JP5950199B2 - Loudspeaker - Google Patents

Description

  The present invention relates to a loudspeaker device such as an interphone used in a house or office.

  A conventional loudspeaker device is described in Patent Document 1. The conventional example described in Patent Document 1 includes an echo canceller and a voice switch in order to prevent unpleasant echoes and howling. When the echo canceller immediately after the start of the call has not converged, it is uncomfortable to operate in a fixed mode in which the total amount of loss (total amount of loss) inserted into the signal path by the voice switch is fixed to a sufficiently large initial value. Echo and howling are suppressed. On the other hand, when the echo canceller is sufficiently converged, a two-way simultaneous call is realized by operating in an update mode in which the voice switch updates the total loss amount as needed.

  Furthermore, in the conventional example described in Patent Document 1, the gain (feedback gain) of the feedback path that causes echo and howling is obtained for each frequency band, and the largest one of the feedback gains obtained for each frequency band is used as the estimated value of the feedback gain. Selected. Therefore, the estimation accuracy of the feedback gain can be improved as compared with the case where the feedback gain is estimated from the time average power of the reference signal (transmitted signal and received signal).

JP 2007-194740 A

  By the way, in the conventional example described in Patent Document 1, when obtaining the feedback gain for each frequency band, the signal level for each frequency band is obtained by subjecting the reference signal to discrete Fourier transform processing. However, the discrete Fourier transform processing requires trigonometric functions and square root operations, and an inexpensive signal processing circuit (such as a DSP) that performs fixed-point arithmetic used for a voice switch has a large amount of processing, so the processing time is long. There is a problem of becoming. On the other hand, when a signal processing circuit capable of high-speed arithmetic processing is used, there is a problem that the manufacturing cost of the loudspeaker device increases.

  The present invention has been made in view of the above problems, and an object thereof is to improve the estimation accuracy of feedback gain while suppressing an increase in manufacturing cost.

The loudspeaker apparatus of the present invention transmits a microphone and a speaker, a reception side signal path for transmitting a reception signal transmitted from the other party's telephone terminal to the speaker, and a transmission signal collected by the microphone. A voice switch for switching a call state between reception and transmission by inserting a loss in a transmission side signal path to be sent to the call terminal on the other side, an echo canceller for suppressing acoustic echo caused by acoustic coupling of the microphone and the speaker, and The voice switch includes a transmission side loss insertion unit that inserts a loss into the signal path on the transmission side, a reception side loss insertion unit that inserts a loss into the signal path on the reception side, a transmission side, and An insertion loss amount control unit that controls a loss amount to be inserted from each loss insertion unit on the receiving side, and the insertion loss amount control unit is connected to an output point of the reception side loss insertion unit. With estimating the acoustic side feedback gain path for feeding back through the sound echo path to an input point of the transmitting end losses inserting portion, the receiving side through a line echo path from the output point of the transmitting end losses insertion portion A feedback gain estimator that estimates the line-side feedback gain of the path that returns to the input point of the loss insertion unit , and calculates the total amount of loss to be inserted into the closed loop based on the estimated values of the feedback gains on the acoustic and line sides A total loss amount calculation unit that monitors a transmission signal and a reception signal to estimate a call state, and the transmission side loss insertion unit and the reception side according to the estimation result and a calculated value of the total loss amount calculation unit In the loudspeaker apparatus having the insertion loss amount distribution processing unit for determining the distribution of each insertion loss amount of the side loss insertion unit, the feedback gain estimation unit transmits the signal inherently to each feedback path on the acoustic side and the line side Transmission time difference to compensate for time difference A signal level calculation unit for each frequency band that obtains a signal level for each frequency band by discriminating the signal corrected by the transmission time difference correction unit into a plurality of frequency bands, and a signal level calculation for each frequency band And an estimated value calculation unit for calculating an estimated value of feedback gain for each frequency band from the signal level calculated by the frequency unit, wherein the signal level calculation unit for each frequency band has a plurality of band passes having different frequency bands as pass bands. This is caused by the first echo canceller that discriminates the signal into a plurality of frequency bands by a filter and suppresses the acoustic echo, and acoustic coupling at the other party's call terminal or wraparound of the transmitted signal via the line echo path A second echo canceller for suppressing line echo, wherein the first echo canceller is an impulse of the acoustic echo path. An adaptive filter that adaptively identifies a response and estimates the acoustic echo from an input signal to the acoustic echo path, and a subtraction that subtracts the acoustic echo estimated by the adaptive filter from an output signal from the acoustic echo path The second echo canceller adaptively identifies an impulse response of the line echo path and estimates the line echo from an input signal to the line echo path, and the adaptive filter A subtractor that subtracts the estimated line echo from an output signal from the line echo path, and the adaptive filter of the first echo canceller and the adaptive filter of the second echo canceller include the insertion loss. depending on the allocation of the insertion loss amount distribution processing section determines, to adjust the timing for updating the filter coefficients And butterflies.

  In this loudspeaker, it is preferable that the feedback gain estimation unit alternately performs the process of estimating the acoustic side feedback gain and the process of estimating the line side feedback gain at regular intervals.

  In this loudspeaker apparatus, the signal level calculation unit for each frequency band detects a double talk state in which a voice component is applied in the middle of a path where the echo canceller estimates a feedback gain. It is preferable to substitute the signal level obtained by the previous filtering process by stopping the operation.

  In this loudspeaker communication apparatus, the echo canceller includes a transmission time difference detection unit that detects a difference in signal transmission time inherent in each of the feedback paths on the acoustic side and the line side, and the transmission time difference correction unit includes the transmission time difference The transmission time difference is preferably corrected using the transmission time difference detected by the detecting means.

  In this loudspeaker, when the estimated value calculation unit estimates a feedback gain for each frequency band from the signal level calculated by the signal level calculation unit for each frequency band, the signal level becomes a predetermined threshold value or less. It is preferable not to calculate an estimated value of the feedback gain in the frequency band.

  In this loudspeaker apparatus, it is preferable that the total loss amount calculation unit calculates a total sum of the loss amounts by adding a predetermined margin to the estimated value of the feedback gain estimated by the feedback gain estimation unit.

  In the voice communication device, the total loss amount calculation unit calculates a total sum of the total loss amounts for each frequency band in which the estimated value of the feedback gain is obtained, and the insertion loss amount distribution processing unit It is preferable that the distribution of the insertion loss amounts of the side loss insertion unit and the reception side loss insertion unit is determined for each frequency band.

  In this loudspeaker apparatus, the adaptive filter of the first echo canceller is configured such that when the insertion loss amount distribution processing unit distributes a relatively high insertion loss amount to the transmission side loss insertion unit, the filter coefficient And the adaptive filter of the second echo canceller updates the filter coefficient when the insertion loss amount distribution processing unit distributes a relatively high insertion loss amount to the reception-side loss insertion unit. It is preferable.

  In this loudspeaker, when each of the adaptive filters of the first echo canceller and the second echo canceller is updating the filter coefficient, the insertion loss amount distribution processing unit is configured to transmit the transmission side and the reception side. It is preferable to continue to update the filter coefficient until a certain period elapses after the insertion loss amount is equally distributed to the respective loss insertion portions.

  In the loudspeaker according to the present invention, the signal level calculation unit for each frequency band discriminates the reference signal into a plurality of frequency bands by a plurality of band pass filters, and therefore feedback is performed while suppressing an increase in manufacturing cost compared to the conventional example. There is an effect that the estimation accuracy of the gain can be improved.

It is a block diagram which shows the insertion loss amount control part in embodiment of the loudspeaker apparatus which concerns on this invention. It is a block diagram which shows the same as the above. It is a flowchart for demonstrating operation | movement of the signal level calculation part classified by frequency band in a prior art example.

  DESCRIPTION OF EMBODIMENTS Hereinafter, embodiments of a loudspeaker device according to the present invention will be described in detail with reference to the drawings.

  In the present embodiment, as shown in FIG. 2, a microphone 1, a speaker 2, a 2-wire to 4-wire conversion circuit 3, a microphone amplifier G1, a line output amplifier G2, a line input amplifier G3, a speaker amplifier G4, amplifiers G5 and G6, audio The switch 10 includes first and second echo cancellers 30A and 30B. The line consists of a two-wire transmission line.

The first echo canceller 30A has a conventionally known configuration including an adaptive filter 31A and a subtracter 32A. Adaptive filter 31A identifies the impulse response of the feedback path (acoustic echo path) H _AC formed by the acoustic coupling between the speaker 2 microphone 1 adaptively. The subtractor 32A suppresses the acoustic echo by subtracting the pseudo echo component (acoustic echo) estimated by the adaptive filter 31A from the output signal of the microphone amplifier G1. The second echo canceller 30B also has a conventionally known configuration including an adaptive filter 31B and a subtractor 32B. The adaptive filter 31B has a feedback path (line echo path) formed by reflection due to impedance mismatch between the 2-wire-to-wire conversion circuit 3 and the transmission path and acoustic coupling between the speaker and the microphone at the other party's call terminal. ) Adaptively identify the impulse response of H _LIN . The subtractor 32B suppresses the line echo by subtracting the pseudo echo component (line echo) estimated by the adaptive filter 31B from the received signal.

The voice switch 10 includes a transmission side loss insertion unit 11 that inserts a loss into the signal path on the transmission side, a reception side loss insertion unit 12 that inserts a loss into the signal path on the reception side, and a transmission side and a reception side. And an insertion loss amount control unit 13 for controlling the loss amount inserted from each of the loss insertion units 11 and 12. The insertion loss amount control unit 13 includes a feedback gain estimation unit 14, a total loss calculation unit 15, and an insertion loss amount distribution processing unit 16. Feedback gain estimation unit 14, the path to be fed back from the output point Rout of the receiving-side loss insertion portion 12 into the input point Tin of the transmitting end losses insertion portion 11 via the acoustic echo path H _AC (hereinafter referred to as acoustic-side feedback path) The acoustic side feedback gain α is estimated. Further, the feedback gain estimation unit 14 returns a path (hereinafter referred to as a line side feedback path) from the output point Tout of the transmission side loss insertion part 11 to the input point Rin of the reception side loss insertion part 12 via the line echo path _HLIN. ) Is estimated. The total loss calculation unit 15 calculates the total loss amount to be inserted into the closed loop based on the estimated values α ′ and β ′ of the feedback gains α and β on the acoustic side and the line side (insertion of the transmission side loss insertion unit 11 The sum of the loss amount and the insertion loss amount of the reception-side loss insertion unit 12 is calculated. The insertion loss amount distribution processing unit 16 monitors the transmission signal and the reception signal to estimate the call state, and according to the estimation result and the calculated value of the total loss amount calculation unit 15, the transmission loss insertion unit 11 and the reception side The distribution of each insertion loss amount of the side loss insertion unit 12 is determined. The first and second echo cancellers 30A and 30B and the voice switch 10 in this embodiment control the DSP (Digital Signal Processor) hardware by software (program) for the echo canceller and voice switch. It is realized by. Therefore, the input / output signals (received signal and transmitted signal) of the voice switch 10 and the first and second echo cancellers 30A and 30B in the following description are sampled at a predetermined sampling period, and are quantized by the A / D converter. It has become.

  The total loss calculation unit 15 calculates a total loss Lt necessary to obtain a desired gain margin MG from the estimated values α ′ and β ′ of the acoustic feedback gain α and the line feedback gain β, and the value Lt Is output to the insertion loss amount distribution processing unit 16.

  The insertion loss amount distribution processing unit 16 monitors the input / output signals of the transmission side loss insertion unit 11 and the input / output signals of the reception side loss insertion unit 12, and the magnitude relationship between the power levels of these signals and the presence or absence of a voice signal, etc. The call state (receiving state, transmitting state, etc.) is determined from the information. Furthermore, the insertion loss amount distribution processing unit 16 distributes the total loss amount Lt to the transmission side loss insertion unit 11 and the reception side loss insertion unit 12 at a rate corresponding to the determined call state. Adjust the amount of insertion loss.

  By the way, the total loss amount calculation unit 15 in the present embodiment has two operation modes, an update mode and a fixed mode. In the update mode, as described above, the sum of the loss amounts to be inserted into the closed loop is calculated and adaptively updated based on the estimated values α ′ and β ′ of the feedback gains α and β estimated by the feedback gain estimation unit 14 as described above. It is an operation mode. On the other hand, the fixed mode is an operation mode in which the total loss amount is fixed to a predetermined initial value. Then, the total loss calculation unit 15 performs a period from the start of the call with the other party's call terminal until the filter coefficients in the adaptive filters 31A and 31B of the first and second echo cancellers 30A and 30B sufficiently converge. Operates in fixed mode. Then, during the period after the filter coefficients have sufficiently converged, the total loss calculation unit 15 operates in the update mode. That is, in the total loss amount calculation unit 15, the estimated values α ′ and β ′ of the acoustic side feedback gain α and the line side feedback gain β both continue for a predetermined time (several hundred milliseconds) from the start of the call for a predetermined threshold value. It is considered that the filter coefficient has sufficiently converged when it falls below ε (for example, a value that is 10 to 15 dB smaller than the estimated values α ′ and β ′ at the start of the call). The total loss amount calculation unit 15 operates in a fixed mode in which the total loss amount is fixed to the initial value before the time point, and after the time point, the total loss amount is calculated based on the estimated values α ′ and β ′. Switch the operation mode to the update mode for adaptive update. Note that the initial value of the total loss amount in the fixed mode is set to a value sufficiently larger than the total loss amount updated as needed in the update mode.

  Thus, when the filter coefficient immediately after the start of the call is not sufficiently converged, the initial total loss amount set to a sufficiently large value is inserted into the closed loop, resulting in unpleasant echoes and howling. Stable half-duplex call can be realized by suppressing the above. In addition, when the time has elapsed since the start of the call and the filter coefficient has sufficiently converged, the operation mode of the total loss calculation unit 15 switches to the update mode, and the total loss inserted in the closed loop is less than the initial value. Since it decreases to a sufficiently low value, two-way simultaneous calls can be realized.

  Next, the feedback gain estimation unit 14 will be described in more detail. FIG. 1 is a block diagram showing in detail the configuration of the feedback gain estimation unit 14 in the insertion loss amount control unit 13. The feedback gain estimation unit 14 includes an acoustic side feedback gain estimation unit 14A that estimates the acoustic side feedback gain α, and a line side feedback gain estimation unit 14B that estimates the line side feedback gain β.

  The acoustic side feedback gain estimator 14A and the line side feedback gain estimator 14B are configured such that, at a certain time n, the input point Tin and the output point Tout of the transmission side loss insertion unit 11 on the transmission path, and the reception side loss insertion unit on the reception path The transmission signal and reception signal Tin (n), Tout (n), Rin (n), Rout (n) are taken in from the 12 input points Rin and output points Rout, respectively, and the acoustic feedback gain α (n) and line side Estimated values α ′ (n) and β ′ (n) for the feedback gain β (n) are output and passed to the total loss calculating unit 15.

  The acoustic-side feedback gain estimation unit 14A is referred to from the reception-side block that processes the reference signal (reception signal) referenced from the output point Rout of the reception-side loss insertion unit 12, and from the input point Tin of the transmission-side loss insertion unit 11. It has a transmission side block for processing a reference signal (transmission signal), and an estimated value calculation unit 54A. The transmission side block includes a reference signal storage unit 52C, a signal level calculation unit 53C for each frequency band, and the like. The reference signal storage unit 52C stores the referenced reference signal, and the signal level calculation unit 53C for each frequency band includes a plurality of bandpass filters (not shown) having different frequency bands as passbands. The signal level for each frequency band is obtained by discriminating the reference signal read from the storage unit 52C into a plurality of frequency bands.

The receiving side block includes a transmission time difference correction unit 51A, a reference signal storage unit 52A, a signal level calculation unit 53A for each frequency band, and the like. Transmission time difference correction unit 51A includes acoustic side feedback path including the acoustical side echo path H _AC corrects the difference in signal transmission time with unique. The reference signal storage unit 52A stores the reference signal (received signal) after the signal transmission time difference is corrected. The signal level calculation unit 53A for each frequency band has a plurality of band pass filters (not shown) whose pass bands are different frequency bands, and discriminates the reference signal read from the reference signal storage unit 52A into a plurality of frequency bands. Thus, the signal level for each frequency band is obtained.

  The estimated value calculation unit 54A calculates each frequency band from the ratio between the signal level for each frequency band calculated by the signal level calculation unit for each frequency band 53C and the signal level for each frequency band calculated by the signal level calculation unit for each frequency band 53A. The estimated value α ′ of the acoustic feedback gain for each is calculated. Then, the estimated value α ′ of the acoustic feedback gain for each frequency band calculated by the estimated value calculating unit 54A is input to the total loss calculating unit 15.

  On the other hand, the line-side feedback gain estimator 14B includes a transmission side block that processes a reference signal (transmission signal) referenced from the output point Tout of the transmission side loss insertion unit 11, and an input point of the reception side loss insertion unit 12. A receiving side block that processes a reference signal (received signal) referenced from Rin, and an estimated value calculation unit 54B. The receiving side block includes a reference signal storage unit 52D, a signal level calculation unit 53D for each frequency band, and the like. The reference signal storage unit 52D stores the referenced reference signal. The signal level calculation unit 53D for each frequency band has a plurality of band pass filters (not shown) having different frequency bands as pass bands, and discriminates the reference signal read from the reference signal storage unit 52D into a plurality of frequency bands. Thus, the signal level for each frequency band is obtained.

The transmission side block includes a transmission time difference correction unit 51B, a reference signal storage unit 52B, a signal level calculation unit 53B for each frequency band, and the like. The transmission time difference correction unit 51B corrects the difference in signal transmission time inherent in the line-side feedback path including the line-side echo path _HLIN . The reference signal storage unit 52B stores the reference signal (transmission signal) after the difference in signal transmission time is corrected. The signal level calculation unit 53B for each frequency band has a plurality of band pass filters (not shown) having different frequency bands as pass bands, and discriminates the reference signal read from the reference signal storage unit 52B into a plurality of frequency bands. Thus, the signal level for each frequency band is obtained.

  The estimated value calculation unit 54B calculates each frequency band from the ratio between the signal level for each frequency band calculated by the signal level calculation unit 53D for each frequency band and the signal level for each frequency band calculated by the signal level calculation unit 53C for each frequency band. The estimated value β ′ of the line-side feedback gain for each is calculated. Then, the estimated value β ′ of the line-side feedback gain for each frequency band calculated by the estimated value calculating unit 54B is input to the total loss calculating unit 15. The above-described units included in the voice switch 10 are realized by controlling DSP hardware with a dedicated program, and are processed by the feedback gain estimation units 14A and 14B on the acoustic side and the line side. Are all treated as digital data obtained by sampling and quantizing analog transmission signals and reception signals.

  Here, in the estimated value calculation units 54A and 54B, when calculating the estimated value of the feedback gain for each frequency band from the signal level calculated by the signal level calculation units 53A to 53D for each frequency band, the signal level of the reference signal is a predetermined level. It is preferable not to calculate the estimated value of the feedback gain in the frequency band that is equal to or lower than the threshold value. In other words, if feedback gain estimation processing is performed on noise components such as ambient noise (background noise), an estimated value far from the actual feedback gain may be calculated. On the other hand, the estimation accuracy of the feedback gain can be improved by not estimating the feedback gain.

Its system until receiving the signal output from the output point Rout of the receiving-side loss insertion portion 12 reaches the input point Tin of the transmitting end losses insertion portion 11 via the acoustic side feedback path including an acoustic echo path H _AC Inherent transmission time is required. Therefore, the transmission time difference correction unit 51A is set in advance by measuring the time for the single pulse generated from the output point Rout of the receiving side loss insertion unit 12 to reach the input point Tin of the transmission side loss insertion unit 11. only the system transmission time D _alpha of which had been delays the reference signal from the output point Rout of the receiving-side loss insertion portion 12. For example, when the sampling frequency of the reference signal is 8 [kHz] and the measured delay time is 12 [msec], a delay processing signal storage unit for 8 × 12 = 96 data (separate from the reference signal storage units 52A to 52D). FIFO (First In First Out) type signal storage unit) is prepared. At time n, the oldest 12 [msec] previous data in the delay processing signal storage unit (not shown) is passed to the reference signal storage unit 52A as DRout (n), and the output of the receiving side loss insertion unit 12 is output. The signal delay is realized by newly accumulating the reference signal Rout (n) referenced from the point Rout in the delay processing signal storage unit.

However, the first and second echo canceller 30A, if 30B has a means for detecting a transmission time (difference transmission time) D _alpha, transmission time difference correction unit 51A, 51B is detected echo canceller 30A, at 30B that it is preferable to perform the correction processing of the transmission time difference by using a transfer time difference D _alpha. Incidentally, an echo canceller 30A, as a method for detecting the transmission time difference D _alpha in 30B, the adaptive filter 31A, since the filter order of the value of the filter coefficient is largest is considered to transmission time difference D _alpha in 31B, transmitted from the filter order a method for determining the time difference D _alpha.

Here, in the signal level calculation method of the signal level calculation unit for each frequency band in the conventional example described in Patent Document 1, the number of data to be processed at one time (the length of the discrete Fourier transform process) is Nf8, and the flowchart of FIG. The description will be given with reference. The signal level calculation unit for each frequency band stores a plurality of (for example, eight) reference signal data DRout (n), DRout (n-1), ..., DRout (n-7) referenced in time series as reference signals. (Step 1), and a discrete Fourier transform process is performed on the read data (step 2). In this discrete Fourier transform process, the real part and the imaginary part are individually calculated for each of the four frequency bands [f0, f1, f2, f3] (step 3), and the square sum of the real part and the imaginary part is further calculated. Calculate the signal level for each frequency band | F _Rout _f0 (n) |, | F _Rout _f1 (n) |, | F _Rout _f2 (n) |, | F _Rout _f3 (n) | (Step 4). That is, since the discrete Fourier transform is an even function with respect to frequency components, if the length Nf = 8, it is sufficient to calculate half of the four components, so that the sampling frequency of the reference signal is 8 [kHz]. When Nf = 8, four frequency band components are calculated: f0: 0 to 0.5 [kHz], f1: 0.5 to 1.5 [kHz], f2: 1.5 to 2.5 [kHz], f3: 2.5 to 3.5 [kHz] I am doing so. 3, reference signal data DRout (n), DRout (n−1),..., DRout (n−7) matrix and signal level | F _{Rout —} f0 (n) | | F _{Rout —} f1 (n) |, | F _{Rout —} f2 (n) |, | F _{Rout —} f3 (n) | Then, after performing the matrix calculation process of step 2, for each band component of [f0, f1, f2, f3], add the square of the real part component and the imaginary part component (step 3), and By performing the square root processing (step 4), the magnitude of the signal level for each band component [| F _Rout _f0 (n) |, | F _Rout _f1 (n) |, | F _Rout _f2 (n) |, | F _{Rout _f3} (n) |] the asking and outputs it to the envelope section.

  With respect to the square root calculation process in step 4, in an algorithm in which the value gradually approaches a true value every time a loop operation such as a Newton-Rapson method that is generally performed is repeated, 2 inputs × 4 bands = A loop operation must be executed in each of the eight square root operations, and the processing load of the DSP realizing the feedback gain estimation unit is heavy.

  On the other hand, the signal level calculation units 53A to 53D for each frequency band in the present embodiment discriminate the reference signal into a plurality of frequency bands by using a plurality of band pass filters having different frequency bands as pass bands. Here, the band-pass filter can be realized using a second-order IIR filter, and the processing necessary for the second-order IIR filter is only a product-sum operation. It is clear that the DSP processing load can be reduced.

  As described above, in the present embodiment, the signal level calculation units 53A to 53D for each frequency band discriminate the reference signal into a plurality of frequency bands by a plurality of band pass filters, so that the manufacturing cost is increased compared to the conventional example. It is possible to improve the estimation accuracy of the feedback gain while suppressing it.

The state the speaker at the other end of the call terminal and the unit audio component is applied in an acoustic echo path H _AC by speaking almost simultaneously, in the so-called double-talk state, and if the stationary noise signal is present Similarly, the estimation accuracy of the feedback gain is lowered. Therefore, in this embodiment, when the double talk state is detected by the double talk detection function of the first echo canceller 30A, the signal level calculation unit 53A for each frequency band stops the filter processing of the band pass filter and performs the previous filter processing. Substitute the signal level obtained in step 1 above. Therefore, it is possible to prevent a decrease in estimation accuracy caused by estimating the feedback gain in the double talk state as described above.

  Here, when the feedback gain estimator 14 is configured by a DSP, interrupt processing for obtaining each feedback gain estimate value on the acoustic side and the line side is performed at each sampling time when the reference signal is A / D converted. The calculation processing burden is considerably increased. Therefore, in contrast to the interrupt processing generated at every A / D sampling time in the DSP, in this embodiment, the acoustic side feedback gain unit 14A and the line side feedback gain unit 14B are alternately operated to reduce the processing load on the DSP. Reduced. Specifically, the transmission time difference correction unit 51 and the reference signal storage unit 52 are always operated, and the signal level calculation unit 54 for each frequency band is operated / stopped exclusively every time an interrupt occurs.

  By the way, the total loss calculation unit 15 selects the feedback gain estimation value of the highest frequency band among the feedback gain estimation values for each frequency band estimated by the feedback gain estimation unit 14, and selects the selected feedback Calculate the total loss from the gain estimate. At this time, the total loss amount calculation unit 15 preferably calculates a total loss amount by adding a predetermined margin to the selected feedback gain estimation value. In this way, howling can be suppressed even when the estimated value of the feedback gain is smaller than the actual feedback gain.

  Note that the total loss amount calculation unit 15 may calculate the total sum of the total loss amounts for each frequency band for which the estimated value of the feedback gain is obtained. At this time, the insertion loss amount distribution processing unit 16 determines the distribution of the insertion loss amounts of the transmission side loss insertion unit 11 and the reception side loss insertion unit 12 for each frequency band. In this way, it is possible to reduce the amount of loss inserted into the frequency band where the estimated feedback gain value is small.

  Incidentally, since the first echo canceller 30A suppresses acoustic echoes, it is not necessary to operate when the voice switch 10 is in the transmission state. Similarly, since the second echo canceller 30B suppresses line echo, there is no need to operate when the voice switch 10 is in the receiving state. Therefore, the adaptive filter 31A of the first echo canceller 30A and the adaptive filter 31B of the second echo canceller 30B distribute the insertion loss amount determined by the insertion loss amount distribution processing unit 16 (the reception state and transmission state of the voice switch 10). ), It is preferable to adjust the timing for updating the filter coefficient. That is, when the insertion loss amount distribution processing unit 16 increases the distribution of the insertion loss amount of the transmission side loss insertion unit 11 (when the voice switch 10 is in the receiving state), the adaptive filter 31A of the first echo canceller 30A is The filter coefficient is updated, and the adaptive filter 31B of the second echo canceller 30B may not update the filter coefficient. On the other hand, when the insertion loss amount distribution processing unit 16 increases the distribution of the insertion loss amount of the reception side loss insertion unit 12 (when the voice switch 10 is in the transmission state), the adaptive filter 31A of the first echo canceller 30A is The adaptive filter 31B of the second echo canceller 30B may update the filter coefficient without updating the filter coefficient.

  As described above, if the first and second echo cancellers 30A and 30B update the filter coefficient in conjunction with the state of the voice switch 10, the filter coefficient is updated regardless of the state of the voice switch 10. As a result, it is possible to improve the convergence accuracy of the filter coefficients and reduce the processing load of the DSP.

  By the way, the voice switch 10 of the present embodiment is switched to three types of states, that is, a reception state, a transmission state, and a neutral state. The neutral state is a state that is neither a reception state nor a transmission state, and the insertion loss amount distribution processing unit 15 distributes the same loss amount to the transmission side loss insertion unit 11 and the reception side loss insertion unit 12. The state that is. Then, each adaptive filter 31A, 31B of the first and second echo cancellers 30A, 30B starts from the time when the voice switch 10 is switched from the state in which the filter coefficient is updated (the reception state or the transmission state) to the neutral state. It is preferable to continue to update the filter coefficient until a certain period has elapsed. In this way, even when the state of the voice switch 10 is frequently switched, the adaptive filters 31A and 31B update the filter coefficients, and the first and second echo cancellers 30A and 30B appropriately suppress echoes. Can do.

13 Insertion loss control unit
14 Feedback gain estimator
15 Total loss calculator
51A, 51B Transmission time difference correction unit
53A to 53D Signal level calculator for each frequency band

Claims (9)

A microphone and a speaker, a reception side signal path for transmitting a reception signal sent from the other party's telephone terminal to the speaker, and a transmission signal collected by the microphone are transmitted to the other party's telephone terminal A voice switch for switching a call state to reception and transmission by inserting a loss in a signal path on the transmission side, and an echo canceller for suppressing an acoustic echo generated by acoustic coupling of the microphone and the speaker. A transmission side loss insertion unit that inserts a loss into the signal path on the transmission side, a reception side loss insertion unit that inserts a loss into the signal path on the reception side, and the loss insertion units on the transmission side and the reception side. An insertion loss amount control unit for controlling the loss amount inserted from the reception side loss insertion unit through the acoustic echo path from the output point of the reception side loss insertion unit. With estimating the acoustic side feedback gain of the path to return to the input point of the transmitter-side loss insertion portion, from the output point of the transmitting end losses insertion portion to the input point of the reception-side loss insertion portion via a line echo path A feedback gain estimator that estimates the line-side feedback gain of the feedback path, and a total loss amount calculator that calculates the total amount of loss to be inserted into the closed loop based on the estimated values of the feedback gains on the acoustic side and the line side; , The transmission signal and the reception signal are monitored to estimate the call state, and each insertion of the transmission side loss insertion unit and the reception side loss insertion unit according to the estimation result and the calculated value of the total loss amount calculation unit In a loudspeaker apparatus having an insertion loss amount distribution processing unit for determining a loss amount distribution,
The feedback gain estimation unit includes a transmission time difference correction unit that corrects a difference in signal transmission time inherent in each of the feedback paths on the acoustic side and the line side, and a signal corrected by the transmission time difference correction unit as a plurality of frequencies. A signal level calculation unit for each frequency band for obtaining a signal level for each frequency band by discriminating into bands, and an estimation for calculating an estimated value of the feedback gain for each frequency band from the signal level calculated by the signal level calculation unit for each frequency band A value calculation unit, and the signal level calculation unit for each frequency band discriminates the signal into a plurality of frequency bands by a plurality of band pass filters having different frequency bands as pass bands ,
A first echo canceller that suppresses the acoustic echo, and a second echo canceller that suppresses line echo caused by acoustic coupling at the counterpart telephone terminal or wraparound of the transmission signal via the line echo path, The first echo canceller adaptively identifies an impulse response of the acoustic echo path and estimates the acoustic echo from an input signal to the acoustic echo path, and the acoustic echo estimated by the adaptive filter And a subtractor that subtracts from the output signal from the acoustic echo path, and the second echo canceller adaptively identifies the impulse response of the line echo path and outputs the signal from the input signal to the line echo path. An adaptive filter for estimating line echo, and the line echo estimated by the adaptive filter A subtracter for subtracting from the output signal from the path, and the insertion filter determined by the insertion loss distribution processor determines the adaptive filter of the first echo canceller and the adaptive filter of the second echo canceller A loudspeaker apparatus that adjusts a timing for updating a filter coefficient in accordance with the distribution of the loss amount .   2. The loudspeaker apparatus according to claim 1, wherein the feedback gain estimation unit alternately performs the process of estimating the acoustic-side feedback gain and the process of estimating the line-side feedback gain at regular intervals.   When the signal level calculation unit for each frequency band detects a double talk state in which a speech component is applied in the course of the path where the echo canceller estimates the feedback gain, the filter processing of the band pass filter is stopped and 3. The loudspeaker apparatus according to claim 1, wherein the signal level obtained by the filtering process is substituted.   The echo canceller includes a transmission time difference detection unit that detects a difference in signal transmission time inherent in each of the feedback paths on the acoustic side and the line side, and the transmission time difference correction unit is a transmission detected by the transmission time difference detection unit. The loudspeaker apparatus according to any one of claims 1 to 3, wherein the transmission time difference is corrected using a time difference.   The estimated value calculation unit, when estimating the feedback gain for each frequency band from the signal level calculated by the signal level calculation unit for each frequency band, the feedback gain of the frequency band for which the signal level is a predetermined threshold or less The loudspeaker apparatus according to any one of claims 1 to 4, wherein the estimated value is not calculated.   6. The total loss amount calculation unit according to claim 1, wherein the total loss amount calculation unit calculates a total sum of the loss amounts by adding a predetermined margin to the estimated value of the feedback gain estimated by the feedback gain estimation unit. The loudspeaker apparatus according to any one of the preceding claims. The total loss amount calculation unit calculates a total sum of the total loss amounts for each frequency band for which the estimated value of the feedback gain is obtained, and the insertion loss amount distribution processing unit includes the transmission side loss insertion unit and the transmission side loss insertion unit The loudspeaker according to any one of claims 1 to 6, wherein the distribution of each insertion loss amount of the reception side loss insertion unit is determined for each frequency band. The adaptive filter of the first echo canceller updates the filter coefficient when the insertion loss amount distribution processing unit distributes a relatively high insertion loss amount to the transmission side loss insertion unit, and The adaptive filter of the 2-echo canceller updates the filter coefficient when the insertion loss amount distribution processing unit distributes a relatively high insertion loss amount to the reception-side loss insertion unit. Item 8. The voice communication device according to any one of Items 1 to 7. When each of the adaptive filters of the first echo canceller and the second echo canceller is updating the filter coefficient, the insertion loss amount distribution processing unit is configured to transmit each loss insertion unit on the transmitting side and the receiving side. the insertion loss equally from changing the allocation to state until a predetermined period has elapsed according to any one of claims 1 to 8, characterized in that to continue updating the filter coefficient with respect to Loudspeaker .

Images