JP4371002B2 - Loudspeaker - Google Patents

Description

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

  In this type of loudspeaker, an uncomfortable echo (acoustic echo or acoustic echo) is caused by an acoustic return path formed by acoustic coupling of a microphone and a speaker, or a line-side return path formed between the other party's call terminal. Line echo) may be heard, or howling occurs at the frequency when a closed loop in which the loop gain in an arbitrary frequency component exceeds 1 is formed in the communication system by the feedback path or the like. In some cases, an echo canceller and a voice switch are provided to prevent the generation of unpleasant echoes and howling as described above.

  The voice switch constantly estimates the call state (sending state, receiving state) and inserts a loss into the signal path on the transmitting side and the receiving side with an appropriate distribution based on the estimation result. The echo canceller adaptively identifies the impulse response of the feedback path and estimates the pseudo echo component of the feedback path from the input signal to the feedback path, and the pseudo echo component estimated by the adaptive filter as the feedback path. And a subtracter that subtracts from the output signal from the. Here, since the echo canceller's adaptive filter normally requires several seconds of learning time to identify the impulse response of the feedback path, the echo suppression effect of the echo canceller cannot be expected sufficiently in the few seconds immediately after the start of the call. There is a possibility that an unpleasant echo or howling may occur because a closed loop is formed in the call system.

  Therefore, the present applicant has already proposed a loudspeaker device that can suppress unpleasant echoes and howling immediately after the start of a call (see Patent Document 1).

  In this conventional example, when the echo canceller immediately after the start of the call has not converged, the voice switch operates in a fixed mode that fixes the total amount of loss (total loss amount) inserted into the signal path to a sufficiently large initial value. Thus, unpleasant echoes and howling are suppressed, and in a state where the echo canceller has sufficiently converged, the voice switch operates in an update mode in which the total loss amount is updated at any time, thereby realizing two-way simultaneous calls.

  The conventional voice switch disclosed in Patent Document 1 includes a transmitter attenuator that inserts a loss into the signal path on the transmitter side, a receiver attenuator that inserts a loss into the signal path on the receiver side, and a transmitter. And an insertion loss amount control unit for controlling a loss amount to be inserted from each attenuator on the talk side and the reception side. Further, the insertion loss amount control unit is configured to return the acoustic side feedback gain of the path that returns from the output point of the receiving side attenuator to the input point of the transmitting side attenuator via the acoustic echo path (hereinafter referred to as “acoustic side feedback path”). And the line-side feedback gain of the path that returns from the output point of the transmitting-side attenuator to the input point of the receiving-side attenuator via the line echo path (hereinafter referred to as “line-side feedback path”) Based on the estimated values of the feedback gains on the acoustic side and the line side, the sum of the loss amounts to be inserted into the closed loop (sum of the insertion loss amount of the transmitting side attenuator and the insertion loss amount of the receiving side attenuator) is calculated. The total loss amount calculation unit and the call state are estimated by monitoring the transmission signal and the reception signal, and each of the transmission side attenuator and the reception side attenuator according to the estimation result and the calculated value of the total loss amount calculation unit An insertion loss amount distribution processing unit that determines the distribution of the insertion loss amount.

  The total loss calculation unit estimates the time average power of the input signal of the transmitting side attenuator in a short time using a rectifier / smoothing device or low-pass filter, and also uses a rectifier / smoothing device or low-pass filter. The time average power of the output signal of the receiver attenuator in a short time is estimated, and the minimum value of the estimated value of the time average power of the output signal of the receiver attenuator is calculated at the maximum delay time assumed in the acoustic side feedback path. The value obtained by dividing the estimated value of the time average power of the input signal of the transmitting side attenuator by this minimum value is used as the estimated value of the acoustic side feedback gain, and the received signal is received using a rectifier / smoothing device, a low-pass filter, etc. Estimate the time average power of the input signal of the side attenuator in a short time, and also estimate the time average power of the output signal of the transmission side attenuator in a short time using a rectifier smoother or low-pass filter, etc. Side return path The minimum value of the average value of the time average power of the output signal of the transmitting side attenuator is obtained at the maximum delay time assumed, and the estimated value of the time average power of the input signal of the receiving side attenuator is divided by this minimum value. The value is an estimated value of the line-side feedback gain. Then, the total loss amount calculation unit calculates the total loss amount necessary to obtain a desired gain margin from the estimated values of the acoustic side feedback gain and the line side feedback gain, and outputs the value to the insertion loss amount distribution processing unit. To do.

The insertion loss distribution processing unit monitors the input / output signals of the transmitting side attenuator and the input / output signals of the receiving side attenuator, and makes a call based on information such as the relationship between the power levels of these signals and the presence / absence of an audio signal. Insertion loss of each attenuator so as to determine the state (receiving state, transmitting state, etc.) and distribute the total loss amount to the transmitting side attenuator and the receiving side attenuator in proportion to the determined call state The amount is adjusted.
JP 2002-359580 A

  By the way, since the feedback path generally has frequency dependency, howling occurs at a specific frequency. Therefore, it is necessary to investigate the frequency dependence and use the maximum value (peak value) for estimating the gain margin and the like. However, in the feedback gain estimation process of the above conventional example, the estimated value is obtained by the ratio of the time average power without considering the frequency dependence. Therefore, the higher the frequency dependence of the gain of the feedback path, the lower the estimation accuracy. It was. Therefore, in order to prevent the howling that occurs when a closed loop in which the round trip gain of the feedback path exceeds 1 is formed in the communication system, the gain margin must be designed on the safe side. However, even if the number does not exceed one time, there is a case where a two-way simultaneous call cannot be realized and a one-way call is made.

  The present invention has been made in view of the above circumstances, and its purpose is to improve the estimation accuracy of the feedback gain, and to control the total loss amount so that the closed loop loop gain does not exceed one time more optimal than before. An object of the present invention is to provide a loudspeaker device that can further increase the opportunities for realizing two-way simultaneous calls.

In order to achieve the above object, the invention of claim 1 provides 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 collected by the microphone. Suppresses acoustic echo caused by acoustic coupling between the microphone and speaker, and a voice switch that switches the call state between receiving and transmitting by inserting loss into the transmitting signal path that transmits the signal and sends it to the other party's telephone terminal An echo canceller, and the voice switch includes a transmission side loss insertion unit that inserts loss into the signal path on the transmission side, a reception side loss insertion unit that inserts loss into the signal path on the reception side, and a transmission An insertion loss amount control unit for controlling the loss amount inserted from each loss insertion unit on the reception side and the reception side. With estimating the acoustic side feedback gain of the path to return to the input point of the transmitter-side loss insertion portion through the input point of the receiving-side insertion loss portion from the output point of the transmitting end losses insertion portion via a line echo path A feedback gain estimator for estimating the line-side feedback gain of the path to be fed back, and a total loss calculator for calculating the sum of the losses to be inserted into the closed loop based on the estimated values of the feedback gains on the acoustic side and the line side Then, the transmission state and the reception signal are monitored to estimate the call state, and the insertion loss amounts 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 calculation unit An update mode for adaptively updating the total loss amount calculation unit by calculating the sum of the loss amounts to be inserted into the closed loop based on the estimated value of each feedback gain, Fixed to fix the total loss to a predetermined initial value The period after the echo canceller has sufficiently converged while operating in the fixed mode during the period from the start of the call with the other party's call terminal until the echo canceller has sufficiently converged In the loudspeaker device operating in the update mode, the feedback gain estimation unit includes a transmission time difference correction unit that corrects a difference in signal transmission time inherent in each feedback path on the acoustic side and the line side, and a transmission time difference correction unit. A signal level calculation unit for each frequency band that obtains a signal level for each frequency band by performing Fourier transform processing on the corrected signal, and a feedback gain for each frequency band from the signal level calculated by the signal level calculation unit for each frequency band A maximum feedback gain selection unit that estimates and selects the maximum feedback gain among the feedback gains for each frequency band, and a maximum feedback gain selected by the maximum feedback gain selection unit. And a smoothing filter unit in which the smoothed value is smoothed in time series and the smoothed value is variable by parameter setting.

  According to a second aspect of the present invention, in the first aspect of the invention, the feedback gain estimation unit includes a reference signal storage unit that stores data obtained by sampling the reference signal of the transmission signal and the reception signal, and a signal level calculation unit for each frequency band. Is characterized in that it performs discrete Fourier transform processing by reading data stored in the reference signal storage unit.

  According to a third aspect of the present invention, in the first or second aspect of the present invention, an envelope unit that obtains a time series average for each frequency band with respect to the signal level for each frequency band calculated by the signal level calculation unit for each frequency band is provided with a feedback gain. The maximum feedback gain selection unit is provided in the estimation unit, and selects the maximum time series average among the time series averages of the signal levels obtained by the envelope unit as a feedback gain.

According to a fourth aspect of the present invention, in the invention according to any one of the first to third aspects, the feedback gain estimation unit alternately estimates the acoustic side feedback gain and the line side feedback gain at regular intervals, and is estimated at different times. The feedback gain estimation is performed by combining the acoustic side feedback gain and the line side feedback gain.

According to a fifth aspect of the present invention, in any one of the first to fourth aspects, the signal level calculation unit for each frequency band performs a process of obtaining a square root when calculating the signal level. A list having a one-to-one correspondence with the square root is provided, and the square root is calculated by referring to the list.

According to a sixth aspect of the present invention, in the second aspect of the invention, the signal level calculation unit for each frequency band may be such that all the data read from the reference signal storage unit for performing the discrete Fourier transform processing is equal to or less than a predetermined threshold value. For example, the discrete Fourier transform process is stopped and the signal level obtained by the previous discrete Fourier transform process is used instead.

The invention according to claim 7 is the invention according to any one of claims 1 to 6 , wherein the signal level calculation unit for each frequency band is in a double talk state in which an audio component is applied in the middle of a path where the echo canceller estimates the feedback gain. If detected, the discrete Fourier transform process is stopped and the signal level obtained by the previous discrete Fourier transform process is used instead.

The invention according to claim 8 is the invention according to any one of claims 1 to 7 , wherein the maximum feedback gain selection unit provides the maximum feedback with respect to the closed loop feedback gain represented by the sum of the acoustic side feedback gain and the line side feedback gain. When selecting a signal level of a frequency band having a gain, a denominator for each frequency band of a feedback gain calculation expression expressed by a determinant is divided.

According to the first aspect of the present invention, since the largest feedback gain obtained for each frequency band is selected as the estimated value of the feedback gain, the feedback gain is estimated from the time average power of the reference signal. The feedback gain estimation accuracy is improved compared to the example, and as a result, it is possible to perform the total loss amount control so that the closed loop loop gain does not exceed one time more optimally than in the past. It is possible to further increase opportunities to make calls, and even in environments where it is difficult to call normally due to ambient noise such as outdoor road noise and indoor TV noise,
A more comfortable call is achieved, and the selected maximum feedback gain is smoothed over time and flat.
Since the smoothed value is variable depending on the parameter setting, howling due to incorrect estimation occurs.
There is an effect that it is possible to provide a loudspeaker device that can be easily designed to prevent life .

  According to the invention of claim 2, in addition to the effect of the invention of claim 1, the execution waiting time of the discrete Fourier transform process until the data subject to the discrete Fourier transform process is accumulated in the reference signal storage unit Thus, continuous discrete Fourier transform processing synchronized with the sampling period of the reference signal can be performed, so that the feedback gain estimation accuracy can be further improved.

  According to the invention of claim 3, in addition to the effect of the invention of claim 1 or 2, since the time series average of the signal level for each frequency band is obtained by the envelope section, the amount of calculation in the discrete Fourier transform processing is suppressed. Therefore, even when the number of data simultaneously performing the discrete Fourier transform process is small, it is possible to prevent a reduction in feedback gain estimation accuracy due to noise application.

According to the invention of claim 4 , in addition to the effects of the invention of any one of claims 1 to 3 , the amount of calculation is reduced compared to the case of simultaneously estimating the acoustic side feedback gain and the line side feedback gain, This has the effect of being able to be implemented on low-spec computers.

According to the invention of claim 5 , in addition to the effect of any one of claims 1 to 4 , the amount of calculation is reduced as compared with the case where the square root calculation is performed using a general Newton-Raphson method or the like. Therefore, there is an effect that it is possible to implement with a low-spec computer.

According to the sixth aspect of the invention, in addition to the effect of the second aspect of the invention, it is possible to prevent a reduction in feedback gain estimation accuracy caused by estimating a feedback gain by a ratio of stationary noise signals. There is.

According to the invention of claim 7 , in addition to the effect of any one of the inventions of claims 1 to 6 , the feedback gain is estimated in a double talk state in which a voice component is applied in the middle of the path for estimating the feedback gain. There is an effect that it is possible to prevent the feedback gain estimation accuracy from being lowered.

According to the invention of claim 8 can be reduced in addition to the effect of any one of the claims 1 to 7, the computation amount required for the maximum value selection from can avoid the division operation by Tsubun There is an effect.

  DESCRIPTION OF THE PREFERRED EMBODIMENTS Embodiments in which the present invention is applied to a voice call device (interphone terminal) will be described in detail below with reference to the drawings. However, the present invention is not limited to this, and may be any loudspeaker device installed in the entire living space.

  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, and a line output amplifier G2 that amplifies a transmission signal to a line (2-line transmission path). The line input amplifier G3, the speaker amplifier G4, the transmission volume adjustment amplifier G5, the reception volume adjustment amplifier G6, the voice switch 10, and the first and second echo cancellers 30A and 30B that amplify the reception signal from the line. Composed.

The first echo canceller 30A includes a well-known structure composed of the adaptive filter 31A and a subtractor 32A, the impulse response of the feedback path (acoustic echo path) H _AC formed by the acoustic coupling between the speaker 2 microphone 1 The acoustic echo is adaptively identified by the adaptive filter 31A and subtracted from the output signal of the microphone amplifier G1 by the subtractor 32A from the pseudo echo component (acoustic echo) estimated from the reference signal (input signal to the speaker amplifier G4). It is to suppress. The second echo canceller 30B also has a conventionally well-known configuration including an adaptive filter 31B and a subtractor 32B. The reflection due to impedance mismatch between the 2-wire-to-wire conversion circuit 3 and the transmission path and the counterpart The impulse response of the feedback path (line echo path) H _LIN formed by the acoustic coupling between the speaker and the microphone in the telephone terminal (for example, intercom system door phone slave unit) is adaptively identified by the adaptive filter 31B and referred to By subtracting the pseudo echo component (line echo) estimated from the signal (input signal to the line output amplifier G2, that is, transmission signal) from the received signal by the subtractor 32B, the line echo is suppressed.

The voice switch 10 includes a transmission side attenuator 11 for inserting a loss in the signal path on the transmission side, a reception side attenuator 12 for inserting a loss in the signal path on the reception side, and attenuations on the transmission side and the reception side. And an insertion loss amount control unit 13 for controlling the loss amount inserted from the devices 11 and 12. The insertion loss amount control unit 13, the path to be fed back from the output point Rout of the receiving side attenuator 12 to the input point Tin of the transmitter-side attenuator 11 via the acoustic echo path H _AC (hereinafter referred to as "acoustic side feedback path" ) On the acoustic side feedback gain α and a path for returning from the output point Tout of the transmitting side attenuator 11 to the input point Rin of the receiving side attenuator 12 via the line echo path H _LIN (hereinafter referred to as “line side”). The feedback gain estimation unit 14 for estimating the line-side feedback gain β of the feedback path) and the loss 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 A total loss amount calculation unit 15 for calculating the total amount (sum of the insertion loss amount of the transmission side attenuator and the insertion loss amount of the reception side attenuator), and the call state is estimated by monitoring the transmission signal and the reception signal In accordance with the estimation result and the calculated value of the total loss calculating unit 15, the transmitting side attenuator 1 1 and an insertion loss amount distribution processing unit 16 for determining the distribution of each insertion loss amount of the receiving side attenuator 12. The first and second echo cancellers 30A and 30B and the voice switch 10 in this embodiment control the DSP (Digital Signal Processor) hardware with software (program) for the echo canceller and voice switch. It is realized by. Accordingly, 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 amount calculation unit 15 calculates a total loss amount Lt necessary to obtain a desired gain margin MG from the estimated values α ′ and β ′ of the acoustic side feedback gain α and the line side feedback gain β, and the value Lt Is output to the insertion loss amount distribution processing unit 16.

  The total loss amount distribution processing unit 16 monitors the input / output signals of the transmitting side attenuator 11 and the input / output signals of the receiving side attenuator 12, and information such as the magnitude relationship between the power levels of these signals and the presence / absence of an audio signal. The communication state (the reception state, the transmission state, etc.) is determined from the transmission state, and the total loss Lt is distributed to the transmission side attenuator 11 and the reception side attenuator 12 at a rate corresponding to the determined call state. The amount of insertion loss of the attenuators 11 and 12 is adjusted.

  By the way, the total loss calculation unit 15 in the present embodiment calculates the amount of loss to be inserted into the closed loop 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. There are two operation modes: an update mode in which the sum is calculated and adaptively updated, and a fixed mode in which the total loss is fixed to a predetermined initial value. During the period until the echo cancellers 30A and 30B sufficiently converge, the fixed mode operates, and during the period after the first and second echo cancellers 30A and 30B sufficiently converge, the update mode operates. 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 ε ( For example, it is considered that the first and second echo cancellers 30A and 30B have sufficiently converged when the estimated values α ′ and β ′ are smaller than the estimated values α ′ and β ′ at the start of the call by 10 to 15 dB. The operation mode is switched to the update mode in which the total loss amount is adaptively updated based on the estimated values α ′ and β ′ based on the estimated values α ′ and β ′. 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 first and second echo cancellers 30A and 30B immediately after the start of the call are not sufficiently converged, the total loss amount calculation unit 15 operating in the fixed mode sets the value sufficiently large. Since the initial total loss amount is inserted into the closed loop, it is possible to suppress the generation of unpleasant echoes (acoustic echoes and line echoes) and howling, and realize a stable half-duplex call. Also, when the first and second echo cancellers 30A and 30B have sufficiently converged after the start of the call, the operation mode of the total loss 15 is switched from the fixed mode to the update mode and inserted into the closed loop. Since the total loss amount to be reduced to a value sufficiently lower than the initial value, a two-way simultaneous call can be realized.

  Here, a specific operation of the total loss amount calculation unit 14 in the update mode will be described with reference to a flowchart of FIG.

  The total loss calculation unit 15 estimates the acoustic side feedback gain α and the estimated value α ′ (n of the line side feedback gain β that are executed by the feedback gain estimation unit 14 at a predetermined sampling period from the time when the fixed mode is changed to the update mode. ), Β ′ (n) product α ′ (n) · β ′ (n) is read (step 1), and from this product α ′ (n) · β ′ (n) and gain margin MG, the gain of the closed loop is read. The total loss desired value Lr (n) required for maintaining the margin at MG [dB] is calculated by the following equation (step 2).

Lr (n) = 20 log | α ′ (n) · β ′ (n) | + MG [dB]
Note that α ′ (n), β ′ (n), and Lr (n) indicate an estimated value of feedback gain and a desired total loss amount calculated by the nth sampling from the update mode transition point, respectively. Further, the total loss amount calculation unit 15 calculates the n-th total loss amount desired value Lr (n) calculated from the above formula and the previous (n−1) th total loss amount value Lt (n−1), that is, the previous time When the desired total loss amount Lr (n) calculated this time is larger than the total loss amount determined by the processing and actually inserted, a slight increase amount Δi [] is added to the previous total loss amount Lt (n−1). dB] is the current total loss amount Lt (n) = Lt (n−1) + Δi (steps 3 and 4), and this time is calculated for the previous total loss amount Lt (n−1). When the total loss desired value Lr (n) is small, a value obtained by subtracting a slight decrease Δd [dB] from the previous total loss Lt (n−1) is set to the current total loss Lt (n) = Lt ( n−1) −Δd (steps 5 and 6).

  In this way, by suppressing the increase / decrease in the total loss amount by the total loss amount calculation unit 15 to a minute value of Δi or Δd, the first or second echo canceller 30 </ b> A immediately after the start of a call with the other party's call terminal is performed. , 30B actively updates the coefficient toward convergence, so that even when the acoustic feedback gain α and the line side feedback gain β are drastically changed, a sense of incongruity can be eliminated.

Next, the feedback gain estimation unit 14 that is the gist of the present invention will be described with reference to FIGS. 1 and 4 to 9. 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. Feedback gain estimation unit 14, the path to be fed back from the output point Rout of the receiving side attenuator 12 to the input point Tin of the transmitter-side attenuator 11 via the acoustic echo path H _AC (hereinafter referred to as "acoustic side feedback path") The acoustic feedback gain estimation unit 14A for estimating the acoustic feedback gain α of the receiver, and the path returning from the output point Tout of the transmitter attenuator 11 to the input point Rin of the receiver attenuator 12 via the line echo path H _LIN It comprises a line-side feedback gain estimator 14B that estimates line-side feedback gain β (hereinafter referred to as “line-side feedback path”).

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

The acoustic-side feedback gain estimator 14 </ b> A receives a reception-side block that processes a reference signal (reception signal) referenced from the output point Rout of the reception-side attenuator 12, and a reference signal referenced from the input point Tin of the transmission-side attenuator 11. And a transmission side block for processing (transmission signal). A reference signal storage unit 52C for storing a reference signal referred to in the transmitting side block, and a signal for each frequency band for obtaining a signal level for each frequency band by reading the reference signal from the reference signal storage unit 52C and performing a discrete Fourier transform process. A level calculation unit 53C and an envelope unit 54C for obtaining a time series average for each frequency band with respect to the signal level for each frequency band calculated by the signal level calculation unit 53C for each frequency band are included. The reference after the transmission time difference correction unit 51A of the acoustic side feedback path including the acoustical side echo path H _AC corrects the difference in signal transmission time with the unique, the difference in signal transmission time has been corrected on the receiving side block A reference signal storage unit 52A that stores a signal (received signal), a signal level calculation unit 53A for each frequency band that obtains a signal level for each frequency band by reading the reference signal from the reference signal storage unit 52A and performing discrete Fourier transform processing; And an envelope unit 54A that obtains a time series average for each frequency band with respect to the signal level for each frequency band calculated by the signal level calculation unit 53A for each frequency band. Further, the acoustic feedback gain estimation unit 14A includes a maximum feedback gain selection unit 55 and a smoothing filter unit 56A. The maximum feedback gain selection unit 55 estimates the feedback gain for each frequency band from the output signals of the envelope units 54A and 54C of the reception side block and the transmission side block, and the maximum feedback among the feedback gains for each frequency band. A process of selecting a gain is performed. The smoothing filter unit 56A performs processing for smoothing the maximum feedback gain selected by the maximum feedback gain selection unit 55 in time series, and varies the smoothed value by parameter setting as will be described later. .

On the other hand, the line-side feedback gain estimator 14B uses a transmission side block for processing a reference signal (transmission signal) referenced from the output point Tout of the transmission side attenuator 11 and an input point Rin of the reception side attenuator 12. A receiving side block for processing the referenced reference signal (receiving signal). The receiver block stores a reference signal storage unit 52D that stores a reference signal that is referred to, and a signal level for each frequency band that is obtained by reading the reference signal from the reference signal storage unit 52D and performing a discrete Fourier transform process. A calculation unit 53D and an envelope unit 54D for obtaining a time series average for each frequency band with respect to the signal level for each frequency band calculated by the signal level calculation unit 53D for each frequency band are included. The transmission side block includes a transmission time difference correction unit 51B that corrects a difference in signal transmission time inherent in the line-side feedback path including the line-side echo path H _LIN , and a signal after the difference in signal transmission time is corrected. A reference signal storage unit 52B for storing a reference signal (transmission signal), and a signal level calculation unit for each frequency band for obtaining a signal level for each frequency band by reading the reference signal from the reference signal storage unit 52B and performing a discrete Fourier transform process 53B and an envelope unit 54B that obtains a time-series average for each frequency band with respect to the signal level for each frequency band calculated by the signal level calculation unit 53B for each frequency band. Further, the line-side feedback gain estimation unit 14B includes a maximum feedback gain selection unit 55 and a smoothing filter unit 56B. The maximum feedback gain selection unit 55 is shared with the acoustic side feedback gain estimation unit 14A, and estimates the feedback gain for each frequency band from the output signals of the envelope units 54B and 54D of the transmitting side block and the receiving side block. Then, a process of selecting the maximum feedback gain among the feedback gains for each frequency band is performed. Further, the smoothing filter unit 56B performs processing for smoothing the maximum feedback gain selected by the maximum feedback gain selection unit 55 in time series, and varies the smoothed value by parameter setting as will be described later. . The units included in the voice switch 10 are realized by controlling the 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.

Receiving signals output from the output point Rout of the receiving side attenuator 12 is to reach to the input point Tin of the transmitter-side attenuator 11 via the acoustic side feedback path including an acoustic echo path H _AC of the system-specific Transmission time is required. Therefore, the transmission time difference correction unit 51A is set in advance by, for example, measuring the time for a single pulse generated from the output point Rout of the receiving side attenuator 12 to reach the input point Tin of the transmitting side attenuator 11. only transmission time D _alpha of have been the system, so as to delay the reference signal from the output point Rout of the receiving side attenuator 12. For example, when the sampling frequency of the reference signal is 8 [kHz] and the measured delay time is 12 [msec], the delay processing signal storage unit for 8 × 12 = 96 data (separate from the reference signal storage units 52A to 52D). First (First In First Out) type signal storage unit) is prepared, and at time n, the oldest 12 [msec] previous data in the delay processing signal storage unit (not shown) is set as DRout (n). The signal delay is realized by passing the signal to the reference signal storage unit 52A and newly accumulating the reference signal Rout (n) referenced from the output point Rout of the reception side attenuator 12 in the delay processing signal storage unit.

  By the way, the signal level calculation units 53A to 53D for each frequency band collect the reference signal data DRout (n), DRout (n-1),..., DRout (n-7) of a certain number Nf (for example, Nf = 8). If the discrete Fourier transform process is performed by reading from the reference signal storage unit 52A (see FIG. 4A), the discrete Fourier transform process can be performed only once in Nf times of the interrupt process generated every sampling time. Therefore, there is a possibility that the estimation accuracy of the feedback gain is lowered. For this purpose, in this embodiment, the reference signal storage units 52A to 52D are first-in first-out (FIFO) types, and the data in the reference signal storage units 52A to 52D is obtained every time an interrupt for sampling occurs as shown in FIG. Shift one by one (replacement into a shift block), and replace reference signal data DRout (n), DRout (n-1), ..., DRout (n-7) from the reference signal storage units 52A to 52D. By reading into the signal level calculation units 53A to 53D for each band and performing the discrete Fourier transform process, the delay is prevented and the estimation accuracy of the feedback gain is improved (see FIG. 6).

Next, a signal level calculation method in the signal level calculation unit 53A for each frequency band will be described with the number of data to be processed at one time (the length of the discrete Fourier transform process) Nf being 8, with reference to the flowchart of FIG. The signal level calculation unit 53A for each frequency band receives a plurality (8 in this embodiment) of reference signal data DRout (n), DRout (n-1),..., DRout (n-7) referenced in time series. Reading from the reference signal storage unit 52A (step 1), and discrete Fourier transform processing 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 | 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. 5, 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 operation processing in 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 further add By performing the square root calculation process (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) |] is output to the envelope section 54A seeking.

  With respect to the square root calculation processing in step 4, an algorithm in which the value gradually approaches a true value every time a loop operation such as a Newton-Rapson method, which is generally performed, is repeated. Even if a method of alternately executing update processing of feedback path estimation gains α ′ and β ′, which will be described later, is performed with length Nf = 8, loop calculation can be performed in each of two inputs × 4 bands = 8 square root operations. In other words, the processing load of the DSP realizing the feedback gain estimation unit 14 is increased. Therefore, in the present embodiment, a list in which the results before and after the square root operation are associated with each other in advance is stored in a nonvolatile storage area, and the square root value is obtained by referring to the list, thereby calculating the DSP. The load is reduced.

Specifically, when obtaining the square root of X (0 ≦ X <1) Y = √X (0 ≦ Y <1), after decomposing it into X = a × 2-b, the even-oddness of b is considered. Y = √a × 2-n (b = 2n, n is a positive integer)
Y = √ (0.5 × a) × 2-n (b = 2n + 1, n is a positive integer)
If the case is divided, an n-bit shift operation may be executed for the part b, and for the remaining part of the variable a, a list in which the variable a is associated with the square root √a is stored in the storage area in advance. If the square root √a is obtained by referring to the list, the desired square root Y can be obtained. However, a general square root calculation algorithm such as Newton-Raphson method may be used if the DSP has sufficient processing capability.

  Here, in a situation where no call is made, only stationary noise components are included in the reference signal data DRout (n), DRout (n-1),..., DRout (n-7). If the feedback gain is estimated by the ratio, the estimation accuracy is lowered. Therefore, in the signal level calculation unit 53A for each frequency band, all signal levels of the reference signal data DRout (n), DRout (n-1), ..., DRout (n-7) read from the reference signal storage unit 52A are predetermined. The discrete Fourier transform process is stopped and the signal level obtained by the previous discrete Fourier transform process is substituted, and the substituted signal level is output to the envelope unit 54A. As described above, the estimation accuracy caused by estimating the feedback gain by the ratio of the stationary noise signal is prevented.

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 discrete Fourier transform process and the previous discrete Fourier transform. The signal level obtained by the processing is substituted, and the substituted signal level is output to the envelope unit 54A, so that the estimation accuracy due to the estimation of the feedback gain in the double talk state is prevented as described above. Yes.

By the way, in the present embodiment, the length Nf of the discrete Fourier transform process is set to a value (Nf = 8) that is considerably smaller than a general value (256 to 1024) in consideration of reducing the burden of calculation processing by the DSP. For this reason, the time length in which the length Nf of the discrete Fourier transform processing is expressed in time units becomes as short as 8/8 [kHz] = 1 [msec] at the sampling frequency (= 8 [kH]). Since it is easily affected by a decrease in estimation accuracy of the feedback gain due to the application, the signal level for each frequency band [| F _Rout _f0 (n) calculated by the signal level calculation unit for each frequency band 53A to improve this. |, | F _Rout _f1 (n) |, | F _Rout _f2 (n) |, | F _Rout _f3 (n) |] is an envelope unit that calculates a time series average for each frequency band. 54A (see Fig. 6) This envelope calculation is performed for signal levels [| F _Rout _f0 (n) |, | F _Rout _f1 (n) |, | F _Rout _f2 (n) |, | F _Rout _f3 (n) |] for each frequency band. The time series average is obtained by performing the calculation represented by the following formula. However, the envelope coefficient ρ in the following equation is the average of the actual measurement values [| F _Rout _f0 (n) |, | F _Rout _f1 (n) |, | F _Rout _f2 (n) |, | F _Rout _f3 (n) |] It is a numerical value that determines the weighting with the values [env_RO_f0 (n), env_RO_f1 (n), env_RO_f2 (n), env_RO_f3 (n)].

env_RO_f0 (n) = (1-ρ) env_RO_f0 (n-1) + ρ | F _Rout _f0 (n) |
env_RO_f1 (n) = (1-ρ) env_RO_f1 (n-1) + ρ | F _Rout _f1 (n) |
env_RO_f2 (n) = (1−ρ) env_RO_f2 (n−1) + ρ | F _Rout _f2 (n) |
env_RO_f3 (n) = (1-ρ) env_RO_f3 (n-1) + ρ | F _Rout _f3 (n) |
For example, if the reference signal is a sine wave signal of 1 [kHz], an 8-point Fourier transform process (time length 1 [msec]) with a discrete Fourier transform length Nf of 8 as shown in FIG. And the frequency spectrum value after the conversion coincides with the 16-point Fourier transform process (time length 2 [msec]) in which the length Nf of the discrete Fourier transform is 16 as shown in FIG. 5) and there is no difference in accuracy between the two. On the other hand, when a noise component is applied to the sine wave signal, the 8-point Fourier transform process (see FIG. 8A) with a shorter time length is more noise component than the 16-point Fourier transform process (see FIG. 8B). Since the contribution of is large, the decrease in accuracy becomes large. However, when the envelope process is used in the 8-point Fourier transform process, the periodicity of past data can be used from the data of the reference signal storage unit 52A processed as shown in FIG. Even in the worst case immediately before exiting from the reference signal storage unit 52A, it can be seen that the accuracy is improved as compared with the 16-point Fourier transform process in which the time length is doubled (however, in FIG. 8C, the envelope coefficient ρ is set to ρ = 1/16).

  Next, the operation of the maximum feedback gain selection unit 55 in the case where the length of the discrete Fourier transform process is Nf = 8 will be described. The maximum feedback gain selection unit 55 sets each frequency band component at time n calculated by the envelope unit 54A from the transmission time difference correction unit 51A using the output point Rout of the reception side attenuator 12 as a reference point in the acoustic side feedback gain estimation unit 14A. The output value [env_RO_f0 (n), env_RO_f1 (n), env_RO_f2 (n), env_RO_f3 (n)] of the envelope unit 54A of the transmitter is obtained from the reference signal storage unit 52C using the input point Tin of the transmission side attenuator 11 as a reference point. By dividing each frequency band component by the output value [env_TI_f0 (n), env_TI_f1 (n), env_TI_f2 (n), env_TI_f3 (n)] of the envelope section 54C for each frequency band component calculated by the envelope section 54C ( The acoustic feedback gain estimation value [env_α'f0 (n), env_α'f1 (n), env_α'f2 (n), env_α'f3 (n)]] is obtained (see FIG. 9A). reference).

env_TI_f0 (n) / env_RO_f0 (n) ≡env_α'f0 (n)
env_TI_f1 (n) / env_RO_f1 (n) ≡env_α'f1 (n)
env_TI_f2 (n) / env_RO_f2 (n) ≡env_α'f2 (n)
env_TI_f3 (n) / env_RO_f3 (n) ≡env_α'f3 (n)
The output of the envelope unit 54B for each frequency band component at time n calculated by the envelope unit 54B from the transmission time difference correction unit 51B with the output point Tout of the transmission side attenuator 11 as a reference point in the line side feedback gain estimation unit 14B. The frequency [env_TO_f0 (n), env_TO_f1 (n), env_TO_f2 (n), env_TO_f3 (n)] calculated from the reference signal storage unit 52D by the envelope unit 54D using the input point Rin of the receiver attenuator 12 as a reference point By dividing each frequency band component by the output value [env_RI_f0 (n), env_RI_f1 (n), env_RI_f2 (n), env_RI_f3 (n)] of the envelope unit 54D for each band component (see the following formula), the line side feedback Gain estimation values [env_β′f0 (n), env_β′f1 (n), env_β′f2 (n), env_β′f3 (n)] are obtained (see FIG. 9B).

env_RI_f0 (n) / env_TO_f0 (n) ≡env_β'f0 (n)
env_RI_f1 (n) / env_TO_f1 (n) ≡env_β'f1 (n)
env_RI_f2 (n) / env_TO_f2 (n) ≡env_β'f2 (n)
env_RI_f3 (n) / env_TO_f3 (n) ≡env_β'f3 (n)
Furthermore, the acoustic side gain estimates [env_α'f0 (n), env_α'f1 (n), env_α'f2 (n), env_α'f3 (n)] and the line side gain estimates [env_β'f0 (n), env_β 'f1 (n), env_β'f2 (n), env_β'f3 (n)] product for each frequency band component
env_α'f0 (n) x env_β'f0 (n)
env_α'f1 (n) x env_β'f1 (n)
env_α'f2 (n) × env_β'f2 (n)
env_α'f3 (n) x env_β'f3 (n)
Of these, the frequency band component having the maximum value is selected. This corresponds to identifying the frequency component with the least margin among the frequency dependence of the howling margin of the closed loop loop gain in the interphone system, and it is possible to accurately estimate the frequency component with the least margin and its level. Become. For example, as shown in FIG. 9C, if the frequency band f1 is a band having the maximum value of the feedback gain estimation value, the maximum feedback is obtained with the components env_α′f1 (n) and env_β′f1 (n) of the frequency band f1. The maximum feedback gain [α1 (n), β1 (n)] at time n selected by the gain selector 55 is expressed as α1 (n) = env_α'f1 (n)
β1 (n) = env_β'f1 (n)
This is output to the acoustic side smoothing filter unit 56A and the line side smoothing filter unit 56B, respectively. However, when the maximum value of the product for each frequency band component is obtained, the amount of calculation can be reduced by dividing the product. In other words, by dividing the product for each frequency band component, the maximum value can be obtained without executing division, and the division processing that the DSP is not good at can be avoided to reduce the burden of arithmetic processing. Can do. Of course, division processing may be performed if the DSP has sufficient processing capacity.

  Finally, the operation of the smoothing filter units 56A and 56B will be described with reference to FIG. However, since the smoothing processing of the acoustic side smoothing filter unit 56A and the line side smoothing filter unit 56B is common, the acoustic side smoothing filter unit 56A will be described below.

  First, before starting the smoothing process, variables UsCounter and DsCounter are initialized to preset parameters Us_Init and Ds_Init, respectively. The acoustic feedback gain estimated value α ′ (n) after the averaging process is also initialized to an initial value (= 0) (step 1). Subsequently, the maximum feedback gain α1 (n) output from the maximum feedback gain selection unit 55 is read, and the acoustic feedback gain estimated value α1 (n) at time n is averaged after the averaging process at time n−1. If the value of α1 (n) is larger (α1 (n)> α '(n-1)) compared to the side feedback gain estimate α' (n-1), the variable UsCounter is decremented and the variable DsCounter is initialized On the other hand, if α1 (n) is smaller (α1 (n) <α '(n-1)), the variable DsCounter is decremented and the variable UsCounter is initialized. If both are equal (α1 (n) = α ′ (n)) and the process is continued (step 2). Then, the variable UsCounter is 0, that is, in step 2, the acoustic feedback gain estimate α1 (n) at time n is the acoustic feedback gain estimate α ′ (n−) after the averaging process at time n−1. 1) If the case of greater than is continued continuously, α ′ (n−1) plus parameter ε is processed as α ′ (n). If variable DsCounter is 0, that is, in step 2 If the acoustic side feedback gain estimated value α1 (n) at time n is smaller than the acoustic side feedback gain estimated value α ′ (n−1) after the averaging process at time n−1, continuously, A process obtained by subtracting the parameter ε from '(n-1) is set as α' (n). Otherwise, α ′ (n) = α ′ (n−1) is held and the previous value is held (step 3).

  Thus, in the smoothing filter unit 56A, the acoustic feedback gain estimation value α ′ (n) after the averaging process can be increased or decreased according to the magnitude relationship between the parameters Us_Init and Ds_Init representing the initial values of the variables UsCounter and DsCounter. If Us_Init> Ds_Init, the acoustic feedback gain estimation value α ′ (n) can be guided in a direction in which the value becomes smaller than the time series average calculated by the envelope unit 54A. If <Ds_Init, the acoustic feedback gain estimation value α ′ (n) can be guided in a direction in which the value becomes larger than the time-series average calculated by the envelope unit 54A. That is, if the feedback gain is estimated to be an excessively large value by time series averaging by the envelope unit 54A, there is a possibility that howling due to erroneous estimation may occur. Therefore, in order to prevent such howling from occurring. It is desirable to guide the acoustic feedback gain estimated value α ′ (n) in a direction in which the value becomes smaller than the time series average calculated by the envelope unit 54A. For example, in this embodiment, (Us_Init, Ds_Init) = (64, 1).

  Although the processing flow using the output point Rout of the reception side attenuator 12 of the acoustic side feedback gain estimation unit 14A as a reference point has been mainly described, the input point Tin of the transmission side attenuator 11 and the transmission side attenuator 11 are described. The same processing is performed except for the presence or absence of the transmission time difference correction unit 51 with respect to the processing flow using the output point Tout and the input point Rin of the reception side attenuator 12 as reference points.

By the way, when the feedback gain estimator 14 is configured by a DSP, performing an interrupt process for obtaining each feedback gain estimated value on the acoustic side and the line side for each sampling time when the A / D conversion of the reference signal is performed by the DSP. The burden of arithmetic processing becomes considerably large. Therefore, in contrast to the interrupt processing generated at the A / D sampling time by the DSP, in this embodiment, the acoustic side feedback gain unit 14A and the line side feedback gain unit 14B are alternately operated to increase the processing load on the DSP. Reduced. Specifically, the transmission time difference correction unit 51, the reference signal storage unit 52, the maximum feedback gain selection unit 55, and the smoothing filter unit 56 are always operated, and the signal level calculation unit 54 and the envelope unit 55 for each frequency band are operated. Operate / stop exclusively every time an interrupt occurs. At that time, for example, the maximum feedback gain selection unit 55 updates the acoustic feedback gain estimation value α ′ (2k + 1) at time 2k + 1, and the acoustic feedback gain estimation value β ′ (2k) at time 2k. When updating, at time 2k + 1, the acoustic feedback gain estimation values [env_α'f0 (2k + 1), env_α'f1 (2k + 1), env_α'f2 (2k + 1), env_α'f3 (2k +1)] and the line-side feedback gain estimate [env_β'f0 (2k), env_β'f1 (2k), env_β'f2 (2k), env_β'f3 (2k)] for each frequency band component
env_α'f0 (2k + 1) × env_β'f0 (2k)
env_α'f1 (2k + 1) × env_β'f1 (2k)
env_α'f2 (2k + 1) × env_β'f2 (2k)
env_α'f3 (2k + 1) × env_β'f3 (2k)
At time 2k, the acoustic feedback gain estimate [env_α'f0 (2k-1), env_α'f1 (2k-1), env_α'f2 (2k-1), env_α'f3 (2k-1 )] And line-side feedback gain estimate [env_β'f0 (2k), env_β'f1 (2k), env_β'f2 (2k), env_β'f3 (2k)] for each frequency band component
env_α'f0 (2k-1) × env_β'f0 (2k)
env_α'f1 (2k-1) × env_β'f1 (2k)
env_α'f2 (2k-1) × env_β'f2 (2k)
env_α'f3 (2k-1) × env_β'f3 (2k)
The frequency band component having the maximum value is selected from The maximum feedback gain [α1 (n), β1 (n)] at time n = 2k + 1 or n = 2k is obtained and output to the acoustic side smoothing filter unit 56A and the line side smoothing filter unit 56B, respectively. . However, if the DSP has sufficient processing capacity, the feedback gain on the acoustic side and the line side may be updated simultaneously in one interrupt process.

It is a block diagram which shows the insertion loss amount control part in one Embodiment of this invention. It is a block diagram which shows the same as the above. It is a flowchart for operation | movement description of a voice switch in the same as the above. It is operation | movement explanatory drawing of the signal level calculation part classified by frequency band in the same as the above. It is a flowchart for operation | movement description of the signal level calculation part classified by frequency band in the same as the above. It is operation | movement explanatory drawing of an envelope part in the same as the above. It is operation | movement explanatory drawing of an envelope part in the same as the above. It is operation | movement explanatory drawing of an envelope part in the same as the above. It is operation | movement explanatory drawing of the maximum feedback gain selection part in the same as the above. It is operation | movement explanatory drawing of the smoothing filter part in the same as the above.

Explanation of symbols

13 Insertion loss amount control unit 14A Acoustic side feedback gain estimation unit 14B Line side feedback gain estimation unit 51A, 51B Transmission time difference correction unit 52A to 52D Reference signal storage unit 53A to 53D Frequency band specific signal level calculation unit 54A to 54D Envelope unit 55 Maximum feedback gain selection unit 56A, 56B Smoothing filter unit

Claims (8)

The microphone and speaker, the receiver side signal path for transmitting the reception signal sent from the other party's telephone terminal to the speaker, and the transmitter side transmitting the transmission signal collected by the microphone to the other party's telephone terminal The voice switch includes a voice switch that switches the call state between receiving and sending by inserting a loss in the signal path, and an echo canceller that suppresses acoustic echo caused by acoustic coupling between the microphone and the speaker. The loss amount inserted from the loss insertion unit on the transmission side, the loss insertion unit on the reception side that inserts the loss on the signal path on the reception side, and the loss insertion unit on the transmission side and the reception side. comprising an insertion loss amount control unit for controlling the insertion loss amount control unit is fed back from the output point of the receiving-side loss insertion portion to the input point of the transmitter-side loss insertion portion via an acoustic echo path Estimate the feedback gain on the acoustic side of the road and estimate the feedback gain on the line side of the path that returns from the output point of the loss insertion part on the transmission side to the input point on the insertion loss part on the reception side via the line echo path A total loss amount calculation unit 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, and monitors the transmission signal and the reception signal to determine the call state And an insertion loss amount distribution processing unit that determines the distribution of each insertion loss amount 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, The total loss amount calculation unit calculates the sum of the loss amounts to be inserted into the closed loop based on the estimated value of each feedback gain and adaptively updates the update mode, and the fixed mode for fixing the total loss amount to a predetermined initial value. There are two modes of operation, and the other party's call The feedback gain estimation is performed in a loudspeaker that operates in the fixed mode during the period from the start of the call with the terminal until the echo canceller sufficiently converges and in the update mode during the period after the echo canceller sufficiently converges. The transmission time difference correction unit that corrects the difference in signal transmission time inherent to each feedback path on the acoustic side and the line side, and the frequency band by performing Fourier transform processing on the signal corrected by the transmission time difference correction unit Estimate the feedback gain for each frequency band from the signal level calculated by the signal level calculation unit for each frequency band and the signal level calculation unit for each frequency band to obtain another signal level, and the largest of the feedback gains for each frequency band The maximum feedback gain selection unit that selects the feedback gain, and the maximum feedback gain selected by the maximum feedback gain selection unit are smoothed in time series and the smoothed value is A loudspeaker apparatus comprising a smoothing filter unit that is variable by parameter setting. The feedback gain estimation unit includes a reference signal storage unit that stores data obtained by sampling the reference signal of the transmission signal and the reception signal, and the signal level calculation unit for each frequency band reads and stores the data stored in the reference signal storage unit. The loudspeaker apparatus according to claim 1, wherein Fourier transform processing is performed. The feedback gain estimator includes an envelope unit that obtains a time series average for each frequency band with respect to the signal level for each frequency band calculated by the signal level calculation unit for each frequency band, and the maximum feedback gain selection unit is determined by the envelope unit. The loudspeaker apparatus according to claim 2, wherein the maximum time-series average among the time-series averages of the signal levels is selected as a feedback gain. The feedback gain estimation unit alternately estimates the acoustic side feedback gain and the line side feedback gain at regular intervals,
Feedback gain by combining acoustic feedback gain and line feedback gain estimated at different times
The loudspeaker apparatus according to claim 1, 2 or 3, wherein estimation is performed . The signal level calculation unit for each frequency band obtains the square root when calculating the signal level.
A list having a one-to-one correspondence between a variable and the square root of the variable,
5. The loudspeaker apparatus according to claim 1 , wherein a square root is calculated by referring to the list . The signal level calculation unit for each frequency band performs reference signal recording in order to perform discrete Fourier transform processing.
If all the data read from the memory is below the specified threshold, the discrete Fourier transform process is
3. The loudspeaker apparatus according to claim 2 , wherein the signal level obtained by the previous discrete Fourier transform process is stopped and substituted . The signal level calculation unit for each frequency band is in the middle of the path where the echo canceller estimates the feedback gain.
When a double-talk state is detected in which audio components are applied, the discrete Fourier transform process is stopped.
The loudspeaker according to any one of claims 1 to 6, wherein the signal level obtained by the previous discrete Fourier transform process is substituted . The maximum feedback gain selector is a closed-loop feedback expressed as the sum of the acoustic feedback gain and the line feedback gain.
Expressed in fractional expression when selecting the signal level of the frequency band with the maximum feedback gain relative to the gain
The loudspeaker apparatus according to any one of claims 1 to 7, wherein a denominator for each frequency band of the feedback gain calculation formula is divided .

Images