WO2010035308A1 - エコー消去装置 - Google Patents
エコー消去装置 Download PDFInfo
- Publication number
- WO2010035308A1 WO2010035308A1 PCT/JP2008/002638 JP2008002638W WO2010035308A1 WO 2010035308 A1 WO2010035308 A1 WO 2010035308A1 JP 2008002638 W JP2008002638 W JP 2008002638W WO 2010035308 A1 WO2010035308 A1 WO 2010035308A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- echo
- signal
- power
- spectrum
- amplitude suppression
- Prior art date
Links
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04M—TELEPHONIC COMMUNICATION
- H04M9/00—Arrangements for interconnection not involving centralised switching
- H04M9/08—Two-way loud-speaking telephone systems with means for conditioning the signal, e.g. for suppressing echoes for one or both directions of traffic
- H04M9/082—Two-way loud-speaking telephone systems with means for conditioning the signal, e.g. for suppressing echoes for one or both directions of traffic using echo cancellers
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04B—TRANSMISSION
- H04B3/00—Line transmission systems
- H04B3/02—Details
- H04B3/20—Reducing echo effects or singing; Opening or closing transmitting path; Conditioning for transmission in one direction or the other
- H04B3/23—Reducing echo effects or singing; Opening or closing transmitting path; Conditioning for transmission in one direction or the other using a replica of transmitted signal in the time domain, e.g. echo cancellers
- H04B3/235—Reducing echo effects or singing; Opening or closing transmitting path; Conditioning for transmission in one direction or the other using a replica of transmitted signal in the time domain, e.g. echo cancellers combined with adaptive equaliser
Definitions
- the present invention relates to an acoustic echo canceller used for voice communication.
- the transfer function of the echo path from the speaker to the microphone is estimated by means such as an adaptive filter to generate a pseudo echo signal. Then, the echo contained in the transmitted voice is removed by subtracting the pseudo echo signal from the transmitted voice signal input to the microphone.
- This acoustic echo canceller has an advantage that only echoes can be removed without damaging the transmitted voice even in a state where the echo and the transmitted voice are overlapped, that is, in a double talk state.
- the echo is directly suppressed by subjecting the audio signal to amplitude suppression processing by an audio switch, a center clipper, a variable attenuator, and the like.
- This acoustic echo canceller has an advantage that the echo can be surely removed without generating a residual echo with a simple configuration.
- the double talk state there is a problem that sound quality is deteriorated by suppressing the transmitted voice together with the echo.
- a general acoustic echo canceller removes the main component of echo by linear echo processing, and the residual echo that remains unerased is only when the transmitted voice is not voiced. It was the structure removed by nonlinear echo processing.
- Such a configuration is described in Non-Patent Document 1, for example.
- the transmitted voice is easily concealed by echoes, so it is difficult to determine whether the transmitted voice is sound.
- the acoustic echo device cannot detect the transmitted voice in the double talk state, and there is a problem that the transmitted voice is damaged by performing excessive nonlinear echo processing.
- Patent Literature 1 discloses a method for flexibly controlling a suppression coefficient used for suppressing the amplitude of an audio signal in nonlinear echo processing.
- the echo suppression device disclosed in Patent Document 1 calculates the power of only the transmission voice by subtracting the power of the echo signal from the power of the transmission signal in which the transmission voice and the echo are superimposed, and the transmission voice power is calculated as the transmission voice power.
- a value obtained by adding a predetermined masking threshold and further normalizing with the power value of the transmission signal is used as the suppression coefficient.
- the echo suppression device performs nonlinear echo processing using this suppression coefficient, so that the suppression is automatically weakened in the double talk state where the transmitted voice is present, and the transmitted voice is not easily damaged.
- the power of the echo signal is estimated by the product of the power of the received signal and the acoustic coupling amount of the echo path.
- the power of the echo signal is estimated by the product of the power of the received signal and the acoustic coupling amount of the echo path. It was difficult to measure accurately due to the influence of noise fluctuations and ambient noise. For this reason, there is always an estimation error in the power of the estimated echo signal. Therefore, there is also an error in the transmission voice power calculated based on the power of the echo signal. In particular, in an environment where the ratio of transmitted voice to echo is very poor, such as hands-free telephones, the effect of this error is significant, and the estimated power of the echo signal is subtracted from the transmitted signal power. In some cases, the estimated value of the transmitted voice power becomes 0 or less. As a result, there has been a problem that excessive echo suppression works in the non-linear echo processing and the transmitted voice is impaired.
- the present invention has been made to solve the above-described problems.
- An object of the present invention is to obtain an echo canceller that realizes the above.
- the echo canceller uses an adaptive filter unit that removes echoes from a transmission input signal in which transmitted voice and echo are mixed and outputs a transmission intermediate signal, and a reception side and a transmission intermediate signal.
- the received single talk detecting unit for detecting the single talk state, the first power calculating unit for calculating the power of the received signal, the second power calculating unit for calculating the power of the transmission intermediate signal, and the received single talk detecting unit
- an acoustic coupling amount estimator estimates and updates the acoustic coupling amount from the power ratio of the transmission intermediate signal and the received signal, and from the received signal power and the acoustic coupling amount
- a residual echo power estimation unit that estimates the residual echo power, and a signal-to-echo that estimates the signal-to-echo ratio from the ratio of the power of the transmission intermediate signal and the residual echo power
- An estimation unit, those having the amplitude suppression coefficient determining unit determines the amplitude suppression coefficient in accordance with a signal-to
- the echo canceller uses an adaptive filter unit that removes echoes from a transmission input signal in which transmitted voice and echo are mixed and outputs a transmission intermediate signal, and a reception side and a transmission intermediate signal.
- a received single talk detector that detects a single talk state, a first frequency converter that converts a received signal into a frequency spectrum, a first power spectrum calculator that calculates a power spectrum from the frequency spectrum of the received signal, and transmission
- a single talk state is detected by a second frequency conversion unit that converts the intermediate signal into a frequency spectrum, a second power spectrum calculation unit that calculates a power spectrum from the frequency spectrum of the transmission intermediate signal, and a received single talk detection unit.
- the amount of acoustic coupling is estimated from the power spectrum ratio of the transmitted intermediate signal and the received signal.
- An acoustic coupling amount estimation unit, a residual echo power spectrum estimation unit for estimating the residual echo power spectrum of the transmission intermediate signal from the power spectrum of the received signal and the acoustic coupling amount, and the power spectrum and residual echo power spectrum of the transmission intermediate signal A signal-to-echo spectrum ratio estimation unit that estimates a signal-to-echo spectrum ratio from the ratio, an amplitude suppression coefficient spectrum determination unit that determines an amplitude suppression coefficient spectrum according to the signal-to-echo spectrum ratio, and an amplitude suppression coefficient spectrum
- An amplitude suppression unit that suppresses the amplitude of the frequency spectrum of the transmission intermediate signal, and an inverse frequency conversion unit that converts the frequency spectrum of the transmission intermediate signal subjected to amplitude suppression into a time-series signal are provided.
- FIG. 1 is a block diagram showing a configuration of an echo cancellation apparatus according to Embodiment 1 of the present invention.
- the echo canceller includes a speaker 101, a microphone 102, an adaptive filter unit 103, a received single talk detecting unit (hereinafter referred to as a received ST detecting unit) 104, first and second power calculating units 105 and 106, a power comparing unit 107, a sound A coupling amount estimation unit 108, a residual echo power estimation unit 109, a signal-to-echo ratio estimation unit 110, an amplitude suppression coefficient determination unit 111, and an amplitude suppression unit 112 are provided.
- r in (k) is a received signal
- s in (k) is a transmission input signal
- k represents time.
- the speaker 101 outputs the received reception signal r in (k) as a reception voice.
- the microphone 102 collects the received voice output from the speaker 101 as an echo, and generates a transmission input signal s in (k) in which the transmitted voice and the echo emitted by the speaker are mixed.
- the adaptive filter unit 103 receives a transmission input signal s in (k) from the microphone 102, sent from the transmitting input signal by linear echo processing s in (k) by removing the echo intermediate signal s mid (k) of Generate. Note that the adaptive filter unit 103 cannot completely remove the echo, and the transmission intermediate signal s mid (k) includes a residual echo.
- the adaptive filter unit 103 may use a generally well-known NLMS (Normalized Least Mean Square) algorithm or the like.
- the reception ST detection unit 104 uses the received signal r in (k) and the transmission intermediate signal s mid (k) including the residual echo input from the adaptive filter unit 103 as a single talk state where the call state is the reception side. That is, a state in which the transmission input signal is almost occupied by echoes is detected. Then, the received ST detection unit 104 outputs a detection flag RST (k) to the acoustic coupling amount estimation unit 108 as a detection result of the single talk state. In the detection flag RST (k), 0 indicates a state other than the received single talk, and 1 indicates the received single talk state. However, it is assumed that the reception ST detection unit 104 determines the reception single talk state only when it can be clearly determined as the reception single talk state.
- the reception ST detection unit 104 ensures that, for example, the reception signal r in (k) is sound and the signal power of the transmission intermediate signal s mid (k) is lower than a predetermined threshold. Only a situation in which there is no speaker voice is detected as a received single talk. In the case of this configuration, it is assumed that second power calculation section 106 outputs the power of the transmission intermediate signal to reception ST detection section 104.
- the first power calculation unit 105 calculates reception signal power for the reception signal r in (k), and the second power calculation unit 106 calculates transmission intermediate signal power for the transmission intermediate signal s mid (k).
- the power comparator 107 receives the received signal power and the transmitted intermediate signal power from the first and second power calculators 105 and 106, and calculates the power ratio sr (k) that is the ratio between them according to the following equation (1). To do.
- the acoustic coupling amount estimation unit 108 receives the detection flag RST (k) from the reception ST detection unit 104.
- RST (k) 1
- the acoustic coupling amount estimation unit 108 is estimated by the following equation (2).
- the coupling amount h (k) is updated. However, (0 ⁇ ⁇ 1).
- h (k) ⁇ ⁇ sr (k) + (1 ⁇ ) h (k ⁇ 1) (2)
- the residual echo power estimation unit 109 receives the received signal power from the first power calculation unit 105 and the estimated acoustic coupling amount from the acoustic coupling amount estimation unit 108.
- the residual echo power included in the transmission intermediate signal according to the following equation (3) The estimated residual echo power is estimated by estimating the power of.
- the residual echo power estimation unit 109 calculates the current residual echo calculated from the estimated acoustic coupling amount and the received signal power according to the following equation (4) instead of the following equation (3).
- the estimated residual echo power may be calculated by adding an estimated value of the past residual echo power obtained by multiplying the estimated value of power by a predetermined coefficient ⁇ .
- the signal-to-echo ratio estimation unit 110 receives the transmission intermediate signal power from the second power calculation unit 106 and the estimated residual echo power from the residual echo power estimation unit 109, respectively, and the signal according to the following definition equation (5)
- the echo ratio ser (k) is calculated.
- a is an arbitrary constant with a> 0.
- a is inserted to prevent the denominator of ser (k) from becoming zero when there is no echo and the estimated residual echo power, which is the denominator, is zero. It is assumed that the estimated residual echo power when there is a sound including an echo is set in advance to be a sufficiently small value.
- the amplitude suppression coefficient determination unit 111 receives the signal-to-echo ratio ser (k) from the signal-to-echo ratio estimation unit 110 and determines the amplitude suppression coefficient g (k) according to the following equation (6).
- b and c are constant parameters whose values are set in advance, and satisfy b> 1 and c> 0, respectively.
- an arbitrary lower limit value may be provided for the amplitude suppression coefficient g (k).
- the constant c is a parameter that controls the gradient of the amplitude suppression gain with respect to the log signal to echo ratio.
- FIG. 2 is a graph showing a linear relationship between the amplitude suppression gain G (k) and the log signal to echo ratio SER (k) defined by the above equation (8), and the amplitude suppression gain characteristic with respect to the log signal to echo ratio axis. 1 is shown.
- the linear relationship between the amplitude suppression gain G (k) and the log signal to echo ratio SER (k) continues even in the region where SER (k) ⁇ 0 dB.
- the power of the residual echo cannot exceed the transmission intermediate signal power, and SER (k)> 0 dB should always be satisfied.
- the estimated residual echo power exceeds the transmission intermediate signal power due to the estimation error of the acoustic coupling amount, and the calculation result is SER (k) ⁇ 0 dB.
- the echo cancellation apparatus of the present embodiment has a linear relationship between the amplitude suppression gain G (k) and the log signal-to-echo ratio SER (k) even in the range of SER (k) ⁇ 0 dB.
- the amplitude suppression gain characteristic is defined (Equation (8)).
- the amplitude suppression coefficient determination unit 111 then converts the expression (8) into an amplitude suppression coefficient g (() according to the above expression (6) in which the relational expression between the amplitude suppression coefficient g (k) and the signal-to-echo ratio ser (k) is corrected. k) is determined.
- the amplitude suppression unit 112 performs nonlinear echo processing by multiplying the transmission intermediate signal s mid (k) by the amplitude suppression coefficient g (k) as shown in the following equation (9), and is included in the transmission intermediate signal s mid (k).
- a transmission output signal s out (k) in which the residual echo is suppressed is generated.
- s out (k) g (k) ⁇ s mid (k) (9)
- the echo cancellation apparatus of the present embodiment is configured so that the amplitude suppression gain G (k) and SER (k) have a linear relationship even in the range of the log signal to echo ratio SER (k) ⁇ 0 dB.
- An amplitude suppression gain characteristic is defined, the amplitude suppression coefficient determination unit 111 determines the amplitude suppression coefficient g (k) according to this definition, and the amplitude suppression unit 112 suppresses the residual echo. Therefore, even when SER (k) ⁇ 0 dB due to the estimation error, it is possible to prevent excessive suppression from acting on the transmission intermediate signal, and to reduce the loss of transmitted voice.
- the echo cancellation apparatus includes the adaptive filter unit 103 that removes the echo from the transmission input signal in which the transmission voice and the echo are mixed and outputs the transmission intermediate signal, the received signal,
- a reception ST detection unit 104 that detects a single talk state on the reception side using a transmission intermediate signal, a first power calculation unit 105 that calculates reception signal power, and a second power calculation unit that calculates transmission intermediate signal power 106, a power comparison unit 107 that calculates the ratio between the transmission intermediate signal power and the reception signal power, and an estimation from the power ratio between the transmission intermediate signal power and the reception signal power when the received ST detection unit 104 detects a single talk state
- An acoustic coupling amount estimation unit 108 that estimates and updates the acoustic coupling amount, and an estimated residual echo power of the transmission intermediate signal from the received signal power and the estimated acoustic coupling amount
- the residual echo power estimation unit 109 to estimate, the signal to echo ratio estimation unit 110 to estimate the signal to echo ratio from the ratio of the transmission intermediate
- the amplitude suppression coefficient determination unit 111 is configured to determine the amplitude suppression coefficient according to the signal-to-echo ratio based on the above equation (6). For this reason, even when the estimated residual echo power exceeds the transmission intermediate signal power due to the estimation error, it is possible to prevent excessive amplitude suppression from acting on the transmission intermediate signal and to reduce the loss of transmitted voice.
- the residual echo power estimation unit 109 adds the residual echo power estimated at the past to the residual echo power estimated from the power of the received signal at this time and the acoustic coupling amount according to the above equation (4). It was configured to have echo power. Therefore, it is possible to estimate the estimated residual echo power considering the echo reverberation component.
- the amplitude suppression coefficient determination unit 111 has a linear relationship between the log signal to echo ratio SER (k) and the amplitude suppression gain G (k) in the range of SER (k) ⁇ 0 dB.
- the amplitude suppression gain G (k) has been determined so that In the echo canceller of this embodiment, this relational expression is defined using a mathematical function in which the amplitude suppression gain G (k) draws a saturation curve with respect to the infinity of the logarithmic signal to echo ratio SER (k) axis. .
- a relational expression between the log signal to echo ratio SER (k) and the amplitude suppression gain G (k) is defined as the following expression (10).
- b, c, and d are constant parameters having values set in advance, and b and c satisfy b> 1 and c> 0, respectively.
- FIG. 3 is a graph showing a characteristic curve of the amplitude suppression gain G (k) defined by the above equation (10). As shown in FIG. 3, the amplitude suppression gain G (k) draws a saturation curve with respect to the infinity of the logarithmic signal-to-echo ratio SER (k) axis.
- the log signal to echo ratio SER (k) is low, that is, the transmission input signal is in an echo only state, or the log signal to echo ratio SER (k) is high, that is, the transmission input signal is In the state where only the speaker voice is present, the fluctuation of the amplitude suppression gain G (k) is small, so even if there is a slight error in the log signal-to-echo ratio SER (k), the operation of the echo canceller is not affected. There is an advantage of being stable.
- the constant b is a parameter for controlling the minimum value of the amplitude suppression gain G (k).
- the constant c is a parameter for controlling the gradient of the amplitude suppression gain G (k) with respect to the log signal / echo ratio SER (k).
- the constant d is a parameter for controlling the center position of the transition region of the amplitude suppression gain G (k) by translating the characteristic curve of FIG. 3 along the SER (k) axis direction.
- the amplitude suppression coefficient determination unit 111 performs the amplitude suppression coefficient g (k) according to the following expression (11) obtained by changing the above expression (10) to the relational expression between the amplitude suppression coefficient g (k) and the signal-to-echo ratio ser (k). ). Since the components of the echo canceller other than the amplitude suppression coefficient determination unit 111 are the same as those in the first embodiment, description thereof is omitted.
- the amplitude suppression coefficient determination unit 111 is configured to determine the amplitude suppression coefficient according to the signal-to-echo ratio based on the above equation (11). Therefore, since the amplitude suppression gain characteristic is a saturation curve, the amplitude suppression gain is not easily affected by the error of the log signal to echo ratio, and the operation of the echo canceller can be stabilized.
- the amplitude suppression coefficient determination unit 111 is configured to determine the amplitude suppression coefficient using an equation. However, in the present embodiment, a plurality of inflection points are formed in the characteristic space. And the amplitude suppression coefficient is determined from characteristic data defining the amplitude suppression gain G (k) using a straight line passing through these inflection points or an approximate curve.
- FIG. 4 is a block diagram showing a configuration of an echo cancellation apparatus according to Embodiment 3 of the present invention.
- the echo canceller includes a speaker 101, a microphone 102, an adaptive filter unit 103, an incoming ST detection unit 104, a first power calculation unit 105, a second power calculation unit 106, a power comparison unit 107, an acoustic coupling amount estimation unit 108, A residual echo power estimation unit 109, a signal-to-echo ratio estimation unit 110, an amplitude suppression coefficient determination unit 411, an amplitude suppression gain characteristic setting unit 412, and an amplitude suppression unit 112 are provided.
- FIG. 4 the same or equivalent parts as in FIG.
- the amplitude suppression gain characteristic setting unit 412 stores characteristic data of the amplitude suppression gain G (k) with respect to the log signal to echo ratio SER (k).
- FIG. 5 is an example of amplitude suppression gain G (k) characteristic data stored in the amplitude suppression gain characteristic setting unit 412.
- the relationship between the signal-to-echo ratio ser (k) and the amplitude suppression coefficient g (k) is logarithmic. It is defined by expression.
- the amplitude suppression gain characteristic setting unit 412 defines a plurality of inflection points set in the characteristic space as shown in FIG. 5 by a straight line or a quadratic or higher function that individually connects these points. Characteristic data representing a characteristic curve defined by a curve or an approximate curve connecting all points is stored.
- the designer sets an arbitrary inflection point in the amplitude suppression gain characteristic setting unit 412, generates a characteristic curve, and stores it as amplitude suppression gain G (k) characteristic data.
- the amplitude suppression gain characteristic can be arbitrarily adjusted.
- the amplitude suppression coefficient determination unit 411 calculates a log signal to echo ratio SER (k) based on the signal to echo ratio ser (k) input from the signal to echo ratio estimation unit 110. Then, the amplitude suppression coefficient determination unit 411 reads the amplitude suppression gain G (k) corresponding to the log signal-to-echo ratio SER (k) from the amplitude suppression gain characteristic setting unit 412, and reads this from the amplitude suppression coefficient g (k). And output to the amplitude suppression unit 112.
- the echo canceller has an amplitude suppression gain characteristic in which the relationship between the signal-to-echo ratio and the amplitude suppression coefficient is defined, and a plurality of variable variables arbitrarily set on the characteristic space.
- An amplitude suppression gain characteristic setting unit 412 that holds as amplitude suppression gain characteristic data defined by a straight line or an approximate curve connecting points is provided.
- An amplitude suppression coefficient corresponding to the signal-to-echo ratio is determined based on the gain characteristic data. For this reason, it is possible to freely adjust the amplitude suppression gain characteristic.
- FIG. 6 is a block diagram showing a configuration of an echo cancellation apparatus according to Embodiment 4 of the present invention.
- the echo canceller includes a speaker 101, a microphone 102, an adaptive filter unit 103, an incoming ST detection unit 104, a first power spectrum calculation unit 605, a second power spectrum calculation unit 606, a power comparison unit 607, and an acoustic coupling amount estimation unit.
- residual echo power spectrum estimation unit 609 residual echo power spectrum estimation unit 609, signal-to-echo spectrum ratio estimation unit 610, amplitude suppression coefficient spectrum determination unit 611, amplitude suppression unit 612, first frequency conversion unit 613, second frequency conversion unit 614, inverse frequency A conversion unit 615 is provided. 6 that are the same as or equivalent to those in FIG. 1 are assigned the same reference numerals, and descriptions thereof are omitted.
- the first frequency conversion unit 613 converts the received reception signal r in (k) into a frequency spectrum to obtain a reception signal spectrum R ( ⁇ , k). ⁇ represents frequency and k represents time. Further, the second frequency conversion unit 614 converts the transmission intermediate signal s mid (k) input from the adaptive filter unit 103 into a frequency spectrum to obtain a transmission intermediate signal spectrum S ( ⁇ , k). For the frequency conversion processing of the first and second frequency conversion units 613 and 614, FFT (Fast Fourier Transform) is used.
- FFT Fast Fourier Transform
- the first power spectrum calculation unit 605 calculates a reception signal power spectrum from the reception signal spectrum R ( ⁇ , k) input from the first frequency conversion unit 613.
- the second power spectrum calculation unit 606 calculates a transmission intermediate signal power spectrum from the transmission intermediate signal spectrum S ( ⁇ , k) input from the second frequency conversion unit 614.
- the power comparison unit 607 receives the reception signal power spectrum and the transmission intermediate signal power spectrum from the first and second power spectrum calculation units 605 and 606, and according to the following equation (12), the power ratio sr ( ⁇ , k) is calculated.
- the acoustic coupling amount estimation unit 608 receives a detection flag RST (k) that is a detection result of the single talk state from the reception ST detection unit 104, and when RST (k) is 1, that is, in the reception single talk state,
- the estimated acoustic coupling amount H ( ⁇ , k) is updated by the following equation (13). However, (0 ⁇ ⁇ 1).
- H ( ⁇ , k) ⁇ ⁇ sr ( ⁇ , k) + (1 ⁇ ) H ( ⁇ , k ⁇ 1) (13)
- the residual echo power spectrum estimation unit 609 receives the received signal power spectrum from the first power spectrum calculation unit 605 and the estimated acoustic coupling amount from the acoustic coupling amount estimation unit 608, and estimates the residual echo power spectrum according to the following equation (14). Is calculated. When the echo reverberation component is considered, the residual echo power spectrum estimation unit 609 replaces the following equation (14) with the following equation (15) and the current residual that is calculated from the estimated acoustic coupling amount and the received power spectrum.
- the estimated residual echo power spectrum may be calculated by giving an estimated value of the residual echo power spectrum in the past by multiplying the estimated value of the echo power spectrum by a predetermined coefficient ⁇ .
- the signal-to-echo spectrum ratio estimation unit 610 receives the transmission intermediate signal power spectrum from the second power spectrum calculation unit 606 and the estimated residual echo power spectrum from the residual echo power spectrum estimation unit 609, respectively, and the signal-to-echo spectrum ratio ser ( ⁇ , k) is calculated according to the following defining formula (16).
- a is an arbitrary constant with a> 0.
- the amplitude suppression coefficient spectrum determination unit 611 receives the signal-to-echo spectrum ratio ser ( ⁇ , k) from the signal-to-echo spectrum ratio estimation unit 610 and uses any one of the above-described first to third embodiments to determine the amplitude.
- a suppression coefficient spectrum g ( ⁇ , k) is determined.
- the amplitude suppression coefficient spectrum determination unit 611 performs amplitude according to the following equation (17), and when the method of the second embodiment is applied, according to the following equation (18).
- a suppression coefficient spectrum g ( ⁇ , k) is determined.
- the echo canceller defines the characteristic of the amplitude suppression gain G ( ⁇ , k) corresponding to the log signal-to-echo spectrum ratio SER ( ⁇ , k).
- An amplitude suppression gain characteristic setting unit that holds amplitude suppression gain G ( ⁇ , k) characteristic data is provided, and the amplitude suppression coefficient spectrum determination unit 611 reads the amplitude suppression gain G ( ⁇ , k) from the amplitude suppression gain characteristic setting unit. This is converted into an amplitude suppression coefficient spectrum g ( ⁇ , k).
- the amplitude suppression coefficient spectrum determination unit 611 determines the amplitude suppression coefficient spectrum g ( ⁇ , k) using the methods of the first to third embodiments, the effects obtained by using each method are as follows. Needless to say, it can also be obtained in the embodiment.
- the amplitude suppression unit 612 performs nonlinear echo processing by multiplying the transmission intermediate signal spectrum S ( ⁇ , k) by the amplitude suppression coefficient spectrum g ( ⁇ , k), and generates a transmission output signal spectrum S out ( ⁇ , k). To do. Then, the inverse frequency conversion unit 615 receives the transmission output signal spectrum S out ( ⁇ , k) from the amplitude suppression unit 612 and performs frequency inverse conversion to generate the transmission output signal s out (k).
- the amplitude suppression coefficient spectrum determining unit 611 calculates the amplitude suppression coefficient spectrum g ( ⁇ , k) that is different for each band in the frequency domain, thereby suppressing the frequency component mainly including the echo component in the transmission input signal.
- the frequency component mainly including the transmitted voice is stored.
- the echo canceller of the fourth embodiment further reduces the loss of speaker speech included in the transmission input signal.
- the echo cancellation apparatus includes the adaptive filter unit 103 that removes the echo from the transmission input signal in which the transmission voice and the echo are mixed and outputs the transmission intermediate signal, the received signal,
- a reception ST detection unit 104 that detects a single talk state on the reception side using a transmission intermediate signal, a first frequency conversion unit 613 that converts a reception signal into a reception signal spectrum that is a frequency spectrum, and a power spectrum from the reception signal spectrum
- a first power spectrum calculation unit 605 that calculates the transmission intermediate signal
- a second frequency conversion unit 614 that converts the transmission intermediate signal into a transmission intermediate signal spectrum that is a frequency spectrum, and a second power spectrum that is calculated from the transmission intermediate signal spectrum
- a power comparison unit 607 that calculates the amount of acoustic coupling and a sound coupling amount that estimates and updates the estimated acoustic coupling amount from the ratio of the transmission intermediate signal power spectrum and the reception signal power spectrum
- the echo cancellation apparatus determines the amplitude suppression coefficient used for the nonlinear echo processing based on the signal-to-echo ratio, the echo cancellation apparatus with little auditory deterioration of the speaker voice And is suitable for use in an echo canceling apparatus such as a hands-free telephone having a low ratio of transmitted voice to echo.
Landscapes
- Engineering & Computer Science (AREA)
- Signal Processing (AREA)
- Computer Networks & Wireless Communication (AREA)
- Cable Transmission Systems, Equalization Of Radio And Reduction Of Echo (AREA)
- Telephone Function (AREA)
Abstract
Description
特許文献1に開示の反響抑圧装置は、送話音声とエコーとが重畳した送信信号のパワーからエコー信号のパワーを減算することによって送話音声のみのパワーを算出し、この送話音声パワーに所定のマスキング閾値を加え、さらに送信信号のパワー値で正規化した値を、抑制係数としていた。反響抑圧装置は、この抑制係数を用いて非線形エコー処理を行うことにより、送話音声が存在するダブルトーク状態では自動的に抑圧が弱められ、送話音声が損なわれにくくなっていた。なお、エコー信号のパワーは、受信信号のパワーとエコー経路の音響結合量との積によって推定されていた。
特に、ハンズフリー電話等の送話音声対エコー比率の非常に悪い環境では、この誤差の影響が顕著であり、送信信号のパワーからエコー信号のパワーの推定値を減算して送話信号のパワーを算出するときに、送話音声パワーの推定値が0以下になってしまうことがあった。この結果、非線形エコー処理において過大なエコー抑圧が働き、送話音声が損なわれてしまうという問題が起こっていた。
なお、特許文献1を例にあげて説明した従来の方法において送話音声が損なわれる課題を回避するためには、前述のマスキング閾値を非常に高く設定してエコー抑圧を軽減させる必要がある。しかし、この場合には弊害としてエコーが充分に抑えられなくなるという問題が起こる。
実施の形態1.
図1は、この発明の実施の形態1に係るエコー消去装置の構成を示すブロック図である。エコー消去装置は、スピーカ101、マイク102、適応フィルタ部103、受話シングルトーク検出部(以下、受話ST検出部)104、第1および第2のパワー算出部105,106、パワー比較部107、音響結合量推定部108、残留エコーパワー推定部109、信号対エコー比推定部110、振幅抑圧係数決定部111、振幅抑圧部112を備える。図1において、rin(k)は受信信号、sin(k)は送信入力信号であり、kは時刻を表す。
スピーカ101は、受信した受信信号rin(k)を受話音声として出力する。マイク102は、スピーカ101が出力した受話音声をエコーとして収音し、話者が発する送話音声とエコーとが混在した送信入力信号sin(k)を生成する。続いて、適応フィルタ部103が、マイク102から送信入力信号sin(k)を受け、線形エコー処理により送信入力信号sin(k)からエコーを除去して送信中間信号smid(k)を生成する。
なお、適応フィルタ部103はエコーを完全に除去することはできず、送信中間信号smid(k)には残留エコーが含まれる。適応フィルタ部103には、一般的によく知られるNLMS(Normalized Least Mean Square)アルゴリズム等を用いればよい。
検知フラグRST(k)は、0が受話シングルトーク以外の状態を示し、1が受話シングルトーク状態を示す。ただし、受話ST検出部104は、明確に受話シングルトーク状態と判断できる状態の場合のみ、受話シングルトーク状態と判定するものとする。
h(k)=α・sr(k)+(1-α)h(k-1) (2)
なお、エコーの残響成分を考慮する場合には、残留エコーパワー推定部109が下記式(3)に代えて下記式(4)に従って、推定音響結合量と受信信号パワーから算出する現時点の残留エコーパワーの推定値に、所定の係数βを乗じた過去時の残留エコーパワーの推定値を加えることによって、推定残留エコーパワーを算出してもよい。
G(k)=20log10(g(k)) (7)
SER(k)=10log10(ser(k))
sout(k)=g(k)・smid(k) (9)
上記実施の形態1のエコー消去装置において、振幅抑圧係数決定部111は、対数信号対エコー比SER(k)と振幅抑圧ゲインG(k)とがSER(k)<0dBの範囲で線形関係となるように振幅抑圧ゲインG(k)を決定していた。本実施の形態のエコー消去装置では、この関係式を、振幅抑圧ゲインG(k)が対数信号対エコー比SER(k)軸の無限遠に対して飽和曲線を描く数学関数を用いて定義する。
上記実施の形態1および2のエコー消去装置では、振幅抑圧係数決定部111が式を用いて振幅抑圧係数を決定する構成であったが、本実施の形態では、特性空間上に複数の変節点を設定してこれら変節点を通る直線、または近似曲線を用いて振幅抑圧ゲインG(k)を定義した特性データから、振幅抑圧係数を決定する構成とする。
図5は、振幅抑圧ゲイン特性設定部412が保存している振幅抑圧ゲインG(k)特性データの一例であり、信号対エコー比ser(k)と振幅抑圧係数g(k)の関係を対数表現で定義したものである。振幅抑圧ゲイン特性設定部412には、図5に示すように特性空間上に設定された複数の変節点に対して、これらの点同士を個々に結ぶ直線もしくは2次以上の関数で定義される曲線、または全ての点を結ぶ近似曲線によって定義された特性曲線を表す特性データが保存されている。本実施の形態のエコー消去装置では、設計者が振幅抑圧ゲイン特性設定部412に任意の変節点を設定し、特性曲線を生成して振幅抑圧ゲインG(k)特性データとして保存しておくことによって、任意に振幅抑圧ゲイン特性を調節できる。
上記実施の形態1~3のエコー消去装置は、送信中間信号を時間領域において振幅抑圧する構成としたが、本実施の形態のエコー消去装置では周波数領域において振幅抑圧する構成とする。図6は、この発明の実施の形態4に係るエコー消去装置の構成を示すブロック図である。エコー消去装置は、スピーカ101、マイク102、適応フィルタ部103、受話ST検出部104、第1のパワースペクトル算出部605、第2のパワースペクトル算出部606、パワー比較部607、音響結合量推定部608、残留エコーパワースペクトル推定部609、信号対エコースペクトル比推定部610、振幅抑圧係数スペクトル決定部611、振幅抑圧部612、第1の周波数変換部613、第2の周波数変換部614、逆周波数変換部615を備える。なお、図6において図1と同一または相当の部分については同一の符号を付し説明を省略する。
また、第2の周波数変換部614は、適応フィルタ部103から入力された送信中間信号smid(k)を、周波数スペクトルに変換し、送信中間信号スペクトルS(ω,k)とする。第1および第2の周波数変換部613,614の周波数変換処理には、FFT(Fast Fourier Transform)を用いる。
H(ω,k)=α・sr(ω,k)+(1-α)H(ω,k-1) (13)
なお、エコーの残響成分を考慮する場合には、残留エコーパワースペクトル推定部609が下記式(14)に代えて下記式(15)に従って、推定音響結合量と受信パワースペクトルから算出する現時点の残留エコーパワースペクトルの推定値に、所定の係数βを乗じた過去時の残留エコーパワースペクトルの推定値を与えることによって、推定残留エコーパワースペクトルを算出してもよい。
振幅抑圧係数スペクトル決定部611は、実施の形態1の方法を適用した場合には下記式(17)に従って、また、実施の形態2の方法を適用した場合には下記式(18)に従って、振幅抑圧係数スペクトルg(ω,k)を決定する。
なお、振幅抑圧係数スペクトル決定部611が上記実施の形態1~3の方法を用いて振幅抑圧係数スペクトルg(ω,k)を決定する場合には、各方法を用いることにより得られる効果が本実施の形態においても得られることは言うまでもない。
Claims (10)
- 送話音声とエコーが混在した送信入力信号からエコーを除去して送信中間信号を出力する適応フィルタ部と、
受信信号と前記送信中間信号を用いて、受話側のシングルトーク状態を検知する受話シングルトーク検出部と、
前記受信信号のパワーを算出する第1のパワー算出部と、
前記送信中間信号のパワーを算出する第2のパワー算出部と、
前記受話シングルトーク検出部によってシングルトーク状態が検知された場合に、前記送信中間信号と前記受信信号のパワー比から音響結合量を推定して更新する音響結合量推定部と、
前記受信信号のパワーと前記音響結合量から、前記送信中間信号の残留エコーパワーを推定する残留エコーパワー推定部と、
前記送信中間信号のパワーと前記残留エコーパワーの比から、信号対エコー比を推定する信号対エコー比推定部と、
前記信号対エコー比に応じて振幅抑圧係数を決定する振幅抑圧係数決定部と、
前記振幅抑圧係数に基づいて前記送信中間信号を振幅抑圧する振幅抑圧部とを備えるエコー消去装置。 - 信号対エコー比と振幅抑圧係数の関係を定義した振幅抑圧ゲイン特性を、特性空間上に任意に設定された複数の変節点を結ぶ直線または近似曲線で定義する振幅抑圧ゲイン特性データとして保持する振幅抑圧ゲイン特性設定部を備え、
振幅抑圧係数決定部は、前記振幅抑圧ゲイン特性設定部に保持された前記振幅抑圧ゲイン特性データに基づいて、信号対エコー比に応じた振幅抑圧係数を決定することを特徴とする請求項1記載のエコー消去装置。
- 残留エコーパワー推定部は、現時点の受信信号のパワーと音響結合量から推定した残留エコーパワーに、過去時に推定した残留エコーパワーを加えて送信中間信号の残留エコーパワーとすることを特徴とする請求項1記載のエコー消去装置。
- 送話音声とエコーが混在した送信入力信号からエコーを除去して送信中間信号を出力する適応フィルタ部と、
受信信号と前記送信中間信号を用いて、受話側のシングルトーク状態を検知する受話シングルトーク検出部と、
前記受信信号を周波数スペクトルに変換する第1の周波数変換部と、
前記受信信号の周波数スペクトルからパワースペクトルを算出する第1のパワースペクトル算出部と、
前記送信中間信号を周波数スペクトルに変換する第2の周波数変換部と、
前記送信中間信号の周波数スペクトルからパワースペクトルを算出する第2のパワースペクトル算出部と、
前記受話シングルトーク検出部によってシングルトーク状態が検知された場合に、前記送信中間信号と前記受信信号のパワースペクトル比から音響結合量を推定して更新する音響結合量推定部と、
前記受信信号のパワースペクトルと前記音響結合量から、前記送信中間信号の残留エコーパワースペクトルを推定する残留エコーパワースペクトル推定部と、
前記送信中間信号のパワースペクトルと前記残留エコーパワースペクトルの比から、信号対エコースペクトル比を推定する信号対エコースペクトル比推定部と、
前記信号対エコースペクトル比に応じて振幅抑圧係数スペクトルを決定する振幅抑圧係数スペクトル決定部と、
前記振幅抑圧係数スペクトルに基づいて前記送信中間信号の周波数スペクトルを振幅抑圧する振幅抑圧部と、
前記振幅抑圧された前記送信中間信号の周波数スペクトルを時系列信号に変換する逆周波数変換部とを備えるエコー消去装置。 - 信号対エコースペクトル比と振幅抑圧係数スペクトルの関係を定義した振幅抑圧ゲイン特性を、特性空間上に任意に設定された複数の変節点を結ぶ直線または近似曲線で定義する振幅抑圧ゲイン特性データとして保持する振幅抑圧ゲイン特性設定部を備え、
振幅抑圧係数スペクトル決定部は、前記振幅抑圧ゲイン特性設定部に保持された前記振幅抑圧ゲイン特性データに基づいて、信号対エコースペクトル比に応じた振幅抑圧係数スペクトルを決定することを特徴とする請求項6記載のエコー消去装置。 - 残留エコーパワースペクトル推定部は、現時点の受信信号のパワースペクトルと音響結合量から推定した残留エコーパワースペクトルに、過去時に推定した残留エコーパワースペクトルを加えて送信中間信号の残留エコーパワースペクトルとすることを特徴とする請求項6記載のエコー消去装置。
Priority Applications (5)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
JP2010530639A JP5036874B2 (ja) | 2008-09-24 | 2008-09-24 | エコー消去装置 |
EP08877055A EP2330752B1 (en) | 2008-09-24 | 2008-09-24 | Echo cancelling device |
US13/058,212 US8792649B2 (en) | 2008-09-24 | 2008-09-24 | Echo canceller used for voice communication |
CN2008801312499A CN102165707B (zh) | 2008-09-24 | 2008-09-24 | 回波消除装置 |
PCT/JP2008/002638 WO2010035308A1 (ja) | 2008-09-24 | 2008-09-24 | エコー消去装置 |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
PCT/JP2008/002638 WO2010035308A1 (ja) | 2008-09-24 | 2008-09-24 | エコー消去装置 |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2010035308A1 true WO2010035308A1 (ja) | 2010-04-01 |
Family
ID=42059326
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/JP2008/002638 WO2010035308A1 (ja) | 2008-09-24 | 2008-09-24 | エコー消去装置 |
Country Status (5)
Country | Link |
---|---|
US (1) | US8792649B2 (ja) |
EP (1) | EP2330752B1 (ja) |
JP (1) | JP5036874B2 (ja) |
CN (1) | CN102165707B (ja) |
WO (1) | WO2010035308A1 (ja) |
Cited By (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2012227566A (ja) * | 2011-04-14 | 2012-11-15 | Nippon Telegr & Teleph Corp <Ntt> | エコー消去装置、その方法及びプログラム |
WO2012153452A1 (ja) * | 2011-05-10 | 2012-11-15 | 三菱電機株式会社 | エコー消去装置およびエコー検出装置 |
EP2662855A4 (en) * | 2011-01-04 | 2015-12-16 | Fujitsu Ltd | VOICE CONTROL DEVICE, VOICE CONTROL METHOD, AND VOICE CONTROL PROGRAM |
JP2016025425A (ja) * | 2014-07-17 | 2016-02-08 | 沖電気工業株式会社 | エコー抑圧装置、エコー抑圧プログラム、エコー抑圧方法及び通信端末 |
JP2016189548A (ja) * | 2015-03-30 | 2016-11-04 | 日本電信電話株式会社 | エコー消去装置、その方法及びプログラム |
JP2017191992A (ja) * | 2016-04-12 | 2017-10-19 | 日本電信電話株式会社 | エコー抑圧装置、その方法、プログラム、及び記録媒体 |
WO2018221206A1 (ja) * | 2017-05-29 | 2018-12-06 | 株式会社トランストロン | エコー抑圧装置、エコー抑圧方法及びエコー抑圧プログラム |
CN109273019A (zh) * | 2017-04-21 | 2019-01-25 | 豪威科技股份有限公司 | 用于回声抑制的双重通话检测的方法及回声抑制 |
Families Citing this family (17)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP5156043B2 (ja) * | 2010-03-26 | 2013-03-06 | 株式会社東芝 | 音声判別装置 |
SG182859A1 (en) * | 2011-01-06 | 2012-08-30 | Creative Tech Ltd | A method, system and apparatus for processing audio signals |
JP5631523B2 (ja) | 2012-04-25 | 2014-11-26 | 三菱電機株式会社 | エコー消去装置 |
CN104050971A (zh) * | 2013-03-15 | 2014-09-17 | 杜比实验室特许公司 | 声学回声减轻装置和方法、音频处理装置和语音通信终端 |
JP6075783B2 (ja) * | 2013-12-09 | 2017-02-08 | 日本電信電話株式会社 | エコー消去装置、エコー消去方法及びプログラム |
US9319783B1 (en) * | 2014-02-19 | 2016-04-19 | Amazon Technologies, Inc. | Attenuation of output audio based on residual echo |
US10504501B2 (en) | 2016-02-02 | 2019-12-10 | Dolby Laboratories Licensing Corporation | Adaptive suppression for removing nuisance audio |
CN107045872B (zh) * | 2016-02-05 | 2020-09-01 | 中国电信股份有限公司 | 通话回声的识别方法和装置 |
US10542153B2 (en) * | 2017-08-03 | 2020-01-21 | Bose Corporation | Multi-channel residual echo suppression |
US10594869B2 (en) * | 2017-08-03 | 2020-03-17 | Bose Corporation | Mitigating impact of double talk for residual echo suppressors |
US10200540B1 (en) | 2017-08-03 | 2019-02-05 | Bose Corporation | Efficient reutilization of acoustic echo canceler channels |
EP3692704B1 (en) | 2017-10-03 | 2023-09-06 | Bose Corporation | Spatial double-talk detector |
US10192567B1 (en) * | 2017-10-18 | 2019-01-29 | Motorola Mobility Llc | Echo cancellation and suppression in electronic device |
US10964305B2 (en) | 2019-05-20 | 2021-03-30 | Bose Corporation | Mitigating impact of double talk for residual echo suppressors |
JP7235117B2 (ja) * | 2019-08-06 | 2023-03-08 | 日本電信電話株式会社 | エコー消去装置、エコー消去方法、プログラム |
CN110971769B (zh) * | 2019-11-19 | 2022-05-03 | 百度在线网络技术(北京)有限公司 | 通话信号的处理方法、装置、电子设备及存储介质 |
CN111968663B (zh) * | 2020-08-12 | 2024-04-16 | 福建星网智慧科技有限公司 | 一种回声残留判断方法 |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPH11331046A (ja) * | 1998-03-16 | 1999-11-30 | Nippon Telegr & Teleph Corp <Ntt> | エコー抑圧方法及び装置並びにエコー抑圧プログラムが記憶されたコンピュータに読取り可能な記憶媒体 |
JP2002084212A (ja) | 2000-09-08 | 2002-03-22 | Nippon Telegr & Teleph Corp <Ntt> | 反響抑圧方法、反響抑圧装置及び反響抑圧プログラム記憶媒体 |
JP2006033789A (ja) * | 2004-06-16 | 2006-02-02 | Nippon Telegr & Teleph Corp <Ntt> | 反響路結合量推定方法、反響路結合量推定装置、反響路結合量推定プログラム、エコー抑圧方法、エコー抑圧装置、エコー抑圧プログラム、エコーサプレス方法、エコーサプレッサ、エコーサプレッサプログラム、通信路上の損失量制御方法、通信路上の損失量制御装置、通信路上の損失制御プログラム、多チャネルエコー抑圧方法、多チャネルエコー抑圧装置、多チャネルエコー抑圧プログラム、記録媒体 |
Family Cites Families (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
EP0667700B1 (en) * | 1994-02-10 | 2004-01-07 | Nippon Telegraph And Telephone Corporation | Echo cancelling method and apparatus using fast projection scheme |
SE505692C2 (sv) * | 1995-12-18 | 1997-09-29 | Ericsson Telefon Ab L M | Förfarande och anordning för ekosläckning genom skattning av effekt hos residualsignal |
JP3404236B2 (ja) | 1996-12-26 | 2003-05-06 | 株式会社神戸製鋼所 | 拡声通話装置 |
JP2001044896A (ja) | 1999-08-03 | 2001-02-16 | Matsushita Electric Ind Co Ltd | 通話装置および通話方法 |
JP3566158B2 (ja) * | 1999-12-07 | 2004-09-15 | 三菱電機株式会社 | エコーキャンセラ装置 |
CN1162247C (zh) * | 2002-02-27 | 2004-08-18 | 王台钦 | 钢筋续接器植焊施工法 |
US7577248B2 (en) * | 2004-06-25 | 2009-08-18 | Texas Instruments Incorporated | Method and apparatus for echo cancellation, digit filter adaptation, automatic gain control and echo suppression utilizing block least mean squares |
JP4247203B2 (ja) | 2005-05-06 | 2009-04-02 | 日本電信電話株式会社 | 自動音量制御方法、自動音量制御装置、プログラム、記録媒体 |
JP4613705B2 (ja) * | 2005-06-15 | 2011-01-19 | 沖電気工業株式会社 | エコーキャンセラ |
JP4559989B2 (ja) | 2006-03-24 | 2010-10-13 | 日本電信電話株式会社 | エコー抑圧方法、エコー抑圧プログラム、記録媒体 |
-
2008
- 2008-09-24 JP JP2010530639A patent/JP5036874B2/ja not_active Expired - Fee Related
- 2008-09-24 EP EP08877055A patent/EP2330752B1/en not_active Not-in-force
- 2008-09-24 CN CN2008801312499A patent/CN102165707B/zh not_active Expired - Fee Related
- 2008-09-24 WO PCT/JP2008/002638 patent/WO2010035308A1/ja active Application Filing
- 2008-09-24 US US13/058,212 patent/US8792649B2/en not_active Expired - Fee Related
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPH11331046A (ja) * | 1998-03-16 | 1999-11-30 | Nippon Telegr & Teleph Corp <Ntt> | エコー抑圧方法及び装置並びにエコー抑圧プログラムが記憶されたコンピュータに読取り可能な記憶媒体 |
JP2002084212A (ja) | 2000-09-08 | 2002-03-22 | Nippon Telegr & Teleph Corp <Ntt> | 反響抑圧方法、反響抑圧装置及び反響抑圧プログラム記憶媒体 |
JP2006033789A (ja) * | 2004-06-16 | 2006-02-02 | Nippon Telegr & Teleph Corp <Ntt> | 反響路結合量推定方法、反響路結合量推定装置、反響路結合量推定プログラム、エコー抑圧方法、エコー抑圧装置、エコー抑圧プログラム、エコーサプレス方法、エコーサプレッサ、エコーサプレッサプログラム、通信路上の損失量制御方法、通信路上の損失量制御装置、通信路上の損失制御プログラム、多チャネルエコー抑圧方法、多チャネルエコー抑圧装置、多チャネルエコー抑圧プログラム、記録媒体 |
Non-Patent Citations (1)
Title |
---|
See also references of EP2330752A4 |
Cited By (17)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
EP2662855A4 (en) * | 2011-01-04 | 2015-12-16 | Fujitsu Ltd | VOICE CONTROL DEVICE, VOICE CONTROL METHOD, AND VOICE CONTROL PROGRAM |
US9271089B2 (en) | 2011-01-04 | 2016-02-23 | Fujitsu Limited | Voice control device and voice control method |
JP2012227566A (ja) * | 2011-04-14 | 2012-11-15 | Nippon Telegr & Teleph Corp <Ntt> | エコー消去装置、その方法及びプログラム |
US9286881B2 (en) | 2011-05-10 | 2016-03-15 | Mitsubishi Electric Corporation | Echo canceler and echo detector |
CN103329450B (zh) * | 2011-05-10 | 2015-02-11 | 三菱电机株式会社 | 回波消除装置以及回波检测装置 |
JP5501527B2 (ja) * | 2011-05-10 | 2014-05-21 | 三菱電機株式会社 | エコー消去装置およびエコー検出装置 |
CN103329450A (zh) * | 2011-05-10 | 2013-09-25 | 三菱电机株式会社 | 回波消除装置以及回波检测装置 |
WO2012153452A1 (ja) * | 2011-05-10 | 2012-11-15 | 三菱電機株式会社 | エコー消去装置およびエコー検出装置 |
JP2016025425A (ja) * | 2014-07-17 | 2016-02-08 | 沖電気工業株式会社 | エコー抑圧装置、エコー抑圧プログラム、エコー抑圧方法及び通信端末 |
JP2016189548A (ja) * | 2015-03-30 | 2016-11-04 | 日本電信電話株式会社 | エコー消去装置、その方法及びプログラム |
JP2017191992A (ja) * | 2016-04-12 | 2017-10-19 | 日本電信電話株式会社 | エコー抑圧装置、その方法、プログラム、及び記録媒体 |
CN109273019A (zh) * | 2017-04-21 | 2019-01-25 | 豪威科技股份有限公司 | 用于回声抑制的双重通话检测的方法及回声抑制 |
CN109273019B (zh) * | 2017-04-21 | 2023-06-16 | 豪威科技股份有限公司 | 用于回声抑制的双重通话检测的方法及回声抑制 |
WO2018221206A1 (ja) * | 2017-05-29 | 2018-12-06 | 株式会社トランストロン | エコー抑圧装置、エコー抑圧方法及びエコー抑圧プログラム |
JP2018201147A (ja) * | 2017-05-29 | 2018-12-20 | 株式会社トランストロン | エコー抑圧装置、エコー抑圧方法及びエコー抑圧プログラム |
RU2732362C1 (ru) * | 2017-05-29 | 2020-09-16 | Транстрон Инк. | Устройство эхоподавления, способ эхоподавления и программа эхоподавления |
US11039245B2 (en) | 2017-05-29 | 2021-06-15 | Transtron Inc. | Echo suppression device, echo suppression method and echo suppression program |
Also Published As
Publication number | Publication date |
---|---|
JPWO2010035308A1 (ja) | 2012-02-16 |
EP2330752A1 (en) | 2011-06-08 |
US20110135105A1 (en) | 2011-06-09 |
US8792649B2 (en) | 2014-07-29 |
JP5036874B2 (ja) | 2012-09-26 |
EP2330752A4 (en) | 2012-02-22 |
EP2330752B1 (en) | 2013-03-20 |
CN102165707B (zh) | 2013-12-04 |
CN102165707A (zh) | 2011-08-24 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
JP5036874B2 (ja) | エコー消去装置 | |
US7856097B2 (en) | Echo canceling apparatus, telephone set using the same, and echo canceling method | |
JP5347794B2 (ja) | エコー抑圧方法およびその装置 | |
US8311234B2 (en) | Echo canceller and communication audio processing apparatus | |
US9723152B2 (en) | Nonlinear echo suppression | |
US20030039353A1 (en) | Echo cancellation processing system | |
US8160239B2 (en) | Echo canceller and speech processing apparatus | |
CN109273019B (zh) | 用于回声抑制的双重通话检测的方法及回声抑制 | |
US20110261950A1 (en) | Adaptive filter and echo canceller having this adaptive filter | |
JP4978352B2 (ja) | エコーキャンセラ | |
US9361874B2 (en) | Varying adaptive filter step size in acoustic echo cancellation | |
US20170310360A1 (en) | Echo removal device, echo removal method, and non-transitory storage medium | |
JP4607015B2 (ja) | エコー抑圧装置 | |
US8369511B2 (en) | Robust method of echo suppressor | |
JP4475155B2 (ja) | エコーキャンセラ | |
KR20090122802A (ko) | 스펙트럼 차감을 이용한 음향학적 반향 제거 방법 및 장치 | |
JP6180689B1 (ja) | エコーキャンセラ装置、エコー消去方法、及びエコー消去プログラム | |
JP2007221219A (ja) | ハウリング防止方法、装置、プログラム、及びこのプログラムを記録した記録媒体 | |
WO2022202012A1 (ja) | エコー抑圧装置、エコー抑圧方法及びエコー抑圧プログラム | |
JP6369192B2 (ja) | エコー抑圧装置、エコー抑圧プログラム、エコー抑圧方法及び通信端末 | |
KR100431965B1 (ko) | 시변 적응알고리즘이 적용된 음향반향 제거장치 및 그 방법 | |
JP6314608B2 (ja) | エコー抑圧装置、エコー抑圧プログラム、及びエコー抑圧方法 | |
US20120224684A1 (en) | Soft attenuation of high-power signals |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
WWE | Wipo information: entry into national phase |
Ref document number: 200880131249.9 Country of ref document: CN |
|
121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 08877055 Country of ref document: EP Kind code of ref document: A1 |
|
WWE | Wipo information: entry into national phase |
Ref document number: 2010530639 Country of ref document: JP |
|
WWE | Wipo information: entry into national phase |
Ref document number: 13058212 Country of ref document: US |
|
WWE | Wipo information: entry into national phase |
Ref document number: 2008877055 Country of ref document: EP |
|
NENP | Non-entry into the national phase |
Ref country code: DE |