調)する技術が開示されている。   Various techniques for suppressing non-target sound components (separating each component) have been proposed in the past for acoustic signals in which target sound components and non-target sound components are mixed. For example, Patent Document 1 discloses that a non-target sound component other than a target sound component is suppressed (target sound) in an acoustic signal by applying a suppression coefficient sequence (for example, a Wiener filter) to a spectrum for each unit period (frame) of the sound signal. Techniques for emphasizing ingredients are disclosed.

特開2012−113190公報JP2012-113190A

しかし、非目的音成分を確実に抑圧し得る完全な処理係数列を推定することは現実には困難であるから、処理係数列を適用した処理後の音響信号には非目的音成分の一部の成分(以下「残存成分」という)が残存し得る。以上の事情を背景として、本発明は、抑圧対象の音響成分のうちの残存成分を聴感的に知覚され難くすることを目的とする。   However, since it is actually difficult to estimate a complete processing coefficient sequence that can reliably suppress the non-target sound component, a part of the non-target sound component is included in the processed acoustic signal to which the processing coefficient sequence is applied. May remain (hereinafter referred to as “residual component”). In view of the above circumstances, an object of the present invention is to make it difficult to perceptually perceive residual components among acoustic components to be suppressed.

以上の課題を解決するために、本発明の音響処理装置は、音響信号の特定成分が抑圧されるように複数の周波数の各々の係数値(例えば係数値g(k,τ))が設定された分離フィルタを、時間軸上で相互に重複する単位期間毎に生成する分離フィルタ生成手段と、複数の周波数の各々における位相を、分離フィルタのうち当該周波数の係数値と単位期間毎に変動する変動値(例えば変動値n(k,τ))とに応じた移相量だけ変化させる移相フィルタを、単位期間毎に生成する移相フィルタ生成手段とを具備する。例えば、移相量は、各周波数の係数値が特定成分を抑圧する度合が大きいほど大きい数値となる制御値(例えば制御値F(k,τ))と変動値との乗算に応じて設定される。   In order to solve the above problems, in the sound processing apparatus of the present invention, coefficient values (for example, coefficient values g (k, τ)) of a plurality of frequencies are set so that specific components of the sound signal are suppressed. The separation filter generating means for generating the separation filter for each unit period that overlaps each other on the time axis, and the phase of each of the plurality of frequencies are varied for each coefficient period of the separation filter and for each unit period. Phase shift filter generation means for generating a phase shift filter that changes by a phase shift amount corresponding to a fluctuation value (for example, fluctuation value n (k, τ)) for each unit period. For example, the amount of phase shift is set according to multiplication of a control value (for example, control value F (k, τ)) and a fluctuation value that becomes a larger numerical value as the degree to which the coefficient value of each frequency suppresses the specific component is larger. The

以上の構成では、音響信号の複数の周波数の各々の移相量が、音響信号の特定成分を抑圧する分離フィルタのうち当該周波数の係数値と単位期間毎に変動する変動値とに応じて設定される。すなわち、音響信号の各周波数の移相量に対する変動値の変動の影響(移相量の変動の度合)が分離フィルタの係数値に応じて周波数毎に制御される。したがって、例えば、音響信号のうち特定成分が優勢な周波数(例えば分離フィルタの係数値が小さい周波数)に対する移相量を、音響信号のうち特定成分以外の音響成分が優勢な周波数に対する移相量と比較して大きく変動させることが可能である。各単位期間は時間軸上で相互に重複する(すなわち、音響信号の各周波数の成分が複数の単位期間にわたる)から、音響信号のうち単位期間毎の移相量の変動が大きい周波数の成分ほど時間軸上の広範囲に分散される。特定成分のエネルギーの最大値は、各周波数成分が時間軸上の広範囲に分散されるほど小さくなる。したがって、以上の構成によれば、分離フィルタの作用の前後で各周波数の位相が維持される構成と比較して、特定成分のうち分離フィルタの係数値では完全には抑圧されない残存成分を聴感的に知覚され難くすることが可能である。   In the above configuration, the phase shift amount of each of the plurality of frequencies of the acoustic signal is set according to the coefficient value of the frequency and the fluctuation value that varies for each unit period among the separation filters that suppress the specific component of the acoustic signal. Is done. That is, the influence of the fluctuation of the fluctuation value on the phase shift amount of each frequency of the acoustic signal (the degree of fluctuation of the phase shift quantity) is controlled for each frequency according to the coefficient value of the separation filter. Therefore, for example, the phase shift amount for a frequency where a specific component is dominant in the acoustic signal (for example, the frequency where the coefficient value of the separation filter is small) is the phase shift amount for a frequency where an acoustic component other than the specific component is dominant in the acoustic signal. It is possible to make a large fluctuation in comparison. Since each unit period overlaps with each other on the time axis (that is, each frequency component of the acoustic signal extends over a plurality of unit periods), the component of the acoustic signal whose frequency shift amount is large for each unit period. Widely distributed on the time axis. The maximum value of the energy of the specific component becomes smaller as each frequency component is dispersed over a wide range on the time axis. Therefore, according to the above configuration, compared to the configuration in which the phase of each frequency is maintained before and after the operation of the separation filter, the remaining components that are not completely suppressed by the coefficient value of the separation filter among the specific components are audibly heard. It is possible to make it difficult to perceive.

本発明の好適な態様において、変動値は、複数の周波数の各々について個別に設定される。以上の構成によれば、複数の周波数にわたる残存成分を効果的に知覚され難くすることが可能である。変動値の好適例は、単位期間毎に生成された乱数(例えば一様乱数)である。以上の構成によれば、乱数を発生させる簡便な構成により、残存成分を時間軸上に効果的に分散させることが可能である。   In a preferred aspect of the present invention, the variation value is individually set for each of the plurality of frequencies. According to the above configuration, it is possible to make it difficult to effectively perceive residual components over a plurality of frequencies. A suitable example of the fluctuation value is a random number (for example, a uniform random number) generated for each unit period. According to the above configuration, it is possible to effectively disperse the remaining components on the time axis with a simple configuration for generating random numbers.

本発明の好適な態様に係る音響処理装置は、分離フィルタと移相フィルタとを音響信号に作用させる信号処理手段を具備する。以上の構成によれば、分離フィルタと移相フィルタとを作用させた音響信号を生成することが可能である。   The acoustic processing apparatus according to a preferred aspect of the present invention includes signal processing means for causing the separation filter and the phase shift filter to act on the acoustic signal. According to the above configuration, it is possible to generate an acoustic signal in which the separation filter and the phase shift filter are operated.

以上の各態様に係る音響処理装置は、音響信号の処理に専用されるDSP(Digital Signal Processor)などのハードウェア(電子回路)によって実現されるほか、CPU(Central Processing Unit)等の汎用の演算処理装置とプログラムとの協働によっても実現される。具体的には、本発明のプログラムは、音響信号の特定成分が抑圧されるように複数の周波数の各々の係数値が設定された分離フィルタを、時間軸上で相互に重複する単位期間毎に生成する分離フィルタ生成処理と、複数の周波数の各々における位相を、分離フィルタのうち当該周波数の係数値と単位期間毎に変動する変動値とに応じた移相量だけ変化させる移相フィルタを、単位期間毎に生成する移相フィルタ生成処理とをコンピュータに実行させる。本発明のプログラムは、コンピュータが読取可能な記録媒体に格納された形態で提供されてコンピュータにインストールされ得る。記録媒体は、例えば非一過性(non-transitory)の記録媒体であり、CD-ROM等の光学式記録媒体(光ディスク)が好例であるが、半導体記録媒体や磁気記録媒体等の公知の任意の形式の記録媒体を包含し得る。また、例えば、本発明のプログラムは、通信網を介した配信の形態で提供されてコンピュータにインストールされ得る。   The sound processing apparatus according to each of the above aspects is realized by hardware (electronic circuit) such as a DSP (Digital Signal Processor) dedicated to processing of an acoustic signal, or a general-purpose operation such as a CPU (Central Processing Unit). This is also realized by cooperation between the processing device and the program. Specifically, the program of the present invention performs separation filters in which coefficient values of a plurality of frequencies are set so that specific components of an acoustic signal are suppressed for each unit period that overlaps each other on the time axis. A separation filter generation process to generate, and a phase shift filter that changes a phase at each of a plurality of frequencies by a phase shift amount corresponding to a coefficient value of the frequency and a variation value that varies for each unit period of the separation filter, A computer is caused to execute a phase shift filter generation process generated for each unit period. The program of the present invention can be provided in a form stored in a computer-readable recording medium and installed in the computer. The recording medium is, for example, a non-transitory recording medium, and an optical recording medium (optical disk) such as a CD-ROM is a good example, but a known arbitrary one such as a semiconductor recording medium or a magnetic recording medium This type of recording medium can be included. For example, the program of the present invention can be provided in the form of distribution via a communication network and installed in a computer.

本発明の第1実施形態に係る音響処理装置のブロック図である。1 is a block diagram of a sound processing apparatus according to a first embodiment of the present invention. フィルタ生成部の具体的な構成のブロック図である。It is a block diagram of the concrete structure of a filter production | generation part.

<第1実施形態>
図1は、本発明の第1実施形態に係る音響処理装置100のブロック図である。図1に示すように、音響処理装置100には信号供給装置12と放音装置14とが接続される。信号供給装置12は、音響信号SXを音響処理装置100に供給する。音響信号SXは、複数の音響成分(例えば楽音や音声)の混合音の波形を示す時間領域信号である。例えば、複数種の楽器の演奏音(歌唱音等の音声も含意する)の混合音を示す音響信号SXが音響処理装置100に供給される。周囲の音響を収音して音響信号SXを生成する収音機器や、可搬型または内蔵型の記録媒体から音響信号SXを取得して音響処理装置100に供給する再生装置や、通信網から音響信号SXを受信して音響処理装置100に供給する通信装置が信号供給装置12として採用され得る。
<First Embodiment>
FIG. 1 is a block diagram of a sound processing apparatus 100 according to the first embodiment of the present invention. As shown in FIG. 1, a signal supply device 12 and a sound emitting device 14 are connected to the sound processing device 100. The signal supply device 12 supplies the acoustic signal SX to the acoustic processing device 100. The acoustic signal SX is a time domain signal indicating a waveform of a mixed sound of a plurality of acoustic components (for example, musical sound and voice). For example, an acoustic signal SX indicating a mixed sound of performance sounds of multiple types of musical instruments (which also imply voices such as singing sounds) is supplied to the acoustic processing device 100. A sound collecting device that collects ambient sounds to generate an acoustic signal SX, a playback device that acquires the acoustic signal SX from a portable or built-in recording medium and supplies the acoustic signal SX to the acoustic processing device 100, or an acoustic from a communication network A communication device that receives the signal SX and supplies the signal SX to the sound processing device 100 may be employed as the signal supply device 12.

第1実施形態の音響処理装置100は、信号供給装置12から供給される音響信号SXに対する音響処理で音響信号SYを生成する信号処理装置である。音響信号SYは、音響信号SXに包含される複数の音響成分のうち1以上の特定の音響成分(以下「第1音響成分」という)を抽出した音響の波形を示す時間領域信号である。すなわち、音響処理装置100は、音響信号SXのうち第1音響成分以外の1以上の音響成分(以下「第2音響成分」という)を抑圧することで音響信号SYを生成する。第1音響成分は、音響信号SXのうち強調対象となる目的音成分とも換言され、第2音響成分は、音響信号SXのうち抑圧対象となる非目的音成分とも換言される。放音装置14(例えばスピーカやヘッドホン)は、音響処理装置100から供給される音響信号SYに応じた音波を放射する。なお、音響信号SYをデジタルからアナログに変換するD/A変換器の図示は便宜的に省略した。   The acoustic processing device 100 according to the first embodiment is a signal processing device that generates an acoustic signal SY by acoustic processing on the acoustic signal SX supplied from the signal supply device 12. The acoustic signal SY is a time domain signal indicating an acoustic waveform obtained by extracting one or more specific acoustic components (hereinafter referred to as “first acoustic component”) from among a plurality of acoustic components included in the acoustic signal SX. That is, the sound processing apparatus 100 generates the sound signal SY by suppressing one or more sound components (hereinafter referred to as “second sound component”) other than the first sound component in the sound signal SX. The first acoustic component is also referred to as a target sound component to be emphasized in the acoustic signal SX, and the second acoustic component is also referred to as a non-target sound component to be suppressed in the acoustic signal SX. The sound emitting device 14 (for example, a speaker or a headphone) emits a sound wave corresponding to the acoustic signal SY supplied from the acoustic processing device 100. The D / A converter that converts the acoustic signal SY from digital to analog is not shown for convenience.

図1に示すように、音響処理装置100は、演算処理装置22と記憶装置24とを具備するコンピュータシステムで実現される。記憶装置24は、演算処理装置22が実行するプログラムや演算処理装置22が使用する各種のデータを記憶する。半導体記録媒体や磁気記録媒体等の公知の記録媒体や複数種の記録媒体の組合せが記憶装置24として任意に採用され得る。音響信号SXを記憶装置24に記憶した構成(したがって信号供給装置12は省略され得る)も好適である。   As shown in FIG. 1, the sound processing device 100 is realized by a computer system including an arithmetic processing device 22 and a storage device 24. The storage device 24 stores a program executed by the arithmetic processing device 22 and various data used by the arithmetic processing device 22. A known recording medium such as a semiconductor recording medium or a magnetic recording medium or a combination of a plurality of types of recording media can be arbitrarily employed as the storage device 24. A configuration in which the acoustic signal SX is stored in the storage device 24 (therefore, the signal supply device 12 can be omitted) is also suitable.

図1の演算処理装置22は、記憶装置24に記憶されたプログラムを実行することで、音響信号SXから音響信号SYを生成するための複数の機能(周波数分析部20,フィルタ生成部30,信号処理部40,波形生成部50)を実現する。なお、演算処理装置22の各機能を複数の集積回路に分散した構成や、専用の電子回路(例えばDSP)が一部の機能を実現する構成も採用され得る。   The arithmetic processing unit 22 in FIG. 1 executes a program stored in the storage device 24 to thereby generate a plurality of functions (frequency analysis unit 20, filter generation unit 30, signal) for generating the acoustic signal SY from the acoustic signal SX. A processing unit 40 and a waveform generation unit 50) are realized. A configuration in which each function of the arithmetic processing unit 22 is distributed over a plurality of integrated circuits, or a configuration in which a dedicated electronic circuit (for example, a DSP) realizes a part of the functions may be employed.

図1の周波数分析部20は、音響信号SXを時間軸上で区分した複数の単位期間(フレーム)の各々についてスペクトル(複素スペクトル)QX(τ)を算定する。任意の1個の単位期間のスペクトルQX(τ)は、周波数軸上の相異なる周波数に対応する複数の周波数成分(複素数)X(k,τ)の系列である。記号kは、周波数軸上の任意の1個の周波数(周波数帯域)を指示する変数であり、記号τは、時間軸上の複数の単位期間の何れかを指示する変数である。相前後する各単位期間は時間軸上で相互に部分的に重複する。スペクトルQX(τ)の生成には例えば短時間フーリエ変換等の公知の周波数分析が任意に利用される。なお、通過帯域が相違する複数の帯域通過フィルタで構成されるフィルタバンクを周波数分析部20として採用することも可能である。また、音響信号SXを記憶装置24に記憶した構成では、記憶装置24から周波数分析部20に音響信号SXが供給される。   The frequency analysis unit 20 in FIG. 1 calculates a spectrum (complex spectrum) QX (τ) for each of a plurality of unit periods (frames) obtained by dividing the acoustic signal SX on the time axis. The spectrum QX (τ) of any one unit period is a series of a plurality of frequency components (complex numbers) X (k, τ) corresponding to different frequencies on the frequency axis. The symbol k is a variable indicating any one frequency (frequency band) on the frequency axis, and the symbol τ is a variable indicating any of a plurality of unit periods on the time axis. Each successive unit period partially overlaps with each other on the time axis. For the generation of the spectrum QX (τ), a known frequency analysis such as a short-time Fourier transform is arbitrarily used. Note that a filter bank composed of a plurality of bandpass filters having different passbands may be employed as the frequency analysis unit 20. In the configuration in which the acoustic signal SX is stored in the storage device 24, the acoustic signal SX is supplied from the storage device 24 to the frequency analysis unit 20.

図1のフィルタ生成部30は、処理フィルタM(τ)を単位期間毎に順次に生成する。本実施形態の処理フィルタM(τ)は、音響信号SX(スペクトルQX(τ))の第2音響成分を抑圧するためのフィルタであり、相異なる周波数(または周波数帯域)に対応する複数の係数値m(k,τ)の系列である。   The filter generation unit 30 in FIG. 1 sequentially generates the processing filter M (τ) for each unit period. The processing filter M (τ) of this embodiment is a filter for suppressing the second acoustic component of the acoustic signal SX (spectrum QX (τ)), and has a plurality of factors corresponding to different frequencies (or frequency bands). It is a series of numerical values m (k, τ).

信号処理部40は、音響信号SXの各単位期間のスペクトルQX(τ)に当該単位期間の処理フィルタM(τ)を作用させることで、音響信号SYのスペクトル(複素スペクトル)QY(τ)を単位期間毎に順次に生成する。具体的には、信号処理部40は、以下の数式(1)で表現される通り、音響信号SXの各単位期間の周波数成分X(k,τ)に対し、当該単位期間についてフィルタ生成部30が生成した処理フィルタM(τ)のうち当該周波数成分X(k,τ)と共通の周波数に対応する係数値m(k,τ)を乗算することで、音響信号SYのスペクトルQY(τ)の周波数成分Y(k,τ)を算定する。

Figure 2014178578
The signal processing unit 40 applies the processing filter M (τ) of the unit period to the spectrum QX (τ) of each unit period of the acoustic signal SX, thereby obtaining the spectrum (complex spectrum) QY (τ) of the acoustic signal SY. Generated sequentially for each unit period. Specifically, as represented by the following formula (1), the signal processing unit 40 performs the filter generation unit 30 for the unit period for the frequency component X (k, τ) of each unit period of the acoustic signal SX. Is multiplied by a coefficient value m (k, τ) corresponding to a common frequency with the frequency component X (k, τ) of the processing filter M (τ) generated by the signal SY. The frequency component Y (k, τ) of is calculated.
Figure 2014178578

波形生成部50は、信号処理部40が単位期間毎に生成するスペクトルQY(τ)から時間領域の音響信号SYを生成する。具体的には、波形生成部50は、各単位期間のスペクトルQY(τ)を例えば短時間逆フーリエ変換で時間領域の信号に変換し、相前後する各単位期間の信号を時間軸上で相互に重複させた状態で加算することにより音響信号SYを生成する。波形生成部50が生成した音響信号SYが放音装置14に供給されることで音響として放射される。   The waveform generator 50 generates a time domain acoustic signal SY from the spectrum QY (τ) generated by the signal processor 40 for each unit period. Specifically, the waveform generation unit 50 converts the spectrum QY (τ) of each unit period into a time domain signal by, for example, short-time inverse Fourier transform, and mutually converts the signal of each unit period on the time axis. The sound signal SY is generated by adding the signals in a state where they are overlapped with each other. The acoustic signal SY generated by the waveform generation unit 50 is supplied to the sound emitting device 14 and is emitted as sound.

第1実施形態のフィルタ生成部30が生成する処理フィルタM(τ)は、分離フィルタG(τ)と移相フィルタP(τ)とで構成される。分離フィルタG(τ)は、音響信号SXの各周波数成分X(k,τ)の強度を変調するフィルタであり、相異なる周波数に対応する複数の係数値g(k,τ)の系列である。また、移相フィルタP(τ)は、音響信号SXの各周波数成分X(k,τ)の位相を変調するフィルタであり、相異なる周波数に対応する複数の係数値p(k,τ)の系列である。処理フィルタM(τ)の各係数値m(k,τ)は、以下の数式(2)で表現される通り、分離フィルタG(τ)の係数値g(k,τ)と移相フィルタP(τ)の係数値p(k,τ)との乗算値に相当する。

Figure 2014178578
The processing filter M (τ) generated by the filter generation unit 30 of the first embodiment includes a separation filter G (τ) and a phase shift filter P (τ). The separation filter G (τ) is a filter that modulates the intensity of each frequency component X (k, τ) of the acoustic signal SX, and is a series of a plurality of coefficient values g (k, τ) corresponding to different frequencies. . The phase shift filter P (τ) is a filter that modulates the phase of each frequency component X (k, τ) of the acoustic signal SX, and has a plurality of coefficient values p (k, τ) corresponding to different frequencies. It is a series. Each coefficient value m (k, τ) of the processing filter M (τ) is expressed by the following equation (2), and the coefficient value g (k, τ) of the separation filter G (τ) and the phase shift filter P This corresponds to a multiplication value of the coefficient value p (k, τ) of (τ).
Figure 2014178578

<フィルタ生成部30>
図2は、フィルタ生成部30の具体的な構成のブロック図である。図2に示すように、フィルタ生成部30は、分離フィルタ生成部32と移相フィルタ生成部34と変動値生成部36とを含んで構成される。分離フィルタ生成部32は分離フィルタG(τ)を生成し、移相フィルタ生成部34は移相フィルタP(τ)を生成する。数式(2)を参照して前述した通り、分離フィルタ生成部32が生成した分離フィルタG(τ)と移相フィルタ生成部34が生成した移相フィルタP(τ)とに応じた処理フィルタM(τ)が信号処理部40による第2音響成分の抑圧(第1音響成分の強調)に適用される。
<Filter generation unit 30>
FIG. 2 is a block diagram of a specific configuration of the filter generation unit 30. As shown in FIG. 2, the filter generation unit 30 includes a separation filter generation unit 32, a phase shift filter generation unit 34, and a fluctuation value generation unit 36. The separation filter generation unit 32 generates a separation filter G (τ), and the phase shift filter generation unit 34 generates a phase shift filter P (τ). As described above with reference to Equation (2), the processing filter M according to the separation filter G (τ) generated by the separation filter generation unit 32 and the phase shift filter P (τ) generated by the phase shift filter generation unit 34. (τ) is applied to suppression of the second acoustic component (emphasis of the first acoustic component) by the signal processing unit 40.

分離フィルタ生成部32は、分離フィルタG(τ)を単位期間毎に順次に生成する。本実施形態の分離フィルタ生成部32は、音響信号SXの第2音響成分が抑圧されるように分離フィルタG(τ)の各係数値g(k,τ)を設定する。具体的には、分離フィルタG(τ)は、各係数値g(k,τ)が以下の数式(3)で表現されるウィナー(Wiener)フィルタである。

Figure 2014178578

数式(3)の記号|X(k,τ)|2は周波数成分X(k,τ)のパワーであり、記号|E(k,τ)|2は、音響信号SXから推定される第1音響成分のパワーである。第1音響成分のパワー|E(k,τ)|2の推定には公知の技術が任意に採用される。数式(1)および数式(2)から理解される通り、係数値g(k,τ)は、周波数成分X(k,τ)の強度に作用し、周波数成分X(k,τ)の利得(スペクトルゲイン)を意味する。数式(3)から理解される通り、第2音響成分が第1音響成分に対して優勢な周波数(例えば第2音響成分の強度が第1音響成分の強度を上回る周波数)の係数値g(k,τ)ほど小さい数値となるように、各係数値g(k,τ)は0以上かつ1以下の範囲内で音響信号SXの音響特性に応じて可変に設定される。 The separation filter generation unit 32 sequentially generates the separation filter G (τ) for each unit period. The separation filter generation unit 32 of the present embodiment sets each coefficient value g (k, τ) of the separation filter G (τ) so that the second acoustic component of the acoustic signal SX is suppressed. Specifically, the separation filter G (τ) is a Wiener filter in which each coefficient value g (k, τ) is expressed by the following formula (3).
Figure 2014178578

The symbol | X (k, τ) | 2 in Equation (3) is the power of the frequency component X (k, τ), and the symbol | E (k, τ) | 2 is the first estimated from the acoustic signal SX. It is the power of the acoustic component. A known technique is arbitrarily employed for estimating the power | E (k, τ) | 2 of the first acoustic component. As understood from the equations (1) and (2), the coefficient value g (k, τ) acts on the intensity of the frequency component X (k, τ) and the gain of the frequency component X (k, τ) ( Spectrum gain). As understood from Equation (3), the coefficient value g (k) of the frequency at which the second acoustic component is dominant over the first acoustic component (for example, the frequency at which the intensity of the second acoustic component exceeds the intensity of the first acoustic component). , τ), the coefficient values g (k, τ) are variably set in the range of 0 or more and 1 or less according to the acoustic characteristics of the acoustic signal SX.

ところで、分離フィルタG(τ)は以上のように生成されるから、音響信号SXのスペクトルQX(τ)に分離フィルタG(τ)を作用させるだけの構成(以下「対比例」という)でも、音響信号SXから第2音響成分を抑圧することは可能である(Y(k,τ)=g(k,τ)X(k,τ))。対比例では、音響信号SXのスペクトルQX(τ)の位相が分離フィルタG(τ)の作用後にも維持される。しかし、第1音響成分のパワー|E(k,τ)|2を音響信号SXから完全に推定することは現実には困難であるから、対比例のもとで音響信号SXに分離フィルタG(τ)を作用させた音響には、第2音響成分の一部(残存成分)が残存し得る。第1実施形態の移相フィルタP(τ)は、分離フィルタG(τ)では抑圧し切れない残存成分を聴感的に知覚され難くするためのフィルタである。 By the way, since the separation filter G (τ) is generated as described above, even in a configuration in which the separation filter G (τ) is simply applied to the spectrum QX (τ) of the acoustic signal SX (hereinafter referred to as “proportional”), It is possible to suppress the second acoustic component from the acoustic signal SX (Y (k, τ) = g (k, τ) X (k, τ)). In contrast, the phase of the spectrum QX (τ) of the acoustic signal SX is maintained even after the action of the separation filter G (τ). However, since it is actually difficult to completely estimate the power | E (k, τ) | 2 of the first acoustic component from the acoustic signal SX, the separation filter G ( A part (residual component) of the second acoustic component may remain in the sound on which τ) is applied. The phase shift filter P (τ) of the first embodiment is a filter for making it difficult to perceptually perceive residual components that cannot be suppressed by the separation filter G (τ).

図2の移相フィルタ生成部34は、音響信号SXに作用させる移相フィルタP(τ)を単位期間毎に順次に生成する。移相フィルタP(τ)は、前述の通り、相異なる周波数(または周波数帯域)に対応する複数の係数値p(k,τ)の系列である。具体的には、移相フィルタP(τ)の各係数値p(k,τ)は、以下の数式(4)で表現される通り、スペクトルQX(τ)における各周波数成分X(k,τ)の位相を移相量θ(k,τ)だけ変化させるように設定される。

Figure 2014178578
The phase shift filter generation unit 34 in FIG. 2 sequentially generates the phase shift filter P (τ) that acts on the acoustic signal SX for each unit period. As described above, the phase shift filter P (τ) is a series of a plurality of coefficient values p (k, τ) corresponding to different frequencies (or frequency bands). Specifically, each coefficient value p (k, τ) of the phase shift filter P (τ) is expressed by the following formula (4), and each frequency component X (k, τ) in the spectrum QX (τ) is expressed. ) Is changed by a phase shift amount θ (k, τ).
Figure 2014178578

数式(4)の記号jは虚数単位を意味する。第1実施形態の移相量θ(k,τ)は、制御値F(k,τ)と変動値n(k,τ)とに応じて可変に設定される。具体的には、移相フィルタ生成部34は、以下の数式(5)で表現される通り、制御値F(k,τ)と変動値n(k,τ)とを乗算することで移相量θ(k,τ)を算定する。

Figure 2014178578
The symbol j in the formula (4) means an imaginary unit. The phase shift amount θ (k, τ) of the first embodiment is variably set according to the control value F (k, τ) and the fluctuation value n (k, τ). Specifically, the phase shift filter generation unit 34 multiplies the control value F (k, τ) and the fluctuation value n (k, τ) by the phase shift as expressed by the following formula (5). The quantity θ (k, τ) is calculated.
Figure 2014178578

図2の変動値生成部36は、数式(5)の変動値n(k,τ)を複数の周波数の各々について単位期間毎に算定する。変動値n(k,τ)は、単位期間毎に刻々と変動する数値であり、例えば単位期間毎に発生する乱数である。例えば、変動値生成部36は、所定範囲(-π≦n(k,τ)≦π)内の各数値が略一様な確率で発生する一様乱数を変動値n(k,τ)として算定する。したがって、変動値n(k,τ)は正数または負数に設定され得る。   The fluctuation value generation unit 36 in FIG. 2 calculates the fluctuation value n (k, τ) of Equation (5) for each of a plurality of frequencies for each unit period. The variation value n (k, τ) is a numerical value that varies every unit period, for example, a random number generated every unit period. For example, the fluctuation value generation unit 36 uses, as the fluctuation value n (k, τ), a uniform random number in which each numerical value within a predetermined range (−π ≦ n (k, τ) ≦ π) is generated with a substantially uniform probability. Calculate. Therefore, the variation value n (k, τ) can be set to a positive number or a negative number.

数式(5)の制御値F(k,τ)は、複数の周波数の各々について単位期間毎に設定される。具体的には、任意の1個の単位期間における任意の1個の周波数の制御値F(k,τ)は、当該単位期間の分離フィルタG(τ)のうち当該周波数の係数値g(k,τ)に応じて設定される。具体的には、移相フィルタ生成部34は、係数値g(k,τ)により音響信号SXが抑圧される度合が大きい(係数値g(k,τ)が小さい)ほど制御値F(k,τ)を大きい数値に設定する。例えば、前掲の数式(6)から理解される通り、所定値(数式(6)の例示では1)から係数値g(k,τ)を減算した数値が制御値F(k,τ)として好適である。   The control value F (k, τ) in Expression (5) is set for each unit period for each of a plurality of frequencies. Specifically, the control value F (k, τ) of any one frequency in any one unit period is the coefficient value g (k) of the frequency in the separation filter G (τ) of the unit period. , τ). Specifically, the phase shift filter generation unit 34 increases the control value F (k) as the degree to which the acoustic signal SX is suppressed by the coefficient value g (k, τ) increases (the coefficient value g (k, τ) decreases). , τ) is set to a large value. For example, as understood from Equation (6) above, a value obtained by subtracting the coefficient value g (k, τ) from a predetermined value (1 in the example of Equation (6)) is suitable as the control value F (k, τ). It is.

以上の説明から理解される通り、本実施形態では、各周波数の移相量θ(k,τ)に対する変動値n(k,τ)の変動の影響(移相量θ(k,τ)の変動の度合)が、分離フィルタG(τ)の係数値g(k,τ)(制御値F(k,τ))に応じて単位期間毎および周波数毎に制御される。係数値g(k,τ)が小さい(第2音響成分が優勢である)ほど制御値F(k,τ)は大きな数値に設定されるから、音響信号SXのうち第2音響成分が優勢な周波数の移相量θ(k,τ)は、第1音響成分が優勢な周波数の移相量θ(k,τ)と比較して大きく変動するという傾向がある。   As understood from the above description, in the present embodiment, the influence of the fluctuation of the fluctuation value n (k, τ) on the phase shift quantity θ (k, τ) of each frequency (the phase shift quantity θ (k, τ) The degree of variation is controlled for each unit period and for each frequency according to the coefficient value g (k, τ) (control value F (k, τ)) of the separation filter G (τ). Since the control value F (k, τ) is set to a larger value as the coefficient value g (k, τ) is smaller (the second acoustic component is more dominant), the second acoustic component is dominant in the acoustic signal SX. The frequency phase shift amount θ (k, τ) tends to vary greatly compared to the phase shift amount θ (k, τ) at which the first acoustic component is dominant.

音響信号SXのうち時間軸上の特定の時点に存在する特定の周波数の残存成分に着目する。各単位期間は時間軸上で相互に重複する(すなわち、音響信号の各周波数の成分が複数の単位期間にわたる)から、残存成分は、複数の単位期間に重複して存在する。各単位期間の残存成分は、当該単位期間の移相フィルタP(τ)の係数値p(k,τ)の作用で移相量θ(k,τ)だけ位相が変化する。移相量θ(k,τ)は、単位期間毎に変動する変動値n(k,τ)に応じて設定されるから、各単位期間の残存成分の位相は単位期間毎に相異なる移相量θ(k,τ)だけ変動する。すなわち、各単位期間の残存成分は、時間軸上の相異なる時点に移動する。波形生成部50では、各単位期間が相互に重複した状態で加算されるから、移相フィルタP(τ)の作用前に時間軸上の1点に存在していた残存成分は、移相フィルタP(τ)の作用後には時間軸上の複数の時点に分散される。すなわち、音響信号SXの残存成分の強度(エネルギー)が時間軸上で分散される。移相フィルタP(τ)は、音響信号SXのうち第2音響成分の残存成分に残響を付与する要素とも換言される。   Attention is paid to a residual component of a specific frequency existing at a specific time point on the time axis in the acoustic signal SX. Since each unit period overlaps with each other on the time axis (that is, the component of each frequency of the acoustic signal extends over a plurality of unit periods), the remaining component exists in a plurality of unit periods. The phase of the remaining component in each unit period changes by the phase shift amount θ (k, τ) by the action of the coefficient value p (k, τ) of the phase shift filter P (τ) in the unit period. Since the phase shift amount θ (k, τ) is set according to the fluctuation value n (k, τ) that varies for each unit period, the phase of the remaining component in each unit period is different for each unit period. It fluctuates by an amount θ (k, τ). That is, the remaining component of each unit period moves to different time points on the time axis. In the waveform generation unit 50, each unit period is added in a state of being overlapped with each other. Therefore, the remaining component existing at one point on the time axis before the action of the phase shift filter P (τ) is the phase shift filter. After the action of P (τ), it is dispersed at a plurality of time points on the time axis. That is, the intensity (energy) of the remaining component of the acoustic signal SX is dispersed on the time axis. In other words, the phase shift filter P (τ) is an element that gives reverberation to the remaining component of the second acoustic component of the acoustic signal SX.

各周波数成分X(k,τ)の移相量θ(k,τ)は、第2音響成分が優勢な周波数ほど大きいから、音響信号SXに移相フィルタP(τ)を作用させることで、音響信号SXの複数の周波数成分X(k,τ)のうち第2音響成分が優勢な周波数の周波数成分X(k,τ)ほど時間軸上の広範囲に分散される。したがって、第1実施形態によれば、音響信号SXの位相が分離フィルタG(τ)の作用後にも維持される対比例と比較して、第2音響成分の残存成分を聴感的に知覚され難くすることができる。   Since the phase shift amount θ (k, τ) of each frequency component X (k, τ) is larger as the frequency where the second acoustic component is dominant, by applying the phase shift filter P (τ) to the acoustic signal SX, Of the plurality of frequency components X (k, τ) of the acoustic signal SX, the frequency component X (k, τ) having the dominant frequency of the second acoustic component is dispersed over a wide range on the time axis. Therefore, according to the first embodiment, the remaining component of the second acoustic component is less likely to be perceptually perceived as compared with the contrast in which the phase of the acoustic signal SX is maintained even after the action of the separation filter G (τ). can do.

<第2実施形態>
本発明の第2実施形態を以下に説明する。なお、以下に例示する各構成において作用や機能が第1実施形態と同等である要素については、第1実施形態の説明で参照した符号を流用して各々の詳細な説明を適宜に省略する。
Second Embodiment
A second embodiment of the present invention will be described below. In addition, about the element in which an effect | action and a function are equivalent to 1st Embodiment in each structure illustrated below, each reference detailed in description of 1st Embodiment is diverted, and each detailed description is abbreviate | omitted suitably.

第1実施形態では、音響信号SXの第1音響成分を抽出した音響信号SYを生成した。第2実施形態の音響処理装置100は、音響信号SY1および音響信号SY2を音響信号SXから選択的に生成する。音響信号SY1は、第1実施形態の音響信号SYと同様に、第1音響成分を目的音成分として音響信号SXから抽出(第2音響成分を非目的音成分として抑圧)した時間領域信号であり、音響信号SY2は、第2音響成分を目的音成分として音響信号SXから抽出(第1音響成分を非目的音成分として抑圧)した時間領域信号である。音響信号SY1または音響信号SY2の何れを生成するかは、例えば利用者からの指示に応じて選択される。   In the first embodiment, the acoustic signal SY obtained by extracting the first acoustic component of the acoustic signal SX is generated. The acoustic processing apparatus 100 according to the second embodiment selectively generates the acoustic signal SY1 and the acoustic signal SY2 from the acoustic signal SX. The acoustic signal SY1 is a time domain signal extracted from the acoustic signal SX with the first acoustic component as the target sound component (suppressing the second acoustic component as the non-target sound component), like the acoustic signal SY of the first embodiment. The acoustic signal SY2 is a time-domain signal extracted from the acoustic signal SX using the second acoustic component as the target sound component (suppressing the first acoustic component as the non-target sound component). Whether the acoustic signal SY1 or the acoustic signal SY2 is generated is selected in accordance with an instruction from the user, for example.

第2実施形態のフィルタ生成部30は、音響信号SXの第1音響成分を抽出するための処理フィルタM1(τ)と音響信号SXの第2音響成分を抽出するための処理フィルタM2(τ)とを選択的に生成する。処理フィルタMi(τ)(i=1,2)は、相異なる周波数に対応する複数の係数値mi(k,τ)の系列であり、分離フィルタGi(τ)と移相フィルタPi(τ)とを含んで構成される。分離フィルタGi(τ)は、音響信号SXの各周波数成分X(k,τ)の強度を変調する複数の係数値gi(k,τ)で構成され、移相フィルタPi(τ)は、各周波数成分X(k,τ)の位相を変調する複数の係数値pi(k,τ)で構成される。   The filter generator 30 of the second embodiment includes a processing filter M1 (τ) for extracting the first acoustic component of the acoustic signal SX and a processing filter M2 (τ) for extracting the second acoustic component of the acoustic signal SX. And are selectively generated. The processing filter Mi (τ) (i = 1, 2) is a series of a plurality of coefficient values mi (k, τ) corresponding to different frequencies, and the separation filter Gi (τ) and the phase shift filter Pi (τ). It is comprised including. The separation filter Gi (τ) is composed of a plurality of coefficient values gi (k, τ) that modulate the intensity of each frequency component X (k, τ) of the acoustic signal SX, and the phase shift filter Pi (τ) It is composed of a plurality of coefficient values pi (k, τ) for modulating the phase of the frequency component X (k, τ).

信号処理部40は、音響信号SXの各単位期間のスペクトルQX(τ)に当該単位期間の処理フィルタMi(τ)を作用させることで、音響信号SYiのスペクトルQYi(τ)を単位期間毎に順次に生成する。具体的には、信号処理部40は、以下の数式(7)の演算で音響信号SYiの周波数成分Yi(k,τ)を算定する。

Figure 2014178578
The signal processing unit 40 applies the processing filter Mi (τ) of the unit period to the spectrum QX (τ) of each unit period of the acoustic signal SX, thereby obtaining the spectrum QYi (τ) of the acoustic signal SYi for each unit period. Generate sequentially. Specifically, the signal processing unit 40 calculates the frequency component Yi (k, τ) of the acoustic signal SYi by the calculation of the following formula (7).
Figure 2014178578

処理フィルタMi(τ)の各係数値mi(k,τ)は、以下の数式(8)で表現される通り、係数値gi(k,τ)と係数値pi(k,τ)との乗算値に相当する。

Figure 2014178578
Each coefficient value mi (k, τ) of the processing filter Mi (τ) is multiplied by the coefficient value gi (k, τ) and the coefficient value pi (k, τ) as expressed by the following equation (8). Corresponds to the value.
Figure 2014178578

分離フィルタG1(τ)の各係数値g1(k,τ)は、第1実施形態の係数値g(k,τ)と同様に前掲の数式(3)で算定され、周波数成分X(k,τ)にて第2音響成分が優勢であるほど小さい数値となる。したがって、数式(7)および数式(8)から理解される通り、処理フィルタM1(k,τ)は、第1実施形態の処理フィルタM(k,τ)と同様に、音響信号SXの第2音響成分を抑圧するとともに、第2音響成分の残存成分(分離フィルタG1(τ)では抑圧し切れない成分)が移相により聴感的に知覚され難くなるように作用する。   Each coefficient value g1 (k, τ) of the separation filter G1 (τ) is calculated by the above equation (3) similarly to the coefficient value g (k, τ) of the first embodiment, and the frequency component X (k, τ) The value becomes smaller as the second acoustic component becomes dominant at τ). Therefore, as can be understood from Equation (7) and Equation (8), the processing filter M1 (k, τ) is similar to the processing filter M (k, τ) of the first embodiment. While suppressing the acoustic component, it acts so that the remaining component of the second acoustic component (a component that cannot be suppressed by the separation filter G1 (τ)) is hardly perceptually perceived by the phase shift.

他方、分離フィルタG2(τ)の各係数値g2(k,τ)は、周波数成分X(k,τ)において第1音響成分が優勢であるほど小さい数値に設定される。例えば、係数値g2(k,τ)は、以下の数式(9)の演算で係数値g1(k,τ)に応じて算定される。

Figure 2014178578

以上の説明から理解される通り、処理フィルタM2(k,τ)は、音響信号SXの第1音響成分を抑圧するとともに、第1音響成分の残存成分(分離フィルタG2(τ)では抑圧し切れない成分)が移相により聴感的に知覚され難くなるように作用する。。 On the other hand, each coefficient value g2 (k, τ) of the separation filter G2 (τ) is set to a smaller value as the first acoustic component becomes dominant in the frequency component X (k, τ). For example, the coefficient value g2 (k, τ) is calculated according to the coefficient value g1 (k, τ) by the calculation of the following formula (9).
Figure 2014178578

As understood from the above description, the processing filter M2 (k, τ) suppresses the first acoustic component of the acoustic signal SX and completely suppresses the remaining component of the first acoustic component (the separation filter G2 (τ)). Not to be perceptually perceived by phase shift. .

第2実施形態においても第1実施形態と同様の効果が実現される。また、第2実施形態では、音響信号SXから第1音響成分を抽出した音響信号SY1に加え、音響信号SXから第2音響成分を抽出した音響信号SY2を生成することができる。   In the second embodiment, the same effect as in the first embodiment is realized. In the second embodiment, in addition to the acoustic signal SY1 obtained by extracting the first acoustic component from the acoustic signal SX, the acoustic signal SY2 obtained by extracting the second acoustic component from the acoustic signal SX can be generated.

<変形例>
以上に例示した各形態には様々な変形が加えられる。具体的な変形の態様を例示すれば以下の通りである。なお、以下の例示から2以上の態様を任意に選択して組合せてもよい。
<Modification>
Various modifications can be made to each of the forms exemplified above. An example of a specific modification is as follows. Two or more aspects may be arbitrarily selected from the following examples and combined.

(1)分離フィルタG(τ)の形式は前述の各形態での例示に限定されない。例えば、各係数値g(k,τ)が以下の数式(10)で表現されるウィナーフィルタを分離フィルタG(τ)として利用することも可能である。

Figure 2014178578

数式(10)の記号max{ }は、括弧内の最大値を選択する演算(すなわち差分|X(k,τ)|2−|I(k,τ)|2を非負値に制限する演算)を意味する。記号|I(k,τ)|2は、音響信号SXから推定される第2音響成分のパワーである。第2音響成分のパワー|I(k,τ)|2の推定には公知の技術が任意に採用される。 (1) The format of the separation filter G (τ) is not limited to the examples in the above-described embodiments. For example, a Wiener filter in which each coefficient value g (k, τ) is expressed by the following formula (10) can be used as the separation filter G (τ).
Figure 2014178578

The symbol max {} in Equation (10) is an operation for selecting the maximum value in parentheses (that is, an operation for limiting the difference | X (k, τ) | 2 − | I (k, τ) | 2 to a non-negative value) Means. The symbol | I (k, τ) | 2 is the power of the second acoustic component estimated from the acoustic signal SX. A known technique is arbitrarily employed for estimating the power | I (k, τ) | 2 of the second acoustic component.

また、第2実施形態では、分離フィルタG1(τ)の係数値g1(k,τ)から分離フィルタG2(τ)の各係数値g2(k,τ)を算定したが、第1音響成分を抑圧するための分離フィルタG2(τ)の各係数値g2(k,τ)を、以下に例示される数式(11)または(12)の演算で算定することも可能である。

Figure 2014178578

以上の説明から理解される通り、分離フィルタG(τ)は、音響信号SXの特定成分(第1音響成分または第2音響成分)が抑圧されるように複数の周波数の各々の係数値g(k,τ)が設定されたフィルタとして包括的に表現される。 In the second embodiment, each coefficient value g2 (k, τ) of the separation filter G2 (τ) is calculated from the coefficient value g1 (k, τ) of the separation filter G1 (τ). It is also possible to calculate each coefficient value g2 (k, τ) of the separation filter G2 (τ) for suppression by the calculation of the following formula (11) or (12).
Figure 2014178578

As understood from the above description, the separation filter G (τ) has the coefficient values g () of each of the plurality of frequencies so that the specific component (the first acoustic component or the second acoustic component) of the acoustic signal SX is suppressed. k, τ) is comprehensively expressed as a set filter.

(2)変動値生成部36が生成する変動値n(k,τ)は一様乱数に限定されない。例えば各数値の発生確率が正規分布に従う正規乱数を変動値n(k,τ)として利用することも可能である。もっとも、変動値n(k,τ)は乱数に限定されない。例えば、単位期間毎に所定量ずつ増加または減少する数値を変動値n(k,τ)として利用することも可能である。また、前述の各形態では、周波数毎に個別に変動値n(k,τ)を設定したが、複数の周波数にわたり共通の変動値n(τ)を単位期間毎に生成することも可能である。以上の説明から理解される通り、前述の各形態の変動値n(k,τ)は、単位期間毎に変動する数値として包括的に表現される。 (2) The fluctuation value n (k, τ) generated by the fluctuation value generator 36 is not limited to a uniform random number. For example, a normal random number in which the occurrence probability of each numerical value follows a normal distribution can be used as the fluctuation value n (k, τ). However, the fluctuation value n (k, τ) is not limited to a random number. For example, a numerical value that increases or decreases by a predetermined amount for each unit period can be used as the fluctuation value n (k, τ). Further, in each of the above-described embodiments, the variation value n (k, τ) is individually set for each frequency. However, a common variation value n (τ) can be generated for each unit period over a plurality of frequencies. . As understood from the above description, the variation value n (k, τ) of each of the above-described forms is comprehensively expressed as a numerical value that varies for each unit period.

(3)第1実施形態では、所定値(1)から係数値g(k,τ)を減算した数値を制御値F(k,τ)として算定した。以上の構成では、係数値g(k,τ)が1に近い場合(第1音響成分が優勢である場合)にも制御値F(k,τ)が有意な正数に設定され、結果的に周波数成分X(k,τ)が移相される。以上の傾向を考慮すると、第1音響成分が優勢である周波数成分X(k,τ)については位相の変動を積極的に抑制する構成が好適である。例えば、係数値g(k,τ)が所定の閾値THを上回る程度に周波数成分X(k,τ)にて第1音響成分が優勢である場合には、移相フィルタ生成部34は、制御値F(k,τ)を強制的に0に設定する。前掲の数式(5)から理解される通り、制御値F(k,τ)が0の周波数成分については移相量θ(k,τ)も0となる。すなわち、係数値g(k,τ)が所定の閾値THを上回る周波数成分X(k,τ)については位相が変更されない。したがって、以上の構成によれば、音響信号SXの第1音響成分の波形歪を抑制できるという利点がある。なお、第2実施形態においても同様の構成が採用される。具体的には、係数値gi(k,τ)が所定の閾値THi(閾値TH1と閾値TH2との異同は不問)を上回る周波数の制御値Fi(k,τ)(移相量θi(k,τ))が0に設定される。 (3) In the first embodiment, a numerical value obtained by subtracting the coefficient value g (k, τ) from the predetermined value (1) is calculated as the control value F (k, τ). In the above configuration, the control value F (k, τ) is set to a significant positive number even when the coefficient value g (k, τ) is close to 1 (when the first acoustic component is dominant), and as a result The frequency component X (k, τ) is phase-shifted. Considering the above tendency, it is preferable that the frequency component X (k, τ), in which the first acoustic component is dominant, be configured to positively suppress the phase variation. For example, when the first acoustic component is dominant in the frequency component X (k, τ) to such an extent that the coefficient value g (k, τ) exceeds a predetermined threshold value TH, the phase shift filter generation unit 34 performs control. The value F (k, τ) is forcibly set to 0. As understood from the above formula (5), the phase shift amount θ (k, τ) is also zero for the frequency component for which the control value F (k, τ) is zero. That is, the phase is not changed for the frequency component X (k, τ) in which the coefficient value g (k, τ) exceeds the predetermined threshold value TH. Therefore, according to the above structure, there exists an advantage that the waveform distortion of the 1st acoustic component of acoustic signal SX can be suppressed. Note that the same configuration is also adopted in the second embodiment. Specifically, the control value Fi (k, τ) (phase shift amount θi (k, τ)) at a frequency at which the coefficient value gi (k, τ) exceeds a predetermined threshold value THi (the difference between the threshold value TH1 and the threshold value TH2 does not matter). τ)) is set to zero.

(4)第2実施形態においては、音響信号SY1および音響信号SY2を選択的に生成したが、音響信号SY1および音響信号SY2を並列に生成することも可能である。以上の構成によれば、例えば音響信号SY1および音響信号SY2の各々に別個の処理(例えば残響付与等の各種の効果付与)を実行したうえで両者を混合するといった特徴的な音響処理が実現される。 (4) In the second embodiment, the acoustic signal SY1 and the acoustic signal SY2 are selectively generated, but the acoustic signal SY1 and the acoustic signal SY2 can also be generated in parallel. According to the above configuration, for example, characteristic acoustic processing is realized in which the acoustic signal SY1 and the acoustic signal SY2 are separately processed (for example, various effects such as reverberation are imparted) and then mixed. The

(5)携帯電話機等の端末装置と通信するサーバ装置で音響処理装置100を実現することも可能である。例えば、音響処理装置100は、端末装置から受信した音響信号SXから音響信号SYを生成して端末装置に送信する。なお、音響信号SXのスペクトルQX(τ)を端末装置から受信する構成(例えば端末装置が周波数分析部20を具備する構成)では音響処理装置100から周波数分析部20が省略され、スペクトルQY(τ)を端末装置に送信する構成(例えば端末装置が波形生成部50を具備する構成)では音響処理装置100から波形生成部50が省略される。また、音響処理装置100が生成した処理フィルタM(τ)を端末装置等の外部装置に転送して音響信号SXの処理に適用する構成では、信号処理部40も音響処理装置100から省略され得る。 (5) The sound processing apparatus 100 can also be realized by a server device that communicates with a terminal device such as a mobile phone. For example, the acoustic processing device 100 generates an acoustic signal SY from the acoustic signal SX received from the terminal device and transmits the acoustic signal SY to the terminal device. In the configuration in which the spectrum QX (τ) of the acoustic signal SX is received from the terminal device (for example, the configuration in which the terminal device includes the frequency analysis unit 20), the frequency analysis unit 20 is omitted from the acoustic processing device 100, and the spectrum QY (τ ) To the terminal device (for example, a configuration in which the terminal device includes the waveform generation unit 50), the waveform generation unit 50 is omitted from the acoustic processing device 100. In the configuration in which the processing filter M (τ) generated by the acoustic processing device 100 is transferred to an external device such as a terminal device and applied to the processing of the acoustic signal SX, the signal processing unit 40 can also be omitted from the acoustic processing device 100. .

100……音響処理装置、12……信号供給装置、14……放音装置、20……周波数分析部、22……演算処理装置、24……記憶装置、30……フィルタ生成部、32……分離フィルタ生成部、34……移相フィルタ生成部、36……変動値生成部、40……信号処理部、50……波形生成部
DESCRIPTION OF SYMBOLS 100 ... Acoustic processing device, 12 ... Signal supply device, 14 ... Sound emission device, 20 ... Frequency analysis part, 22 ... Arithmetic processing device, 24 ... Memory | storage device, 30 ... Filter production | generation part, 32 ... ... Separation filter generator 34... Phase shift filter generator 36. Fluctuation value generator 40. Signal processor 50. Waveform generator

Claims (4)

音響信号の特定成分が抑圧されるように複数の周波数の各々の係数値が設定された分離フィルタを、時間軸上で相互に重複する単位期間毎に生成する分離フィルタ生成手段と、
前記複数の周波数の各々における位相を、前記分離フィルタのうち当該周波数の係数値と単位期間毎に変動する変動値とに応じた移相量だけ変化させる移相フィルタを、単位期間毎に生成する移相フィルタ生成手段と
を具備する音響処理装置。
Separation filter generation means for generating a separation filter in which coefficient values of each of a plurality of frequencies are set so as to suppress a specific component of an acoustic signal for each unit period overlapping each other on a time axis;
A phase shift filter that changes the phase at each of the plurality of frequencies by a phase shift amount corresponding to a coefficient value of the frequency and a fluctuation value that varies for each unit period of the separation filter is generated for each unit period. A sound processing apparatus comprising: a phase shift filter generating unit.
前記移相量は、各周波数の前記係数値が前記特定成分を抑圧する度合が大きいほど大きい数値となる制御値と前記変動値との乗算に応じて設定される
請求項1の音響処理装置。
The acoustic processing apparatus according to claim 1, wherein the phase shift amount is set according to multiplication of a control value that becomes a larger numerical value as the degree to which the coefficient value of each frequency suppresses the specific component is larger and the fluctuation value.
前記変動値は、単位期間毎に生成された乱数である
請求項1または請求項2の音響処理装置。
The sound processing apparatus according to claim 1, wherein the fluctuation value is a random number generated for each unit period.
前記分離フィルタと前記移相フィルタとを前記音響信号に作用させる信号処理手段
を具備する
請求項1から請求項3の何れかの音響処理装置。
The sound processing device according to claim 1, further comprising: a signal processing unit that causes the separation filter and the phase shift filter to act on the sound signal.
JP2013053485A 2013-03-15 2013-03-15 Sound processor Pending JP2014178578A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
JP2013053485A JP2014178578A (en) 2013-03-15 2013-03-15 Sound processor

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
JP2013053485A JP2014178578A (en) 2013-03-15 2013-03-15 Sound processor

Publications (1)

Publication Number Publication Date
JP2014178578A true JP2014178578A (en) 2014-09-25

Family

ID=51698560

Family Applications (1)

Application Number Title Priority Date Filing Date
JP2013053485A Pending JP2014178578A (en) 2013-03-15 2013-03-15 Sound processor

Country Status (1)

Country Link
JP (1) JP2014178578A (en)

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPH10149198A (en) * 1996-11-21 1998-06-02 Matsushita Electric Ind Co Ltd Noise reduction device
WO2008111462A1 (en) * 2007-03-06 2008-09-18 Nec Corporation Noise suppression method, device, and program
JP2008257049A (en) * 2007-04-06 2008-10-23 Yamaha Corp Noise suppressing device and program
JP2010237703A (en) * 1997-12-08 2010-10-21 Mitsubishi Electric Corp Sound signal processing device and sound signal processing method

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPH10149198A (en) * 1996-11-21 1998-06-02 Matsushita Electric Ind Co Ltd Noise reduction device
JP2010237703A (en) * 1997-12-08 2010-10-21 Mitsubishi Electric Corp Sound signal processing device and sound signal processing method
WO2008111462A1 (en) * 2007-03-06 2008-09-18 Nec Corporation Noise suppression method, device, and program
JP2008257049A (en) * 2007-04-06 2008-10-23 Yamaha Corp Noise suppressing device and program

Similar Documents

Publication Publication Date Title
JP5641186B2 (en) Noise suppression device and program
AU2015295518B2 (en) Apparatus and method for enhancing an audio signal, sound enhancing system
JP2008257049A (en) Noise suppressing device and program
JP6064566B2 (en) Sound processor
US9418677B2 (en) Noise suppressing device, noise suppressing method, and a non-transitory computer-readable recording medium storing noise suppressing program
CN112712816B (en) Training method and device for voice processing model and voice processing method and device
KR102578008B1 (en) Nonlinear adaptive filterbank for psychoacoustic frequency range expansion.
CN113614828A (en) Method and apparatus for fingerprinting audio signals via normalization
JP5915281B2 (en) Sound processor
JP5454330B2 (en) Sound processor
JP5609157B2 (en) Coefficient setting device and noise suppression device
Mu et al. An objective analysis method for perceptual quality of a virtual bass system
JP4533126B2 (en) Proximity sound separation / collection method, proximity sound separation / collection device, proximity sound separation / collection program, recording medium
JP2014178578A (en) Sound processor
Mu et al. A timbre matching approach to enhance audio quality of psychoacoustic bass enhancement system
JP2009182759A (en) Howling suppression apparatus and program
JP2010020013A (en) Noise suppression estimation device and program
Mu Perceptual quality improvement and assessment for virtual bass system
JP2015169901A (en) Acoustic processing device
JP2018072723A (en) Acoustic processing method and sound processing apparatus
JP5463924B2 (en) Sound processor
JP6036141B2 (en) Sound processor
JP2013182161A (en) Acoustic processing device and program
JP2015169900A (en) Noise suppression device
JP6565206B2 (en) Audio processing apparatus and audio processing method

Legal Events

Date Code Title Description
RD04 Notification of resignation of power of attorney

Free format text: JAPANESE INTERMEDIATE CODE: A7424

Effective date: 20150410

A621 Written request for application examination

Free format text: JAPANESE INTERMEDIATE CODE: A621

Effective date: 20160120

A977 Report on retrieval

Free format text: JAPANESE INTERMEDIATE CODE: A971007

Effective date: 20170306

A131 Notification of reasons for refusal

Free format text: JAPANESE INTERMEDIATE CODE: A131

Effective date: 20170321

A02 Decision of refusal

Free format text: JAPANESE INTERMEDIATE CODE: A02

Effective date: 20170926