WO2010035658A1 - 信号処理方法、信号処理装置、および信号処理プログラム - Google Patents
信号処理方法、信号処理装置、および信号処理プログラム Download PDFInfo
- Publication number
- WO2010035658A1 WO2010035658A1 PCT/JP2009/066068 JP2009066068W WO2010035658A1 WO 2010035658 A1 WO2010035658 A1 WO 2010035658A1 JP 2009066068 W JP2009066068 W JP 2009066068W WO 2010035658 A1 WO2010035658 A1 WO 2010035658A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- signal
- received signal
- delayed
- signal processing
- received
- Prior art date
Links
- 238000003672 processing method Methods 0.000 title claims abstract description 22
- 238000012545 processing Methods 0.000 title claims description 178
- 230000003111 delayed effect Effects 0.000 claims abstract description 234
- 230000003044 adaptive effect Effects 0.000 claims abstract description 136
- 230000008859 change Effects 0.000 claims abstract description 56
- 230000004807 localization Effects 0.000 claims abstract description 41
- 238000002592 echocardiography Methods 0.000 claims abstract description 26
- 238000004458 analytical method Methods 0.000 claims description 48
- 238000012937 correction Methods 0.000 claims description 44
- 230000008825 perceptual sensitivity Effects 0.000 claims description 40
- 238000000034 method Methods 0.000 claims description 24
- 238000005070 sampling Methods 0.000 claims description 13
- 230000008569 process Effects 0.000 claims description 8
- 230000007274 generation of a signal involved in cell-cell signaling Effects 0.000 claims description 4
- 230000009467 reduction Effects 0.000 claims description 3
- 230000035945 sensitivity Effects 0.000 abstract description 2
- 238000010586 diagram Methods 0.000 description 27
- 230000015572 biosynthetic process Effects 0.000 description 19
- 230000014509 gene expression Effects 0.000 description 19
- 238000003786 synthesis reaction Methods 0.000 description 19
- 230000006866 deterioration Effects 0.000 description 17
- 230000006870 function Effects 0.000 description 17
- 230000001934 delay Effects 0.000 description 11
- 230000000694 effects Effects 0.000 description 9
- 238000006243 chemical reaction Methods 0.000 description 8
- 230000010363 phase shift Effects 0.000 description 6
- 230000005540 biological transmission Effects 0.000 description 5
- 230000009466 transformation Effects 0.000 description 4
- 238000004364 calculation method Methods 0.000 description 2
- 125000004122 cyclic group Chemical group 0.000 description 2
- 238000006073 displacement reaction Methods 0.000 description 2
- 230000036039 immunity Effects 0.000 description 2
- 230000006872 improvement Effects 0.000 description 2
- 230000001902 propagating effect Effects 0.000 description 2
- 230000002195 synergetic effect Effects 0.000 description 2
- 238000009825 accumulation Methods 0.000 description 1
- 230000000903 blocking effect Effects 0.000 description 1
- 230000000295 complement effect Effects 0.000 description 1
- 230000003247 decreasing effect Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000008447 perception Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 230000013707 sensory perception of sound Effects 0.000 description 1
- 238000000844 transformation Methods 0.000 description 1
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
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L21/00—Speech or voice signal processing techniques to produce another audible or non-audible signal, e.g. visual or tactile, in order to modify its quality or its intelligibility
- G10L21/02—Speech enhancement, e.g. noise reduction or echo cancellation
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04M—TELEPHONIC COMMUNICATION
- H04M3/00—Automatic or semi-automatic exchanges
- H04M3/42—Systems providing special services or facilities to subscribers
- H04M3/56—Arrangements for connecting several subscribers to a common circuit, i.e. affording conference facilities
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04R—LOUDSPEAKERS, MICROPHONES, GRAMOPHONE PICK-UPS OR LIKE ACOUSTIC ELECTROMECHANICAL TRANSDUCERS; DEAF-AID SETS; PUBLIC ADDRESS SYSTEMS
- H04R1/00—Details of transducers, loudspeakers or microphones
- H04R1/10—Earpieces; Attachments therefor ; Earphones; Monophonic headphones
- H04R1/1083—Reduction of ambient noise
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04R—LOUDSPEAKERS, MICROPHONES, GRAMOPHONE PICK-UPS OR LIKE ACOUSTIC ELECTROMECHANICAL TRANSDUCERS; DEAF-AID SETS; PUBLIC ADDRESS SYSTEMS
- H04R3/00—Circuits for transducers, loudspeakers or microphones
- H04R3/02—Circuits for transducers, loudspeakers or microphones for preventing acoustic reaction, i.e. acoustic oscillatory feedback
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L21/00—Speech or voice signal processing techniques to produce another audible or non-audible signal, e.g. visual or tactile, in order to modify its quality or its intelligibility
- G10L21/02—Speech enhancement, e.g. noise reduction or echo cancellation
- G10L21/0208—Noise filtering
- G10L2021/02082—Noise filtering the noise being echo, reverberation of the speech
Definitions
- the present invention relates to a signal processing method, a signal processing device, and a signal processing program.
- Non-Patent Document 1 discloses a linear combination type signal processing apparatus that eliminates echoes generated by propagation of a received signal through a spatial acoustic path in a system having a plurality of received signals and a single or a plurality of transmitted signals.
- FIG. 18 shows a block diagram of a linearly coupled multi-channel signal processing apparatus when the number of channels is 2, that is, for stereo signals.
- the linear combination type has a problem of coefficient indefiniteness in which the coefficient of the adaptive filter converges to an indefinite value other than a value (correct solution) equal to the characteristic of the echo path. The converged filter coefficient value depends on the cross-correlation of the filter input signal.
- a correct change in coefficient value due to a change in cross-correlation means that the echo cancellation capability is reduced even when there is no fluctuation in the echo path. Therefore, residual echo is perceived, and the call quality deteriorates.
- Patent Document 2 discloses a method that can uniquely determine the coefficient.
- the number of conditional expressions used for calculating the coefficient of the adaptive filter has increased due to the introduction of the delayed received signal, and there is no problem that the solution that is the coefficient of the adaptive filter becomes indefinite. . Therefore, the coefficients of the adaptive filter converge to an optimal value that is uniquely determined.
- movement of a sound image may be perceived when switching between a received signal and a delayed received signal.
- JP 04-284732 A Japanese Patent Laid-Open No. 11-004183 JP 2000-078061 A
- Patent Document 3 and Non-Patent Document 3 have a slower convergence speed than a linearly coupled signal processing apparatus.
- the convergence speed when the convergence speed is increased, the movement of the sound image localization may be perceived, and the subjective sound quality of the received signal is deteriorated. Therefore, the method disclosed in Patent Document 3 cannot simultaneously achieve a short convergence time and a high subjective sound quality.
- the present invention is a signal processing program for causing a computer to execute reception processing for receiving a plurality of reception signals and echo reduction processing for reducing a plurality of echoes generated by the plurality of reception signals, Delay received signal generation processing for generating a delayed received signal by delaying at least one received signal of the plurality of received signals, and generating the pseudo echo by inputting the received signal and the delayed received signal to the adaptive filter And a pseudo echo subtracting process for subtracting the pseudo echo from each of the plurality of received signals, and the frequency of inputting the received signal and the delayed received signal to the adaptive filter
- a signal processing program is characterized in that control is performed based on perceptual sensitivity to a change in localization of a received signal.
- the signal processing method, the signal processing apparatus, and the signal processing program of the present invention generate a delayed received signal by delaying at least one received signal, and operate the adaptive filter using the received signal and the delayed received signal as inputs. . Since both the received signal and the delayed received signal are used, there is no problem that the number of conditional expressions for obtaining the adaptive filter coefficient increases and the solution becomes indefinite. Therefore, the coefficients of the adaptive filter converge to an optimal value that is uniquely determined.
- the analysis circuit 350 receives the reception signal 1 and the reception signal 2 and calculates the perceptual sensitivity to the movement of the sound image localized by these reception signals.
- the analysis circuit 350 generates a clock signal corresponding to the obtained perceptual sensitivity and supplies it to the filter 310.
- the filter 310 determines the generation frequency of the delayed reception signal based on the supplied clock signal.
- the perceptual sensitivity is high, the frequency of switching between the received signal and the delayed received signal is low, and when the perceptual sensitivity is low, the clock signal is generated so that the frequency of switching between the received signal and the delayed received signal is increased and supplied to the filter 310. .
- the analysis circuit 350 does not perform the correlation calculation when the signal power or amplitude is extremely small. By excluding small signal samples that are susceptible to unwanted effects due to the added noise, the immunity to noise can be increased.
- the signal supplied to the input terminal 3100 is transmitted to the delay element 3101 1 and the coefficient multiplier 3102 0 .
- Coefficient multiplier 3102 0 multiplies the input received signal sample by coefficient value c 0 and transmits the product to adder 3103 1 .
- the delay element 3101 1 delays the received signal sample by one sample and transmits it to the coefficient multiplier 3102 1 .
- the coefficient multiplier 3102 1 multiplies the input received signal sample by the coefficient value c 1 and transmits the product to the adder 3103 1 .
- the adder 3103 1 adds the outputs of the coefficient multiplier 3102 0 and the coefficient multiplier 3102 1 and outputs the sum to the output terminal 3104 as a delayed received signal.
- a clock signal is supplied to the input terminal 3105 from the analysis circuit 350 of FIG. 1 and is transmitted to the coefficient multipliers 3102 0 , 3102 1 , and 3102 2 .
- the coefficient multipliers 3102 0 , 3102 1 , and 3102 2 change the coefficient values based on the clock signal supplied from the input terminal 3105.
- c 0 (k) is shown in FIG. i is an arbitrary natural number. Periodically take 1 and 0 every M (integer) samples. Further, as is clear from Equation 1 , c 1 (k) changes as shown in a diagram obtained by vertically inverting FIG. That is, c 0 (k) and c 1 (k) are exclusive, and any input of the adder 3103 is 0. Therefore, the output of the adder 3103 is equal to either the received signal or the delayed received signal, which is equivalent to switching the received signal or the delayed received signal every M samples.
- the maximum value of c 0 (k) can be set to an arbitrary value, but at that time, the change in amplitude is compensated so that the same output as when the maximum value of c 0 (k) is 1 is obtained. Need to scale the output.
- Coefficient multiplier 3102 1 multiplies the output of delay element 3101 1 by coefficient value c 1 and transmits the product to adder 3103 1 .
- Adder 3103 1 adds the outputs of coefficient multiplier 3102 0 and coefficient multiplier 3102 1 , and transmits the sum to adder 3103 2 .
- the delay element 3101 2 delays the output of the delay element 3101 1 by one sample and transmits it to the coefficient multiplier 3102 2 .
- a clock signal is supplied to the input terminal 3105 from the analysis circuit 350 of FIG. 1 and is transmitted to the coefficient multipliers 3102 0 , 3102 1 , and 3102 2 .
- the coefficient multipliers 3102 0 , 3102 1 , and 3102 2 change the coefficient values based on the clock signal supplied from the input terminal 3105.
- FIG. 5 shows an example of the coefficients c 0 (k), c 1 (k), and c 2 (k) of the coefficient multipliers 3102 0 , 3102 1 , and 3102 2 .
- the coefficients c 0 (k), c 1 (k) and c 2 (k) take 1 exclusively, the received signals corresponding to the respective coefficient multipliers are output as delayed received signals as output terminals. Obtained in 3104.
- FIG. 6 is a block diagram illustrating a third configuration example of the filter 310. It is configured as an L-tap FIR filter having c 0 , c 1 ,..., C L ⁇ 1 as coefficients.
- the reception signal 1 of FIG. 1 is supplied to the input terminal 3100 of FIG.
- the signal obtained at the output terminal 3104 in FIG. 6 is a delayed received signal.
- Coefficient multipliers 3102 L-1 multiplies the coefficient value c L-1 to the output of the delay element 3101 L-1, and transmits the product to the adder 3103 L-1.
- Adder 3103 L-1 adds the outputs of adder 3103 L-2 and coefficient multiplier 3102 L-1 , and outputs the sum to output terminal 3104 as a delayed received signal.
- a clock signal is supplied to the input terminal 3105 from the analysis circuit 350 in FIG. 1, and is transmitted to the coefficient multipliers 3102 0 , 3102 1 ,..., 3102 L ⁇ 1 .
- the coefficient multipliers 3102 0 , 3102 1 ,..., 3102 L ⁇ 1 change the coefficient values based on the clock signal supplied from the input terminal 3105.
- the coefficients c 0 (k), c 1 (k), ..., c L-1 (k) of the coefficient multipliers 3102 0 , 3102 1 , ..., 3102 L-1 are connected to each tap of the filter 310 in parallel. I think that it is. That is, c 0 (k), c 1 (k),..., C L-1 (k) take a non-zero value exclusively, and when either is non-zero, the others are zero. As described with reference to FIGS. 3A and 5, c 0 (k), c 1 (k),..., C L ⁇ 1 (k) take non-zero exclusively, A received signal that has received a delay corresponding to the coefficient multiplier is obtained as a delayed received signal at the output terminal 3104.
- the FIR filter is assumed as the configuration of the filter 310.
- a combination of a variable delay circuit and a switch Other structures such as a variable delay circuit and a variable weighting mixing circuit may be used.
- a plurality of variable delay circuits give different delays to the received signal to generate a plurality of delayed received signals, and these multiple delayed received signals and received signals are switched by a switch, or appropriately mixed by a variable weighting mixing circuit. By doing so, the same function as the time-varying coefficient FIR filter can be realized.
- the frequency at which the received signal and the delayed received signal are input to the adaptive filter is controlled based on the perceptual sensitivity with respect to the localization of the plurality of received signals. For this reason, the received signal and the delayed received signal can be input to the adaptive filter at a frequency that is not subjectively perceived in accordance with the state of the signal, and deterioration in subjective sound quality can be reduced.
- FIG. 7 shows a second embodiment when the number of reception signals and transmission signals is two in the signal processing apparatus of the present invention.
- the difference from the best embodiment described with reference to FIGS. 1 to 6 is that a delay processing circuit 301 is provided instead of the delay processing circuit 300.
- this difference will be described in detail.
- the delay processing circuit 301 delays the received signal 1 and the call 2 to generate delayed received signals.
- the adaptive filters 121 and 123 and the digital-analog (DA) converter 18 and the adaptive filters 122 and 124 and the digital- Each signal is transmitted to an analog (DA) converter 19.
- the delay processing circuit 301 includes filters 310 and 320 and an analysis circuit 351.
- Filters 310 and 320 generate delayed received signals by delaying received signals 1 and 2, respectively.
- the filters 310 and 320 may output the received signals 1 and 2 as they are without delaying.
- the frequency with which the outputs of the filters 310 and 320 change between the delayed received signal and the received signal 1 or between the delayed received signal and the received signal 2 is controlled by a clock signal supplied from the analysis circuit 351. The higher the frequency, the faster the convergence of the adaptive filters 121 and 123 and the adaptive filters 122 and 124 to which the switching signal is supplied.
- the analysis circuit 351 receives the reception signal 1 and the reception signal 2 and calculates the perceptual sensitivity to the movement of the sound image localized by these reception signals.
- the analysis circuit 351 generates a clock signal corresponding to the obtained perceptual sensitivity and supplies it to the filters 310 and 320.
- the filters 310 and 320 determine the generation frequency of the delayed reception signal based on the supplied clock signal.
- the clock signals supplied to the filters 310 and 320 are out of phase with each other. This phase shift will be described later with reference to FIG.
- the configuration of the filter 320 is exactly the same as the configuration of the filter 310 described with reference to FIGS.
- other structures such as a combination of a variable delay circuit and a switch, a variable delay circuit and a variable weighting mixing circuit can be used as long as the received signal and the delayed received signal can be switched and output with time.
- a plurality of variable delay circuits give different delays to the received signal to generate a plurality of delayed received signals, and these multiple delayed received signals and received signals are switched by a switch, or appropriately mixed by a variable weighting mixing circuit. By doing so, the same function as the time-varying coefficient FIR filter can be realized.
- the maximum value of the relative delay of the output signal of the filter 310 with respect to the output signal of the filter 320 is equal to the maximum value of the relative delay of the output signal of the filter 320 with respect to the output signal of the filter 310, switching to the delayed reception signal is performed.
- the left and right shift amounts of the sound image localization generated by the above are equal, and the sound image is perceived as changing symmetrically with time.
- the relative delay of the output signal of the filter 310 with respect to the output signal of the filter 320 is 1
- the relative delay of the output signal of the filter 320 with respect to the output signal of the filter 310 is 1.
- Such a symmetrical change in the sound image localization is perceived as blurring of the sound image, so that the subjective sound quality is less deteriorated than the asymmetrical sound image movement to the left or right.
- the convergence time can be changed.
- the coefficients c 0 (k), c 1 (k), and c 2 (k) of the coefficient multipliers 3102 0 , 3102 1 , and 3102 2 are changed from non-zero to zero.
- the change (or vice versa) can be set to be proportional to time, or can be set to be any smooth curve or straight line connecting non-zero and zero.
- two or more received signals are delayed to generate a delayed received signal, and the received signal and the delayed received signal are input to the adaptive filter.
- the coefficients of the adaptive filter converge to an optimal value that is uniquely determined.
- the frequency at which the received signal and the delayed received signal are input to the adaptive filter is controlled based on the perceptual sensitivity with respect to the localization of the plurality of received signals. For this reason, the received signal and the delayed received signal can be input to the adaptive filter at a frequency that is not subjectively perceived in accordance with the state of the signal, and deterioration in subjective sound quality can be reduced.
- the number of the conditional expressions is further increased, and the convergence time of the solution to the optimum value can be shortened.
- the difference between the left and right channels in the maximum relative delay of the delayed signal of the channel reproduced by the left and right speakers farthest from the center with respect to the received signal takes into account the deviation of the sound image due to left and right asymmetry in the left and right speaker arrangements.
- the amplitude correction circuit 400 corrects the amplitude of the delayed reception signal that is the output of the delay processing circuit 301 to generate an amplitude correction delayed reception signal, the adaptive filters 121 and 123, the digital-analog (DA) converter 18, and The data is transmitted to the adaptive filters 122 and 124 and the DA converter 19, respectively.
- DA digital-analog
- the amplitude correction of the delayed reception signal in the amplitude correction circuit 400 is performed when the output of the delay processing circuit 301 is equal to the delayed reception signal obtained by delaying the reception signal 1 or 2.
- Amplitude correction corrects the relative relationship of the amplitude of signals between a plurality of channels, and cancels the displacement of the sound image localization that occurs when a delayed received signal is used instead of the received signal.
- correction can be performed in all channels so that the total power of all channels is equal to that before correction. By keeping the total power of all channels constant, subjective discomfort can be eliminated when the correction is switched.
- the amplitude correction circuit 400 includes amplitude processing circuits 410 and 420.
- the amplitude processing circuit 410 corrects the amplitude of the delayed received signal generated by delaying the received signal 1 to generate an amplitude corrected delayed received signal.
- the amplitude processing circuit 420 corrects the amplitude of the delayed received signal generated by delaying the received signal 2 to generate an amplitude corrected delayed received signal.
- the amplitude processing circuits 410 and 420 can be configured exactly the same.
- a clock signal is supplied from the analysis circuit 351 included in the delay processing circuit 301 to the amplitude processing circuits 410 and 420. These clock signals are used to apply amplitude correction in accordance with the generation timing of the delay signal in the delay processing circuit 301.
- FIG. 11 is a block diagram illustrating a configuration example of the amplitude processing circuit 410.
- a multiplier 4101 having g 0 as a coefficient is configured.
- a delayed reception signal obtained by delaying the reception signal 1 is supplied to the input terminal 4100 in FIG.
- the multiplier 4101 multiplies the signal supplied to the input terminal 4100 by g 0 and transmits it to the output terminal 4104.
- the signal obtained at the output terminal 4104 in FIG. 11 is a signal obtained by multiplying the delayed reception signal supplied to the input terminal 4100 by g 0 .
- the amplitude processing circuit 420 can be obtained by using g 1 instead of g 0 as the coefficient of the multiplier 4101 in FIG. 11, which is a block diagram showing a configuration example of the amplitude processing circuit 410.
- g 0 and g 1 take values of 1 when the received signals 1 and 2 are supplied to the amplitude processing circuits 410 and 420, respectively, and take values other than 1 (g 0 bar and g 1 bar).
- the g 0 bar and the g 1 bar are set to values that compensate for the displacement of the sound image localization due to the delayed received signal.
- the total power of all channels can be set to be equal to that before correction. By keeping the total power of all channels constant, subjective discomfort can be eliminated when the correction is switched.
- C is a positive constant. Therefore, before the amplitude correction, when the powers of the reception signal 1 and the reception signal 2 are P 1 bar [dB] and P 2 bar [dB], respectively, the power P of the reception signal 1 and the reception signal 2 after amplitude correction.
- the received signal is processed by the delay processing circuit 301 to generate a delayed received signal, and the amplitude of the delayed received signal is corrected by the amplitude correcting circuit 400 to obtain the amplitude corrected delayed received signal.
- the configuration for generating and supplying to the adaptive filters 121, 123, 122, 124 has been described.
- the order of processing the received signals is changed, the amplitude of the received signals is corrected by the amplitude correction circuit 400 to generate an amplitude corrected received signal, and the amplitude corrected received signal is processed by the delay processing circuit 301 to receive the amplitude corrected delayed reception.
- a configuration in which a signal is generated and supplied to the adaptive filters 121, 123, 122, and 124 is also possible. Since the configurations and operations of the delay processing circuit 301 and the amplitude correction circuit 400 at that time have already been described in detail, they are omitted here.
- two or more received signals are delayed to generate a delayed received signal, and the received signal and the delayed received signal are input to the adaptive filter.
- the adaptive filter Since both the received signal and the delayed received signal are used, there is no problem that the number of conditional expressions for obtaining the adaptive filter coefficient increases and the solution becomes indefinite. Therefore, the coefficients of the adaptive filter converge to an optimal value that is uniquely determined.
- the frequency at which the received signal and the delayed received signal are input to the adaptive filter is controlled based on the perceptual sensitivity with respect to the localization of the plurality of received signals. For this reason, the received signal and the delayed received signal can be input to the adaptive filter at a frequency that is not subjectively perceived in accordance with the state of the signal, and deterioration in subjective sound quality can be reduced.
- the number of the conditional expressions is further increased, and the convergence time of the solution to the optimum value can be shortened.
- the difference between the left and right channels in the maximum relative delay of the delayed signal of the channel reproduced by the left and right speakers farthest from the center with respect to the received signal takes into account the deviation of the sound image due to left and right asymmetry in the left and right speaker arrangements.
- the sound image movement caused by the introduction of the delayed received signal is canceled by the amplitude correction processing for the input signal, the sound quality deterioration of the received signal that is directly supplied to the speaker and received is reduced, and good sound quality can be maintained.
- FIG. 12 shows a fourth embodiment in which the number of reception signals and transmission signals is two in the signal processing apparatus of the present invention.
- the difference from the third embodiment described with reference to FIGS. 10 and 11 is that the output signal of the amplitude correction circuit 400 is processed by the nonlinear processing circuit 500 and then supplied to the adaptive filters 121, 123, 122, and 124. It is a point.
- the non-linear processing circuit 500 performs non-linear processing on the amplitude-corrected delayed received signal that is the output of the amplitude correcting circuit 400 to generate a non-linear amplitude-corrected delayed received signal, the adaptive filters 121 and 123, and the digital-analog (DA) converter 18. , And the adaptive filters 122 and 124 and the DA converter 19, respectively.
- the non-linear amplitude correction delay reception signal has a smaller cross-correlation between the plurality of channels than the amplitude correction delay reception signal. Therefore, the convergence of the adaptive filters 121, 123, 122, and 124 can be further accelerated.
- the nonlinear processing circuit 500 includes nonlinear amplitude processing circuits 510 and 520.
- the nonlinear amplitude processing circuit 510 generates a nonlinear amplitude-corrected delayed reception signal by nonlinearly processing the amplitude of the amplitude-corrected delayed reception signal obtained by delaying the reception signal 1 and correcting the amplitude.
- the nonlinear amplitude processing circuit 520 generates a nonlinear amplitude-corrected delayed reception signal by nonlinearly processing the amplitude of the amplitude-corrected delayed reception signal obtained by delaying the reception signal 2 and correcting the amplitude.
- the non-linear amplitude processing circuits 510 and 520 can have exactly the same configuration.
- FIG. 13 is a block diagram showing a configuration example of the nonlinear amplitude processing circuit 510.
- the nonlinear amplitude processing circuit 510 includes a coefficient multiplier 512, a polarity determination circuit 513, a multiplier 514, and an adder 515.
- the input terminal 511 is supplied with an amplitude correction delayed reception signal that is an output of the amplitude correction circuit 400 of FIG.
- the amplitude-corrected delayed reception signal is transmitted to the coefficient multiplier 512, the polarity determination circuit 513, and the adder 515.
- the coefficient multiplier 512 multiplies the input signal by ⁇ and outputs it.
- the polarity determination circuit 513 outputs 1 when the polarity of the signal supplied to the input is positive, and outputs 0 when it is negative.
- the multiplier 514 is supplied with the outputs of the coefficient multiplier 512 and the polarity determination circuit 513, and transmits the product of both to the adder 515.
- the other input terminal of the adder 515 is supplied with the amplitude correction delayed reception signal as it is. That is, the output of the adder 515 with respect to the signal sample x (k) at the input terminal 511 is (1 + ⁇ ) x (k) when the polarity of the input signal is positive, and x (k) when the polarity is negative.
- This signal becomes an output signal of the nonlinear amplitude processing circuit 510. That is, nonlinear amplitude processing circuit 510 constitutes a half-wave rectifier circuit.
- the nonlinear amplitude processing circuit 520 can have the same configuration as the nonlinear amplitude processing circuit 510.
- the received signal is processed by the delay processing circuit 301 to generate a delayed received signal, and the amplitude of the delayed received signal is corrected by the amplitude correcting circuit 400 to obtain the amplitude corrected delayed received signal.
- the configuration of generating and processing the amplitude-corrected delayed received signal by the nonlinear amplitude processing circuit 500 to generate the nonlinear amplitude-corrected delayed received signal and supplying it to the adaptive filters 121, 123, 122, and 124 has been described.
- the received signal is processed in the order of amplitude correction, delay, nonlinear processing and the order of nonlinear processing, delay, amplitude correction after changing the order of processing the received signal, and then the adaptive filters 121, 123, 122, 124, It is also possible to use a configuration for supplying to the battery. Since the configurations and operations of the delay processing circuit 301, the amplitude correction circuit 400, and the nonlinear processing circuit 500 at that time have already been described in detail, they are omitted here.
- two or more received signals are delayed to generate a delayed received signal, and the amplitude of the delayed received signal is corrected to correct the amplitude corrected delayed received signal.
- the adaptive filter is operated with the received signal and the nonlinear amplitude-corrected delayed received signal as inputs. Since both the received signal and the non-linear amplitude-corrected delayed received signal are used, there is no problem that the number of conditional expressions for obtaining the adaptive filter coefficient increases and the solution becomes indefinite. Therefore, the coefficients of the adaptive filter converge to an optimal value that is uniquely determined.
- the frequency at which the received signal and the delayed received signal are input to the adaptive filter is controlled based on the perceptual sensitivity with respect to the localization of the plurality of received signals. For this reason, the received signal and the delayed received signal can be input to the adaptive filter at a frequency that is not subjectively perceived in accordance with the state of the signal, and deterioration in subjective sound quality can be reduced.
- the number of the conditional expressions is further increased, and the convergence time of the solution to the optimum value can be shortened.
- the difference between the left and right channels in the maximum relative delay of the delayed signal of the channel reproduced by the left and right speakers farthest from the center with respect to the received signal takes into account the deviation of the sound image due to left and right asymmetry in the left and right speaker arrangements.
- the sound image movement caused by the introduction of the delayed received signal is canceled by the amplitude correction processing for the input signal, the sound quality deterioration of the received signal that is directly supplied to the speaker and received is reduced, and good sound quality can be maintained. . Furthermore, the convergence time can be further shortened by the synergistic effect of the nonlinear processing and the introduction of the delayed received signal.
- FIG. 14 shows a fifth embodiment in which the number of received signals and transmitted signals is two in the signal processing apparatus of the present invention.
- the difference from the fourth embodiment described with reference to FIGS. 12 and 13 is that the nonlinear processing circuit 500 is a nonlinear processing circuit 501.
- the non-linear processing circuit 501 includes non-linear amplitude processing circuits 530 and 540.
- the non-linear amplitude processing circuit 530 performs non-linear processing on the amplitude-corrected delayed received signal whose amplitude is corrected after delaying the received signal 1 using the amplitude-corrected delayed received signal whose amplitude is corrected after delaying the received signal 2.
- the non-linear amplitude processing circuit 540 performs non-linear processing on the amplitude-corrected delayed received signal whose amplitude is corrected after delaying the received signal 2 by using the amplitude-corrected delayed received signal whose amplitude is corrected after delaying the received signal 1.
- the non-linear amplitude processing circuits 530 and 540 can have exactly the same configuration.
- FIG. 15 is a block diagram illustrating a configuration example of the nonlinear amplitude processing circuit 530.
- the nonlinear amplitude processing circuit 530 includes a coefficient multiplier 512, a polarity determination circuit 513, a multiplier 514, and an adder 515.
- the input terminal 531 is supplied with an amplitude-corrected delayed reception signal obtained by delaying the reception signal 1 from the output of the amplitude correction circuit 400 in FIG. 14 and correcting the amplitude.
- the input terminal 537 is supplied with an amplitude-corrected delayed received signal obtained by delaying the received signal 2 from the output of the amplitude correcting circuit 400 in FIG. 14 and correcting the amplitude.
- the amplitude-corrected delayed reception signal generated from the reception signal 1 is transmitted to the polarity determination circuit 513 and the adder 515.
- the amplitude-corrected delayed reception signal generated from the reception signal 2 is transmitted to the coefficient multiplier 512.
- the coefficient multiplier 512 multiplies the input signal by ⁇ and outputs it.
- the polarity determination circuit 513 outputs 1 when the polarity of the signal supplied to the input is positive, and outputs 0 when it is negative.
- the multiplier 514 is supplied with the outputs of the coefficient multiplier 512 and the polarity determination circuit 513, and transmits the product of both to the adder 515.
- the non-linear amplitude processing circuit 530 has a configuration in which the input of the coefficient multiplier 512 in the non-linear amplitude processing circuit 510 is changed from an amplitude-corrected delayed received signal generated from the received signal 1 to an amplitude-corrected delayed received signal generated from the received signal 2. It has become.
- the nonlinear amplitude processing circuit 540 can have the same configuration as the nonlinear amplitude processing circuit 530. In this configuration, since a signal generated from a received signal of another channel is used for nonlinear processing, the amount of change from the signal before nonlinear processing increases, and the effect of reducing the correlation between channels increases.
- the received signal is processed by the delay processing circuit 301 to generate a delayed received signal, and the amplitude of the delayed received signal is corrected by the amplitude correcting circuit 400 to obtain the amplitude corrected delayed received signal.
- the configuration of generating and processing the amplitude correction delayed reception signal by the nonlinear amplitude processing circuit 501 to generate the nonlinear amplitude correction delay reception signal and supplying it to the adaptive filters 121, 123, 122, and 124 has been described.
- two or more received signals are delayed to generate a delayed received signal, the amplitude of the delayed received signal is corrected, and the amplitude corrected delayed received signal is corrected.
- the frequency at which the received signal and the delayed received signal are input to the adaptive filter is controlled based on the perceptual sensitivity with respect to the localization of the plurality of received signals. For this reason, the received signal and the delayed received signal can be input to the adaptive filter at a frequency that is not subjectively perceived in accordance with the state of the signal, and deterioration in subjective sound quality can be reduced.
- the number of the conditional expressions is further increased, and the convergence time of the solution to the optimum value can be shortened.
- the difference between the left and right channels in the maximum relative delay of the delayed signal of the channel reproduced by the left and right speakers farthest from the center with respect to the received signal takes into account the deviation of the sound image due to left and right asymmetry in the left and right speaker arrangements.
- the convergence time can be further shortened by a synergistic effect of nonlinear processing using a plurality of channels of received signals and introduction of delayed received signals.
- FIG. 16 shows a sixth embodiment in which the number of received signals and transmitted signals is two in the signal processing apparatus of the present invention.
- the frequency analysis / synthesis circuit 600 is provided in front of the delay processing circuit 301, the DA converters 18 and 19, and This is that a frequency analysis / synthesis circuit 610 is provided after the AD converters 20 and 21. Therefore, the delay processing circuit 301, the adaptive filters 121, 122, 123, and 124, and the subtractors 129 and 130 all operate on the narrowband signal that has been band-divided.
- the frequency analysis / synthesis circuit 600 divides the received signals 1 and 2 into bands and transmits them to the delay processing circuit 301.
- the frequency analysis synthesis circuit 600 also band-synthesizes the outputs of the subtractors 129 and 130 to form full-band output signals 16 and 17.
- the frequency analysis / synthesis circuit 610 performs band synthesis on the output of the delay processing circuit 301 and transmits it to the DA converters 18 and 19.
- the frequency analysis / synthesis circuit 610 also band-divides the outputs of the AD converters 20 and 21 and transmits them to the subtracters 129 and 130.
- the delay processing circuit 301 adds a delay to the band-divided signal and outputs it as a band-divided delayed reception signal.
- an optimum delay can be given to each band-divided signal. Accordingly, the degree of freedom in selecting a relative delay as large as possible with the allowable amount of sound image movement described with reference to FIG. 1 increases, leading to improvement in subjective sound quality.
- the frequency analysis function in the frequency analysis synthesis circuits 600 and 610 can be realized by applying frequency conversion to input signal samples divided into frames.
- frequency conversion Fourier transform, cosine transform, KL (Kalunen label) transform, and the like are known.
- Non-patent document 9 (1990, "Digital Coding of Waveforms", Prentice Hall (DIGITAL CODING OF OF WAVEFORMS, PRINCIPLES AND AND APPLICATIONS TO SPEECH AND VIDEO, PRENTICE-HALL, 1990.)). It is well known that other transformations such as Hadamard transformation, Haar transformation, and wavelet transformation can be used.
- the frequency analysis function can also be realized by applying the above-described conversion to the result obtained by weighting the input signal samples of the frame with the window function W.
- window functions such as Hamming, Hanning (Han), Kaiser, and Blackman are known. A more complicated window function can also be used.
- Non-patent documents 10 and 11 disclose techniques related to these window functions. Furthermore, it is also widely practiced to overlap (overlap) a part of two or more consecutive frames. In this case, the above-described frequency conversion is used for the signal that is overlapped and windowed.
- the technology related to blocking and conversion with overlap is disclosed in Non-Patent Document 10 (1975, “Digital Signal Processing”, Prentice Hall (DIGITAL SIGNAL PROCESSING, PRENTICE-HALL, 1975)). ing.
- the frequency analysis function of the frequency analysis / synthesis circuits 600 and 610 may be configured by a band division filter bank.
- the band division filter bank is composed of a plurality of band pass filters. Each frequency band of the band division filter bank may be equally spaced or unequal. By dividing the band at unequal intervals, the time resolution can be reduced by dividing the band into a narrow band in the low band and the time resolution can be increased by dividing the band into a wide band in the high band.
- Typical examples of unequal interval division include octave division in which the band is successively halved toward the low band and critical band division corresponding to human auditory characteristics.
- the frequency synthesis function of the frequency analysis / synthesis circuits 600 and 610 is composed of a frequency conversion that realizes the frequency analysis function of the frequency analysis / synthesis circuits 600 and 610 and an inverse conversion corresponding thereto.
- the frequency analysis function of the frequency analysis synthesis circuits 600 and 610 includes weighting by the window function W, the frequency synthesized signal is multiplied by the window function W.
- the frequency analysis function of the frequency analysis synthesis circuits 600 and 610 is configured by a band division filter bank
- the frequency synthesis function of the frequency analysis synthesis circuits 600 and 610 is configured by a band synthesis filter bank.
- a technique related to the band synthesis filter bank and its design method is disclosed in Non-Patent Document 11.
- a frequency division analysis is performed on two or more received signals to generate a band division reception signal, and the band division reception signal is delayed to obtain a band division delay.
- a reception signal is generated, and an adaptive filter is operated with the band division reception signal and the band division delay reception signal as inputs. Since both the band division reception signal and the band division delay reception signal are used, there is no problem that the number of conditional expressions for obtaining the adaptive filter coefficient increases and the solution becomes indefinite. Therefore, the coefficients of the adaptive filter converge to an optimal value that is uniquely determined.
- the number of the conditional expressions is further increased, and the convergence time of the solution to the optimum value can be shortened.
- the difference between the left and right channels in the maximum relative delay of the delayed signal of the channel reproduced by the left and right speakers farthest from the center with respect to the received signal takes into account the deviation of the sound image due to left and right asymmetry in the left and right speaker arrangements.
- the present invention is not limited to a plurality of received signals and a single or a plurality of received signals. This is applicable to a general case where there are two or more transmission signals.
- An example is an acoustic echo in which a received signal propagates through a spatial acoustic path from a speaker and eliminates an acoustic echo recorded by a microphone.
- an echo other than an acoustic echo for example, an echo caused by a line crosstalk, etc. Applicable.
- a cyclic adaptive filter may be used instead of the non-cyclic adaptive filter.
- a subband adaptive filter or a transform domain adaptive filter may also be used.
- the seventh embodiment of the present invention includes a computer 1000 that operates under program control.
- the computer 1000 performs processing according to any of the above-described best embodiments and the second to sixth embodiments of the present invention on the received signals received from the input terminals 1 and 2, and outputs the output signals 16 and 17 operates based on a program for outputting a signal from which echoes are eliminated.
- the first embodiment receives a plurality of received signals, and subtracts pseudo echoes generated by a plurality of adaptive filters that receive the plurality of received signals from a plurality of echoes generated by the plurality of received signals.
- a signal processing method for reducing the plurality of echoes, wherein a delay reception signal is generated by delaying at least one reception signal among the plurality of reception signals, and the reception signal and the delay reception signal are obtained. Controlling the frequency of inputting the received signal and the delayed received signal to the adaptive filter based on perceptual sensitivity to a localization change of the plurality of received signals. This is a characteristic signal processing method.
- the second embodiment is characterized in that, in the above-mentioned embodiment, at least one of the delayed reception signals is an amplitude-corrected delayed reception signal whose amplitude is corrected.
- the fourth embodiment is characterized in that, in the above-described embodiment, the received signal is decomposed into a plurality of frequency components and delayed for each of the plurality of frequency components to generate a delayed received signal.
- the fifth embodiment is characterized in that, in the above embodiment, the perceptual sensitivity to the localization change is obtained by the similarity of the received signals.
- the sixth embodiment is characterized in that, in the above embodiment, the perceptual sensitivity to the change in localization is obtained by the power of the received signal.
- the seventh embodiment is characterized in that, in the above-described embodiment, the relative delay of the delayed received signal with respect to the received signal is generated so as to take a plurality of values that change with time.
- the eighth embodiment is characterized in that, in the above-described embodiment, the relative delay is an integral multiple of a sampling period.
- the ninth embodiment is characterized in that, in the above-described embodiment, the delayed reception signal is generated by processing the reception signal with a filter having a plurality of time-varying coefficients that alternately take zero and non-zero values.
- the tenth embodiment is characterized in that, in the above-described embodiment, the plurality of time-varying coefficients take mutually exclusive zero values.
- a plurality of reception signals are received, and a pseudo echo generated by a plurality of adaptive filters that receive the plurality of reception signals is subtracted from a plurality of echoes generated by the plurality of reception signals.
- a signal processing device for reducing the plurality of echoes wherein a linear processing circuit that generates a delayed reception signal by delaying at least one reception signal among the plurality of reception signals, the reception signal, and the delay
- An adaptive filter that receives a received signal and generates a pseudo echo, a plurality of subtractors that generate a signal with reduced echo by subtracting the pseudo echo from the mixed signal, and a change in localization of the plurality of received signals
- An analysis circuit for obtaining a perceptual sensitivity to the adaptive filter, wherein the adaptive filter converts the received signal and the delayed received signal based on the perceptual sensitivity. It controls the frequency of the input is a signal processing device and controls the coefficients of the plurality of adaptive filters to minimize the output of said plurality of subtracters.
- the thirteenth embodiment is characterized in that, in the above-mentioned embodiment, an amplitude correction circuit for generating an amplitude-corrected delayed reception signal by correcting the amplitude of at least one of the delayed reception signals is provided.
- the fourteenth embodiment is characterized in that, in the above-described embodiment, a non-linear processing circuit that generates a non-linear processing signal by performing non-linear processing on at least one of the signals input to the plurality of adaptive filters is provided.
- a frequency analysis circuit that decomposes the received signal into a plurality of frequency components, and a linear processing circuit that generates a delayed received signal by delaying the plurality of frequency components. It is characterized by comprising.
- the sixteenth embodiment is characterized in that, in the above-described embodiment, the analysis circuit obtains the perceptual sensitivity to the localization change by the similarity of the received signals.
- the seventeenth embodiment is characterized in that, in the above-described embodiment, the analysis circuit obtains the perceptual sensitivity to the localization change by the power of the received signal.
- the linear processing circuit performs processing such that a relative delay of the delayed received signal with respect to the received signal takes a plurality of values that change with time.
- the nineteenth embodiment is characterized in that, in the above-described embodiment, the linear processing circuit performs processing such that the relative delay is an integral multiple of a sampling period.
- the twentieth embodiment is characterized in that, in the above-described embodiment, the linear processing circuit includes a filter having a plurality of time-varying coefficients that alternately take zero and non-zero values.
- the twenty-first embodiment is characterized in that, in the above-described embodiment, the plurality of time-varying coefficients take mutually exclusive zero values.
- the twenty-second embodiment is characterized in that, in the above-described embodiment, the plurality of time-varying coefficients take non-zero values exclusively from each other.
- a twenty-third embodiment is a signal processing program for causing a computer to execute reception processing for receiving a plurality of reception signals and echo reduction processing for reducing a plurality of echoes generated by the plurality of reception signals.
- a delayed received signal generation process for generating a delayed received signal by delaying at least one received signal of the plurality of received signals; and the received signal and the delayed received signal are input to the adaptive filter to generate a pseudo echo
- a pseudo echo generation process to generate, and a pseudo echo subtraction process to subtract the pseudo echo from each of the plurality of received signals, and the frequency at which the received signal and the delayed received signal are input to the adaptive filter
- It is a signal processing program characterized by controlling based on the perceptual sensitivity with respect to the change of the localization of the received signal.
- the twenty-fourth embodiment is characterized in that, in the above-described embodiment, at least one of the delayed reception signals is an amplitude-corrected delayed reception signal whose amplitude is corrected.
- the twenty-fifth embodiment is characterized in that, in the above-described embodiment, at least one of the signals inputted to the plurality of adaptive filters is a nonlinear processed signal.
- the twenty-sixth embodiment is characterized in that, in the above-described embodiment, the received signal is decomposed into a plurality of frequency components and delayed for each of the plurality of frequency components to generate a delayed received signal.
- the twenty-ninth embodiment is characterized in that, in the above embodiment, the relative delay of the delayed received signal with respect to the received signal is generated so as to take a plurality of values that change with time.
- the 30th embodiment is characterized in that, in the above embodiment, the relative delay is an integral multiple of a sampling period.
- the thirty-first embodiment is characterized in that, in the above-described embodiment, the delayed received signal is generated by processing the received signal with a filter having a plurality of time-varying coefficients that alternate between zero and non-zero values.
- the thirty-second embodiment is characterized in that, in the above-described embodiment, the plurality of time-varying coefficients take mutually exclusive zero values.
- the thirty-third embodiment is characterized in that, in the above-described embodiment, the plurality of time-varying coefficients take non-zero values exclusively from each other.
Landscapes
- Engineering & Computer Science (AREA)
- Signal Processing (AREA)
- Physics & Mathematics (AREA)
- Acoustics & Sound (AREA)
- Multimedia (AREA)
- Health & Medical Sciences (AREA)
- Otolaryngology (AREA)
- General Health & Medical Sciences (AREA)
- Computational Linguistics (AREA)
- Quality & Reliability (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Human Computer Interaction (AREA)
- Telephone Function (AREA)
- Cable Transmission Systems, Equalization Of Radio And Reduction Of Echo (AREA)
- Filters That Use Time-Delay Elements (AREA)
Abstract
Description
<数1>
c1(k)=1-c0(k)
に従って与える。
係数乗算器31020は、入力された受信信号サンプルに係数値c0を乗算し、その積を加算器31031に伝達する。遅延素子31011は受信信号サンプルを1サンプル遅延させて、係数乗算器31021及び遅延素子31012に伝達する。
図5に、係数乗算器31020、31021、及び31022の係数c0(k)、c1(k)及びc2(k)の例を示す。係数c0(k)、c1(k)及びc2(k)が排他的に1をとることによって、それぞれの係数乗算器に対応した遅延を受けた受信信号が、遅延受信信号として出力端子3104に得られる。図3(A)に対応した (B)と(C)のように、図5に示した係数c0(k)、c1(k)及びc2(k)がゼロと非ゼロの値を変化する際に、滑らかに変化するように設定することもできる。滑らかな値の変化によって、受信信号と遅延受信信号を相互に切り替える際に生じる音像が滑らかに移動し、音像移動を知覚されにくくする効果がある。また、前記切り替えに際して、クリック音が知覚されることを避ける効果もある。これらは、主観音質の改善に有効である。
係数乗算器31020は、入力された受信信号サンプルに係数値c0を乗算し、その積を加算器31031に伝達する。遅延素子31011は受信信号サンプルを1サンプル遅延させて、係数乗算器31021及び遅延素子31012に伝達する。
<数2>
P1+P2=C
という関係が成立しなければならない。ここに、Cは正定数である。従って、振幅補正前に、受信信号1と受信信号2の電力がそれぞれP1バー[dB]とP2バー[dB]であるときに、振幅補正後の受信信号1と受信信号2の電力P1[dB]とP2[dB]は、
<数3>
P1 =P1バー - ΔP/2
P2 =P2バー - ΔP/2
の関係を満足しなければならない。ここに、ΔPは、電力補正量である。このため、乗算器4101の係数g0バーとg1バーの値は、数3から、
<数4>
g0バー = 10-ΔPi/40
g1バー = 10ΔPi/40
によって決定することができる。但し、ΔPiは、受信信号をiサンプル遅延させたときに、これを補償するために必要になる電力補償係数である。
3,4 スピーカ
5,6,7,8 エコー
9,10 マイクロフォン
11 話者
12,13 送信信号
14,15 混在信号
16,17 信号処理装置の出力信号
18,19 ディジタル・アナログ変換器
20,21 アナログ・ディジタル変換器
121,122,123,124 適応フィルタ
125,126,127,128 擬似エコー
129,130 減算器
300,301 遅延処理回路
310,320 フィルタ
330,430 クロック変更回路
350,351 分析回路
400 振幅補正回路
410,420 振幅処理回路
500,501 非線形処理回路
510,520,530,540 非線形振幅処理回路
511,531,3100,3105,4100,4105 入力端子
513 極性判定回路
514 乗算器
515,3103 加算器
516,536,3104,4104 出力端子
600,610 周波数分析合成回路
1000 コンピュータ
3101 遅延素子
3102,4101 係数乗算器
Claims (33)
- 複数の受信信号を受信し、前記複数の受信信号により生成される複数のエコーから、前記複数の受信信号を入力とする複数の適応フィルタによって生成された擬似エコーを差し引くことによって、前記複数のエコーを低減する信号処理方法であって、
前記複数の受信信号のうち少なくとも1つの受信信号を遅延させて遅延受信信号を生成し、
前記受信信号と前記遅延受信信号とを前記適応フィルタに入力して擬似エコーを生成し、
前記受信信号と前記遅延受信信号とを前記適応フィルタに入力する頻度を前記複数の受信信号の定位の変化に対する知覚感度に基づいて制御する
ことを特徴とする信号処理方法。 - 前記遅延受信信号のうち少なくとも一つの信号が振幅補正された振幅補正遅延受信信号であることを特徴とする請求項1に記載の信号処理方法。
- 前記複数の適応フィルタに入力する信号のうち少なくとも一つの信号が非線形処理された非線形処理信号であることを特徴とする請求項1又は2に記載の信号処理方法。
- 前記受信信号を複数の周波数成分に分解し、該複数の周波数成分毎に遅延させて遅延受信信号を生成することを特徴とする請求項1から請求項3のいずれか1項に記載の信号処理方法。
- 前記定位の変化に対する知覚感度が、受信信号の類似度によって求められることを特徴とする請求項1から請求項4のいずれか1項に記載の信号処理方法。
- 前記定位の変化に対する知覚感度が、受信信号のパワーによって求められることを特徴とする請求項1から請求項4のいずれか1項に記載の信号処理方法。
- 前記遅延受信信号の前記受信信号に対する相対的な遅延が、時間と共に変化する複数の値をとるように生成することを特徴とする請求項1から請求項6のいずれか1項に記載の信号処理方法。
- 前記相対的な遅延が、サンプリング周期の整数倍であることを特徴とする請求項7に記載の信号処理方法。
- 前記遅延受信信号の生成は、ゼロと非ゼロの値を交互にとる複数の時変係数を有するフィルタで受信信号を処理することによって行うことを特徴とする請求項1から請求項8のいずれか1項に記載の信号処理方法。
- 前記複数の時変係数が、相互に排他的にゼロの値をとることを特徴とする請求項9記載の信号処理方法。
- 前記複数の時変係数が、相互に排他的に非ゼロの値をとることを特徴とする請求項9又は10に記載の信号処理方法。
- 複数の受信信号を受信し、前記複数の受信信号により生成される複数のエコーから、前記複数の受信信号を入力とする複数の適応フィルタによって生成された擬似エコーを差し引くことによって、前記複数のエコーを低減する信号処理装置であって、
前記複数の受信信号のうち少なくとも1つの受信信号を遅延させて遅延受信信号を生成する線形処理回路と、
前記受信信号と前記遅延受信信号とを受けて擬似エコーを生成する適応フィルタと、
前記擬似エコーを前記混在信号から差し引くことによってエコーが低減された信号を生成する複数の減算器と、
前記複数の受信信号の定位の変化に対する知覚感度を求める分析回路と、
を具備し、
前記知覚感度に基づいて前記受信信号と前記遅延受信信号とを前記適応フィルタに入力する頻度を制御し、前記複数の減算器の出力を最小とするように前記複数の適応フィルタの係数を制御することを特徴とする信号処理装置。 - 前記遅延受信信号のうち少なくとも一つの信号を振幅補正して振幅補正遅延受信信号を生成する振幅補正回路を具備することを特徴とする請求項12に記載の信号処理装置。
- 前記複数の適応フィルタに入力する信号のうち少なくとも一つの信号を非線形処理して非線形処理信号を生成する非線形処理回路を具備することを特徴とする請求項12又は13に記載の信号処理装置。
- 前記受信信号を複数の周波数成分に分解する周波数分析回路と、
前記複数の周波数成分毎に遅延させて遅延受信信号を生成する線形処理回路と、
を具備することを特徴とする請求項12から請求項14のいずれか1項に記載の信号処理装置。 - 前記分析回路は、前記定位の変化に対する知覚感度を、受信信号の類似度によって求めることを特徴とする請求項12から請求項15のいずれか1項に記載の信号処理装置。
- 前記分析回路は、前記定位の変化に対する知覚感度が、受信信号のパワーによって求めることを特徴とする請求項12から請求項15のいずれか1項に記載の信号処理装置。
- 前記線形処理回路は、前記遅延受信信号の前記受信信号に対する相対的な遅延が、時間と共に変化する複数の値をとるような処理を行うことを特徴とする請求項12から請求項17のいずれか1項に記載の信号処理装置。
- 前記線形処理回路は、前記相対的な遅延がサンプリング周期の整数倍となるような処理を行うことを特徴とする請求項18に記載の信号処理装置。
- 前記線形処理回路は、ゼロと非ゼロの値を交互にとる複数の時変係数を有するフィルタを具備することを特徴とする請求項12から請求項19のいずれか1項に記載の信号処理装置。
- 前記複数の時変係数が、相互に排他的にゼロの値をとることを特徴とする請求項20記載の信号処理装置。
- 前記複数の時変係数が、相互に排他的に非ゼロの値をとることを特徴とする請求項20又は21に記載の信号処理装置。
- コンピュータに、複数の受信信号を受信する受信する受信処理と、前記複数の受信信号により発生する複数のエコーを低減するエコー低減処理とを実行させる信号処理プログラムであって、
前記複数の受信信号のうち少なくとも1つの受信信号を遅延させて遅延受信信号を生成する遅延受信信号生成処理と、
前記受信信号と前記遅延受信信号とを前記適応フィルタに入力して擬似エコーを生成する擬似エコー生成処理と、
前記複数の受信信号それぞれから前記擬似エコーを差し引く擬似エコー差引き処理と、
を含み、
前記受信信号と前記遅延受信信号とを前記適応フィルタに入力する頻度を前記複数の受信信号の定位の変化に対する知覚感度に基づいて制御することを特徴とする信号処理プログラム。 - 前記遅延受信信号のうち少なくとも一つの信号が振幅補正された振幅補正遅延受信信号であることを特徴とする請求項1に記載の信号処理プログラム。
- 前記複数の適応フィルタに入力する信号のうち少なくとも一つの信号が非線形処理された非線形処理信号であることを特徴とする請求項23又は25に記載の信号処理プログラム。
- 前記受信信号を複数の周波数成分に分解し、該複数の周波数成分毎に遅延させて遅延受信信号を生成することを特徴とする請求項23から請求項25のいずれか1項に記載の信号処理プログラム。
- 前記定位の変化に対する知覚感度が、受信信号の類似度に基づいて求められることを特徴とする請求項23から請求項26のいずれか1項に記載の信号処理プログラム。
- 前記定位の変化に対する知覚感度が、受信信号のパワーに基づいて求められることを特徴とする請求項23から請求項26のいずれか1項に記載の信号処理プログラム。
- 前記遅延受信信号の前記受信信号に対する相対的な遅延が、時間と共に変化する複数の値をとるように生成することを特徴とする請求項23から請求項28のいずれか1項に記載の信号処理プログラム。
- 前記相対的な遅延が、サンプリング周期の整数倍であることを特徴とする請求項29に記載の信号処理プログラム。
- 前記遅延受信信号の生成は、ゼロと非ゼロの値を交互にとる複数の時変係数を有するフィルタで受信信号を処理することによって行うことを特徴とする請求項23から請求項30のいずれか1項に記載の信号処理プログラム。
- 前記複数の時変係数が、相互に排他的にゼロの値をとることを特徴とする請求項31記載の信号処理プログラム。
- 前記複数の時変係数が、相互に排他的に非ゼロの値をとることを特徴とする請求項31又は32に記載の信号処理プログラム。
Priority Applications (3)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN200980137410.8A CN102165708B (zh) | 2008-09-26 | 2009-09-15 | 信号处理方法、信号处理装置及信号处理程序 |
US13/121,144 US9154635B2 (en) | 2008-09-26 | 2009-09-15 | Signal processing method, signal processing device, and signal processing program |
JP2010530815A JP5472643B2 (ja) | 2008-09-26 | 2009-09-15 | 信号処理方法、信号処理装置、および信号処理プログラム |
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
JP2008247273 | 2008-09-26 | ||
JP2008-247273 | 2008-09-26 |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2010035658A1 true WO2010035658A1 (ja) | 2010-04-01 |
Family
ID=42059659
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/JP2009/066068 WO2010035658A1 (ja) | 2008-09-26 | 2009-09-15 | 信号処理方法、信号処理装置、および信号処理プログラム |
Country Status (4)
Country | Link |
---|---|
US (1) | US9154635B2 (ja) |
JP (1) | JP5472643B2 (ja) |
CN (1) | CN102165708B (ja) |
WO (1) | WO2010035658A1 (ja) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102223157A (zh) * | 2010-04-16 | 2011-10-19 | 联芯科技有限公司 | 残留回声消除器和残留回声消除方法 |
Families Citing this family (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US9363386B2 (en) * | 2011-11-23 | 2016-06-07 | Qualcomm Incorporated | Acoustic echo cancellation based on ultrasound motion detection |
WO2014103066A1 (ja) * | 2012-12-28 | 2014-07-03 | 共栄エンジニアリング株式会社 | 音源分離方法、装置、及びプログラム |
JP2015084466A (ja) * | 2013-10-25 | 2015-04-30 | ソニー株式会社 | サンプリングポイント調整装置および方法、並びにプログラム |
JP6670224B2 (ja) * | 2016-11-14 | 2020-03-18 | 株式会社日立製作所 | 音声信号処理システム |
JP6845202B2 (ja) * | 2018-10-11 | 2021-03-17 | ファナック株式会社 | 数値制御方法及び処理装置 |
CN110265050B (zh) * | 2019-05-29 | 2021-06-04 | 广州小鹏汽车科技有限公司 | Aec音频控制系统及其时钟协商方法 |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2000078061A (ja) * | 1998-09-02 | 2000-03-14 | Nec Corp | 多チャネルエコー除去方法及び装置 |
JP2000308176A (ja) * | 1999-02-17 | 2000-11-02 | Nippon Telegr & Teleph Corp <Ntt> | 多チャンネル拡声通信エコー消去方法及びその装置並びにプログラム記録媒体 |
JP2002261661A (ja) * | 2001-02-27 | 2002-09-13 | Nippon Telegr & Teleph Corp <Ntt> | 多チャネルエコー消去方法、その装置、そのプログラム及びその記録媒体 |
Family Cites Families (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2792252B2 (ja) | 1991-03-14 | 1998-09-03 | 日本電気株式会社 | 多チャンネルエコー除去方法および装置 |
JP3397269B2 (ja) * | 1994-10-26 | 2003-04-14 | 日本電信電話株式会社 | 多チャネル反響消去方法 |
JP2850826B2 (ja) * | 1996-01-11 | 1999-01-27 | 日本電気株式会社 | 多チャネルエコー除去方法及び装置 |
JP3152292B2 (ja) * | 1997-04-15 | 2001-04-03 | 日本電気株式会社 | 多チャネルエコー除去方法及び装置 |
US20020181698A1 (en) | 2000-01-19 | 2002-12-05 | Shinya Takahashi | Sound communication device and echo processor |
JP3568922B2 (ja) | 2001-09-20 | 2004-09-22 | 三菱電機株式会社 | エコー処理装置 |
US7536007B2 (en) | 2004-12-30 | 2009-05-19 | Alcatel Lucent | Echo cancellation using adaptive IIR and FIR filters |
CN100550951C (zh) | 2005-09-01 | 2009-10-14 | 华为技术有限公司 | 一种回声消除装置和方法 |
CN101043560A (zh) | 2006-03-22 | 2007-09-26 | 北京大学深圳研究生院 | 回声消除器及回声消除方法 |
CN101106405A (zh) | 2006-07-12 | 2008-01-16 | 北京大学深圳研究生院 | 回声消除器、回声消除方法及其双端通话检测系统 |
-
2009
- 2009-09-15 JP JP2010530815A patent/JP5472643B2/ja not_active Expired - Fee Related
- 2009-09-15 US US13/121,144 patent/US9154635B2/en not_active Expired - Fee Related
- 2009-09-15 WO PCT/JP2009/066068 patent/WO2010035658A1/ja active Application Filing
- 2009-09-15 CN CN200980137410.8A patent/CN102165708B/zh not_active Expired - Fee Related
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2000078061A (ja) * | 1998-09-02 | 2000-03-14 | Nec Corp | 多チャネルエコー除去方法及び装置 |
JP2000308176A (ja) * | 1999-02-17 | 2000-11-02 | Nippon Telegr & Teleph Corp <Ntt> | 多チャンネル拡声通信エコー消去方法及びその装置並びにプログラム記録媒体 |
JP2002261661A (ja) * | 2001-02-27 | 2002-09-13 | Nippon Telegr & Teleph Corp <Ntt> | 多チャネルエコー消去方法、その装置、そのプログラム及びその記録媒体 |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102223157A (zh) * | 2010-04-16 | 2011-10-19 | 联芯科技有限公司 | 残留回声消除器和残留回声消除方法 |
CN102223157B (zh) * | 2010-04-16 | 2013-12-18 | 联芯科技有限公司 | 残留回声消除器和残留回声消除方法 |
Also Published As
Publication number | Publication date |
---|---|
JP5472643B2 (ja) | 2014-04-16 |
CN102165708A (zh) | 2011-08-24 |
JPWO2010035658A1 (ja) | 2012-02-23 |
US20110175676A1 (en) | 2011-07-21 |
US9154635B2 (en) | 2015-10-06 |
CN102165708B (zh) | 2014-06-25 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
JP5472642B2 (ja) | 信号処理方法、信号処理装置、および信号処理プログラム | |
JP5472643B2 (ja) | 信号処理方法、信号処理装置、および信号処理プログラム | |
JP3152292B2 (ja) | 多チャネルエコー除去方法及び装置 | |
JP4644715B2 (ja) | 音響的エコー消去用オーディオシステムと方法 | |
KR101250124B1 (ko) | 에코 억제 필터를 위한 제어 정보를 계산하는 장치 및 방법 및 지연 값을 계산하는 장치 및 방법 | |
JP5288723B2 (ja) | マルチチャネルの反響補償 | |
US7693291B2 (en) | Multi-channel frequency-domain adaptive filter method and apparatus | |
JP2007068125A (ja) | 信号処理の方法及び装置並びにコンピュータプログラム | |
WO2007148509A1 (ja) | エコー抑圧装置 | |
JPWO2009051132A1 (ja) | 信号処理システムと、その装置、方法及びそのプログラム | |
US6577731B1 (en) | Method and apparatus of cancelling multi-channel echoes | |
Molés-Cases et al. | Personal sound zones by subband filtering and time domain optimization | |
JP2004349806A (ja) | 多チャネル音響エコー消去方法、その装置、そのプログラム及びその記録媒体 | |
JP5937451B2 (ja) | エコー消去装置、エコー消去方法及びプログラム | |
JP4549535B2 (ja) | 計算を簡単にした周波数濾波用デジタル処理装置 | |
JP4041770B2 (ja) | 音響エコー消去方法、その装置、プログラム及びその記録媒体 | |
JPH08110794A (ja) | 信号分離方法 | |
JPH08321796A (ja) | 音響エコーキャンセラ | |
US9787825B1 (en) | Cyclically variable signal delays for acoustic echo cancellation | |
JP2850826B2 (ja) | 多チャネルエコー除去方法及び装置 | |
JP6075783B2 (ja) | エコー消去装置、エコー消去方法及びプログラム | |
WO2023214571A1 (ja) | ビームフォーミング方法、ビームフォーミングシステム | |
JPH07154307A (ja) | 音響反響除去装置 | |
JPS62154826A (ja) | エコ−制御方式 |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
WWE | Wipo information: entry into national phase |
Ref document number: 200980137410.8 Country of ref document: CN |
|
121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 09816073 Country of ref document: EP Kind code of ref document: A1 |
|
WWE | Wipo information: entry into national phase |
Ref document number: 2010530815 Country of ref document: JP |
|
WWE | Wipo information: entry into national phase |
Ref document number: 13121144 Country of ref document: US |
|
NENP | Non-entry into the national phase |
Ref country code: DE |
|
122 | Ep: pct application non-entry in european phase |
Ref document number: 09816073 Country of ref document: EP Kind code of ref document: A1 |