TWI399968B - Speakerphone apparatus - Google Patents

Description

Speakerphone

The present invention relates to a speakerphone apparatus for a home or office.

A speakerphone device is known in the art, which realizes hands-free communication without a handset device when communicating: using a speaker to amplify a voice signal and output it to a farther distance from the communication terminal The calling party; and uses a microphone to collect the voice of the calling party and transmit the voice to his/her counterpart's communication terminal. Here, the speaker amplifies the voice signal transmitted from the other party's communication terminal, and the microphone collects the voice from the calling party.

In this speakerphone device, an acoustic feedback path can be established by acoustic coupling between the microphone and the speaker, via which an unpleasant echo (acoustic echo) can be heard. Furthermore, if a closed loop is formed under a frequency component via the feedback path in the communication system, and the gain of the loop exceeds 1, then howling may occur at the frequency. . Therefore, an echo canceller and a voice switch are provided to prevent unpleasant echoes and howling.

The voice switcher continuously evaluates the communication state (transmission state, reception state), and provides the loss into the transmission signal path and the reception signal path by appropriately allocating the loss based on the evaluation result. The echo canceller includes: an adaptive filter for identifying a pulse response in the feedback path and estimating a suspected echo component in the feedback path based on a signal input to the adaptive filter; A subtractor for subtracting the estimated echo component from the signal input to the subtractor via the feedback path. In general, the adaptive filter requires several seconds of learning time in order to identify the impulse response of the feedback path. Therefore, the echo cannot be effectively suppressed within a few seconds immediately after the start of communication. However, since a closed loop has been established in the communication system, an unpleasant echo or howling may occur therein.

In this regard, a speakerphone device capable of suppressing the occurrence of an uncomfortable echo or howling after the start of communication has been proposed (for example, see Japanese Patent Application Laid-Open No. 2002-359580).

In this conventional example, although the echo canceller does not converge immediately after the start of communication, the voice switcher operates in a fixed mode in which the total amount of loss (total loss) provided in the signal path is fixed sufficiently large The initial value, thereby suppressing an unpleasant echo or howling. On the other hand, after the echo canceller is fully converged, the voice switcher operates in an update mode in which the total amount of loss is updated as needed, and thus full duplex communication can be achieved.

Meanwhile, if digital communication (such as internet protocol (IP) communication, etc.) is performed between a plurality of speakerphone devices, it may be handled by a packet or used to suppress jitter in voice data transmission ( Jitter) The communication buffer processing is delayed. In general, it is known that when a delay occurs in the transmission of voice data between a plurality of speakerphone devices, as the delay time becomes longer, the speaker's tolerance to echo is deteriorated during communication, and thus even a small The echo can also make the speaker feel uncomfortable.

This unpleasant echo can be prevented by setting an initial value greater than the initial value of the conventional analog communication in the fixed mode. However, the degradation of the echo tolerance may exceed a few tenths of a mile, which cannot be covered by the cancellation of the echo canceller in the update mode, and therefore can only be communicated in half-duplex communication instead of full duplex. Communication in communication. In addition, when the ambient noise level is high, the following problem may occur: the voice switcher continues to be turned on only one side of the transmission side and the reception side, and thus the voice from the one side is not transmitted to the other One side.

On the other hand, when digital communication (for example, IP communication or the like) is performed between the speakerphone device and the communication device using the earpiece, the speakerphone device can be enabled by an echo suppressor equipped therein. Residual echo attenuation occurring in the transmission signal path, thereby achieving full duplex communication. However, if there is a high level of noise in the signal transmitted from the speakerphone device to the counterpart communication device, the echo suppressor not only attenuates the residual echo but also attenuates the noise for a short time. This short decay to noise is delayed and perceived by the other party via the handset. Therefore, the other party using the handset may feel an uncomfortable disconnection feeling during communication.

Further, if the other party uses the earpiece, the above-mentioned feeling of disconnection can be prevented by setting the attenuation amount of the echo suppressor to be smaller than when the opponent uses the speakerphone device. However, in this case there is a disadvantage of hearing residual echoes through the earpiece.

In view of the above, the present invention provides a speakerphone device capable of achieving full-duplex communication even when a transmission delay occurs in a voice transmitted to the other party, and can prevent disconnection through the earpiece even when the other party uses the earpiece a feeling of.

According to an aspect of the present invention, a speakerphone device includes: a microphone and a speaker; and a voice switcher for switching between transmission and reception in a communication state by: a signal path on the receiving side and Providing a loss in the transmission signal path, the reception signal from the communication terminal of the other party is transmitted to the speaker via the reception side signal path, and the transmission signal collected by the microphone is transmitted to the communication terminal of the other party via the transmission signal path; A canceller for suppressing acoustic echo caused by acoustic coupling between the microphone and the speaker; and a double talk detector for detecting double talk.

The voice switcher includes a transmission side loss insertion unit for providing a loss in a transmission signal path; a reception side loss insertion unit for providing a loss in a reception side signal path; and for controlling in the transmission side and the reception side The insertion loss amount control unit of the corresponding loss amount is provided.

Further, the insertion loss amount control unit includes: a total loss amount calculator for estimating that an output terminal of the path from the receiving side loss insertion unit is fed back to an input of the transmission side loss insertion unit via an acoustic echo path An acoustic feedback gain of the terminal, and calculating a sum of the amount of loss provided to the feedback path based on the estimate of the acoustic feedback gain; and an insertion loss amount allocation processor for monitoring the transmission signal and the received signal The communication state is estimated, and based on the estimated communication state and the sum calculated by the total loss amount calculator, the respective insertion loss amounts assigned to the transmission side loss insertion unit and the reception side loss insertion unit are determined.

Further, the total loss amount calculator has an update mode and a fixed mode in which the total loss amount is updated in an adaptive manner by calculating a total loss amount provided in the feedback path based on a corresponding estimate of the feedback gain, And in the fixed mode, the total loss is fixed to the initial value. Therefore, when the echo canceller does not sufficiently converge after the start of communication, the total loss amount calculator operates in the fixed mode, and after the echo canceller completely converges, the total loss amount calculator operates in the update mode.

The speakerphone device further includes an echo suppressor for attenuating the residual echo by combining a voice switcher and/or a double talk detector to provide a specific amount of attenuation in the transmission signal path.

With the above configuration, even if a transmission delay occurs in the voice transmitted between the plurality of communication devices, the echo suppressor effectively attenuates the residual echo due to the transmission delay in the transmission signal path, thereby achieving full-duplex communication. .

If the insertion loss amount distribution processor estimates that the voice switcher is in the receiving state, the echo suppressor may insert a specific amount of attenuation in the transmission signal path, otherwise no specific attenuation amount is inserted.

Therefore, the inflection occurring in the speech can be suppressed, and only the uncomfortable echo is attenuated after the communication. Therefore, a pleasant communication can be achieved.

When the double talk detector is detecting double talk, the echo suppressor may not insert the amount of attenuation in the transmission signal path, otherwise the attenuation amount is inserted.

Therefore, the inflection occurring in the speech due to the erroneous attenuation of the speech to be transmitted can be suppressed, and only the uncomfortable echo is attenuated after the communication. Therefore, a pleasant communication can be achieved.

If the double talk detector does not detect the double talk and the insertion loss allocation processor estimates that the voice switch is in the receiving state, the echo suppressor may insert the specific attenuation amount in the transmission signal path.

By doing so, it is possible to prevent the echo suppressor from erroneously providing no amount of attenuation and occurrence of an uncomfortable residual echo; and thus accurately attenuating only unpleasant echoes during communication.

Preferably, when the echo canceller does not sufficiently converge after the start of communication, the echo suppressor inserts the specific attenuation amount in the transmission signal path only when the insertion loss amount distribution processor estimates that the voice switcher is in the receiving state. .

In general, when the echo canceller does not converge sufficiently, the double talk detector cannot detect double talk. Therefore, if the insertion loss allocation processor estimates that the voice switcher is in the receiving state, the echo suppressor inserts the specific attenuation amount in the transmission signal path regardless of the detection result from the double-talk detector, and thus starts the communication. An unpleasant echo is attenuated when the post-echo canceller does not converge sufficiently.

When the echo suppressor is inserting the specific attenuation amount in the transmission signal path, if the insertion loss amount distribution processor estimates that the voice switcher is not in the receiving state or the double-talking detector detects the double-talking, the echo suppressor may Immediately stop inserting the attenuation into the transmit signal path.

By doing so, it is possible to prevent the inflection in the speech due to the erroneous attenuation of the transmission signal by the echo suppressor.

The echo canceller can further include: an adaptive filter for adaptively identifying the characteristics of the echo path; a subtractor for subtracting the output of the adaptive filter from the near-end signal; and a double-talk detector Used to detect double-talk. In addition, when the double talk detector detects a double talk, the echo canceller may not update the coefficients of the adaptive filter.

With this configuration, the double-talk detector provided in the echo canceller can be used to achieve configuration simplification and cost reduction.

Corresponding processing in the echo canceller, voice switcher, and echo suppressor is preferably performed in the order described in the transmission signal path.

With this configuration, since the echo suppressor is placed behind the voice switcher in the transmission signal path, even if the echo suppressor erroneously attenuates the transmission signal, the inflection in the voice switcher can be prevented, thereby achieving a pleasant Communication environment.

The echo suppressor can adjust the amount of attenuation based on the acoustic feedback gain estimated by the total loss calculator.

With this configuration, even if the speech from the near-end speaker is erroneously attenuated by the echo suppressor, the occurrence of the voice inflection can be suppressed.

The echo suppressor can adjust the amount of attenuation based on the signal level average of the received signal.

With this configuration, even if the speech from the near-end speaker is erroneously attenuated by the echo suppressor, the occurrence of the voice inflection can be suppressed.

When the signal level of the received signal is lower than the threshold, the echo suppressor can fix the amount of attenuation inserted in the transmission signal path to a standard value, and when the signal level of the received signal is a threshold or higher, the echo is returned. The suppressor adjusts the amount of attenuation from a range above the upper limit of the standard value to the standard value.

With this configuration, the occurrence of a voice inflection can be suppressed.

The echo canceller may further comprise: a signal level average (SLA) operation unit for calculating a signal level average of the received signal; and an adaptive filter for recognizing the echo path in an adaptive manner. a feature; and a subtractor for subtracting the output of the adaptive filter from the near-end signal. Further, the echo suppressor can adjust the amount of attenuation based on the signal level average calculated by the signal level average operation unit.

This makes it possible to achieve a simplification of the configuration and a reduction in cost.

The echo suppressor can insert a specific amount of attenuation in the transmission signal path only when the double talk detector does not detect the double talk and the insertion loss allocation processor estimates that the voice switch is in the receiving state. The loss allocation processor estimates that the voice switcher is neither in the receiving state nor in the transmitting state, or when the insertion loss allocation processor estimates that the voice switcher is in the receiving state and the double-talk detector detects the double-talking The echo suppressor monotonically reduces the amount of attenuation inserted in the transmitted signal path.

With this configuration, even if a transmission delay occurs in the voice transmitted to the other party's communication device, the echo suppressor can effectively attenuate the residual echo due to the transmission delay in the transmission signal path, thereby achieving full-duplex communication. Further, when a transmission signal is not transmitted from the speakerphone device to the communication device using the earpiece, the echo suppressor monotonically reduces the amount of attenuation inserted in the transmission signal path. This makes it possible to prevent the speaker at the communication device using the earpiece from feeling uncomfortable because the surrounding noise is not temporarily attenuated.

When the communication state estimated by the insertion loss amount distribution processor becomes the transmission state when the attenuation amount is monotonically decreasing, the echo suppressor can immediately reduce the attenuation amount to zero.

By doing so, it is possible to prevent the voice (transmission voice) emitted from the near-end speaker from being erroneously attenuated by the echo suppressor and preventing the inflection in the voice heard by the counterpart speaker.

When the amount of attenuation inserted in the transmission signal path is monotonically decreasing, the echo suppressor is when the double-talking detector does not detect the double-talk and the communication state estimated by the insertion loss amount distribution processor is the receiving state. A specific amount of attenuation can be inserted again in the transmission signal path.

With this configuration, the echo suppressor can be prevented from erroneously providing no amount of attenuation and an unpleasant residual echo occurs; and only the unpleasant echo is accurately attenuated during communication.

The echo suppressor can be constant when the communication state estimated by the insertion loss amount allocation processor is neither the reception state nor the transmission state, or when the communication state is the reception state and the double-talk detector detects the double-talk conversation. The reduction ratio reduces the amount of attenuation inserted in the transmission signal path.

In this case, for example, when compared with the stepwise mode reduction, the period (transition period) required for the attenuation amount to decrease to zero can be relatively shortened.

When the double talk detector does not detect the double talk and the communication state estimated by the insertion loss allocation processor is the receiving state, the echo suppressor may insert a specific attenuation amount in the transmission signal path, and the echo suppressor estimates the The level of ambient noise superimposed in the transmitted signal, and the amount of attenuation inserted in the transmitted signal path is reduced when the ambient noise level becomes high based on the estimated ambient noise level.

With this configuration, even if a transmission delay occurs in the voice transmitted to the other party's communication device, the echo suppressor can effectively attenuate the residual echo due to the transmission delay in the transmission signal path, thereby achieving full-duplex communication. Further, since the echo suppressor reduces the amount of attenuation inserted in the transmission signal path as the noise level becomes higher, the instantaneous attenuation of noise can be suppressed, thus making it difficult to feel the feeling of disconnection. In this case, if the echo suppressor reduces the amount of attenuation, the attenuation effect of the residual echo is lowered. However, in actual use, the residual echo is covered by ambient noise and does not cause trouble in communication.

The echo suppressor can compare the back volume estimated according to the received signal level with the attenuation amount according to the surrounding noise level superimposed in the transmission signal, and insert one of the smaller values of the volume and the attenuation amount in the transmission signal path. .

This makes it possible to sufficiently attenuate the residual echo while preventing the feeling of disconnection from being sensed.

When the ambient noise level exceeds a certain level, the echo suppressor can adjust the amount of attenuation to a constant value based on the noise level.

With this configuration, residual echoes are prevented from being attenuated when the ambient noise level is very high.

Hereinafter, embodiments of the present invention will be described in more detail with reference to the accompanying drawings, which form a part.

First embodiment

Referring to FIG. 1, the speakerphone device of the present embodiment includes: a microphone 1; a speaker 2; a voice switcher 10, and a voice switcher 10 by receiving a signal path (shown by a broken line A in FIG. 1) and transmitting a signal path ( The insertion loss in FIG. 1 is shown by the broken line B and changes between transmission and reception in the communication state, and the reception signal from the communication terminal of the other party is transmitted to the speaker 2 via the reception signal path, and the transmission is collected by the microphone 1. The signal is transmitted to the communication terminal of the other party via the transmission signal path; the echo canceller 20 suppresses the acoustic echo caused by the acoustic coupling between the microphone 1 and the speaker 2; and the microphone amplifier G1 for amplifying the output of the microphone 1 Signal (transmission signal); speaker amplifier G2 for amplifying the input signal (received signal) of the microphone 2.

Further, the speakerphone apparatus includes: an amplifier G3 for adjusting a transmission amount inserted between the echo canceller 20 and the voice switcher 10 in the transmission signal path; and an echo suppressor 30 by transmitting a signal A specific amount of attenuation is inserted into the path to attenuate the residual echo.

The echo canceller 20 has a well-known adaptive filter 21 and a subtractor 22. The echo canceller 20 recognizes the impulse response of the feedback path (acoustic echo path) Hac formed by the acoustic coupling between the speaker 2 and the microphone 1 in an adaptive manner by the adaptive filter 21, and by using the subtractor 22 The estimated acoustic echo is subtracted from the output signal of the microphone amplifier G1 to suppress the echo component (acoustic echo) estimated from the reference signal (the input signal to the speaker amplifier G2).

The voice switcher 10 includes: a transmission signal attenuator 11 for inserting a loss in the transmission signal path; a reception signal attenuator 12 for inserting a loss in the reception signal path; and an insertion loss amount controller 13 for controlling The amount of loss of the corresponding transmission signal attenuator 11 and the received signal attenuator 12 is inserted.

The insertion loss amount controller 13 has a total loss amount calculator 14 that estimates a path fed back from the output point Rout of the reception signal attenuator 12 to the input point Tin of the transmission signal attenuator 11 via the acoustic echo path Hac (hereinafter referred to as acoustics) The acoustic feedback gain α of the feedback path C is calculated and the sum of the losses to be inserted (the sum of the insertion loss amounts of the transmission signal attenuator 11 and the reception signal attenuator 12) is calculated based on the estimate α' of the acoustic feedback gain α.

Further, the insertion loss amount controller 13 includes an insertion loss amount (ILA) allocation processor 15 which estimates the communication state by monitoring the transmission signal and the reception signal, and based on the estimated communication state and the calculated total loss amount. The allocation of the respective insertion loss amounts for the transmission signal attenuator 11 and the reception signal attenuator 12 is determined.

In addition, the echo canceller 20 and the voice can be implemented by controlling a hardware (for example, a digital signal processor (DSP)) and a software (for example, a program for an echo canceller and a voice switch). Switcher 10. Therefore, the signals (transmission signals and reception signals) input to and output from the voice switcher 10 and the echo canceller 20 are performed by an A/D converter (analog to digital converter) not shown at a specific sampling period. Sampling and quantification.

The total loss amount calculator 14 estimates the corresponding average power of the signal input to the transmission signal attenuator 11 and the signal output from the reception signal attenuator 12 in a short period of time by using a rectifier, a low-pass filter or the like. The total loss calculator 14 also obtains an average power estimate of the output signal of the received signal attenuator 12 at the maximum delay time expected in the acoustic echo path Hac. Subsequently, the total loss amount calculator 14 sets the estimate α' of the acoustic feedback gain to a value obtained by estimating the average power of the output signal of the received signal attenuator 12 at the maximum delay time. Divided by the average power estimate of the input signal of the transmission signal attenuator 11.

Further, the total loss amount calculator 14 calculates the total loss amount Lt by using the estimate α' of the acoustic feedback gain α and the required gain margin MG, and outputs the total loss amount Lt to the insertion loss amount. The processor 15 is assigned.

The insertion loss amount distribution processor 15 monitors the corresponding input/output signals of the transmission signal attenuator 11 and the reception signal attenuator 12, and determines the communication based on the magnitude relationship in the power levels of the input/output signals and the presence of the speech signal. Status (receive status, transmission status). Further, the insertion loss amount distribution processor 15 adjusts the insertion loss amounts of the respective attenuators 11 and 12 so that the total loss amount Lt can be proportionally distributed to the transmission signal attenuator 11 and the reception signal attenuator based on the determined communication state. 12.

In the present embodiment, the total loss amount calculator 14 has an update mode and a fixed mode in which the sum of the loss amounts to be inserted is calculated and adaptively updated based on the estimate α' of the acoustic feedback gain α, In fixed mode, the total loss is fixed at an initial value. The total loss amount calculator 14 operates in the fixed mode before the echo canceller 20 is fully converged after the start of communication, and then, after the echo canceller is fully converged, the total loss amount calculator 14 operates in the update mode.

That is, once the value of the estimate α' of the acoustic feedback gain α is maintained within a certain period of time after the start of communication (for example, hundreds of milliseconds or longer), it is kept below the threshold ε (for example, an estimate α from the start of communication). The small loss amount calculator 14 considers that the echo canceller 20 has sufficiently converged and switches to operate in the update mode thereafter, and until now it has been operating in the fixed mode. Further, in the fixed mode, the initial value is set to a value sufficiently larger than the total loss amount updated in the update mode as occasion demands.

Therefore, although the echo canceller 20 does not sufficiently converge immediately after the start of communication, the total loss amount calculator 14 operates in the fixed mode, and sets the total loss amount to an initial value having a sufficiently large value, and thus can be suppressed Uncomfortable echo (acoustic echo) or howling occurs and achieves stable half-duplex communication.

Further, when the echo canceller 20 completely converges after a reasonable time elapses from the communication start time, the operation mode of the total loss amount calculator 14 is changed from the fixed mode to the update mode, and the total loss amount to be inserted is reduced to be sufficiently low. The value of the initial value, and thus full duplex communication can be achieved.

Hereinafter, the specific operation of the total loss amount calculator 14 in the update mode will be explained with reference to FIG.

The total loss amount calculator 14 performs an estimation process of the acoustic feedback gain α in a specific sampling period after the time point from the fixed mode to the update mode, and calculates an estimate α' of the acoustic feedback gain α (step 1). Based on the estimated α' and the gain margin MG, the required total loss amount Lr(n) is calculated by using the following formula (step 2) so that the gain margin MG [dB] of the closed loop can be maintained:

Lr(n)=20 log|α'(n)|+MG[dB]

Here, α'(n) and Lr(n) are the estimated α' of the acoustic feedback gain α and the total amount of loss required, respectively, which is the nth sampling from the time point when the fixed mode is changed to the update mode. computational. Further, when the required total loss amount Lr(n) is larger than the previous total loss amount Lt(n-1) determined at the (n-1)th sampling and actually inserted, the total loss amount calculator 14 will be slightly increased. The amount Δi [dB] is added to the previous total loss amount Lt(n-1) (steps 3 and 4). That is, the total loss amount calculator 14 sets the total loss amount Lt(n) to Lt(n-1) + Δi.

Further, when the total amount of loss Lr(n) is smaller than the previous total loss amount Lt(n-1), the total loss amount calculator 14 subtracts a small amount of decrease Δd from the previous total loss amount Lt(n-1) [ dB], and the total loss is set as the result of the subtraction (steps 5 and 6). That is, Lt(n)=Lt(n-1)-Δd.

As described above, since the total loss amount calculator 14 increases and decreases the total loss amount by the small values Δi and Δd, respectively, even if the acoustic feedback gain α is actively updating the coefficient (filter coefficient) due to the echo canceller 20 Convergence and wide fluctuations (just as after the communication terminal with the other party has just started communication) can also eliminate the uncomfortable feeling during the listening process.

As described above with respect to the conventional example, if digital communication such as IP communication is performed between a plurality of speakerphone devices, delay may occur in voice data transmission due to packet processing or communication buffer processing for suppressing jitter. In general, it is known that when a delay occurs in the transmission of voice data between a plurality of speakerphone devices, as the amount of delay time becomes longer, the speaker's tolerance to echo is deteriorated during communication, and thus the speaker can I noticed even a small echo and felt uncomfortable. This unpleasant echo can be prevented by setting an initial value larger than the initial value of the conventional example for the fixed mode.

However, the amount of degradation of the echo tolerance may sometimes reach a few tenths of a mile, which cannot be completely covered by the cancellation of the echo canceller in the update mode. In this case, communication is performed at half duplex instead of full duplex. Further, when the noise level is high, a problem may occur in which the voice switcher continues to be turned on only one side of the transmission side and the reception side, and therefore, the voice from one side is not transmitted to the other side.

In this regard, in the present embodiment, the echo suppressor 30 is additionally provided in the transmission signal path behind the echo canceller 20 and the voice switcher 10. The echo suppressor 30 attenuates the acoustic echo (hereinafter referred to as "residual echo") which should have been suppressed by the echo canceller 20. The echo suppressor 30 needs to effectively attenuate the residual echo while transmitting the speech signal (transmission signal) without attenuating.

The echo suppressor 30 incorporates an attenuation amount into the transmission signal path in conjunction with the voice switcher 10, and more specifically, as illustrated in the flow chart of FIG. That is, the echo suppressor 30 continues to monitor (step 11) the state of the voice switcher 10 (the insertion loss amount allocation processor 15 estimates the communication state (reception state or transmission state)). When the voice switcher 10 is in the receiving state, the echo suppressor 30 considers that no voice signal is transmitted to the transmission signal path, and attenuates it by multiplying the signal input from the voice switcher 10 by a specific attenuation coefficient. And output it (step 12).

Further, when the voice switcher 10 is in the transmission state, the echo suppressor 30 considers that there is no residual echo to be cancelled or any voice signal (transmission signal) to be transmitted, and the signal input from the voice switcher 10 is pressed without attenuation. Output as it is (step 13).

According to the first embodiment of the present invention, even if a transmission delay occurs in a voice transmitted between a plurality of communication devices, the echo suppressor effectively attenuates the residual echo due to the transmission delay in the transmission signal path, thereby achieving Full duplex communication.

At this point, when the echo suppressor 30 inserts an attenuation amount in the transmission signal path when the voice switcher 10 is in the transmission state instead of the reception state, for example, the speaker from the near end side (in this embodiment, the use) The speech emitted by the speaker of the speakerphone device is attenuated, which is undesirable and may be heard by the speaker at the other party's communication device like a voice inflection, where the volume of the voice becomes lower or higher. . However, in the present embodiment, when the voice switcher 10 is in the receiving state, the echo suppressor 30 inserts an attenuation amount in the transmission signal path, and when the voice switcher 10 is in the transmission state, the echo suppressor 30 is not present. The amount of attenuation is inserted in the transmission signal path. Therefore, only the uncomfortable echo (residual echo) can be attenuated without a speech inflection.

Furthermore, if the echo suppressor 30 is placed in front of the voice switcher 10 in the transmission signal path and erroneously attenuates the transmission signal, a voice call may occur in the voice switcher 10 because the transmission signal is erroneously attenuated by the echo suppressor 30. fold. However, in the present embodiment, since the echo suppressor 30 is disposed behind the echo canceller 20 and the voice switcher 10 in the transmission signal path, and the corresponding processing is performed in the stated order, even if the echo suppressor 30 is in error The transmission signal is attenuated, and the occurrence of a voice inflection can also be suppressed in the voice switcher 10. Therefore, a pleasant environment can be provided for communication.

Meanwhile, the present inventors have proposed a speakerphone apparatus in which an attenuator for insertion loss is provided behind a voice switcher in a transmission signal path (see Japanese Patent Application Laid-Open No. 2003-324371). However, in this document, the attenuator inserts a certain amount of loss in the transmission signal path only when the temporary power estimate of the transmitted signal remains below a threshold for a certain period of time. Therefore, if the echo has a large volume above the threshold, the attenuator does not insert a loss, and an unpleasant echo may be transmitted to the other party.

Furthermore, even if the power level of the speech from the speaker at the proximal end is below the threshold, the attenuator may accidentally attenuate the speech signal due to insertion loss. However, in the speakerphone apparatus according to the embodiment, the echo suppressor 30 operates to incorporate an amount of attenuation in conjunction with the voice switcher 10, thereby preventing the above from occurring.

Second embodiment

Fig. 4 is a block diagram schematically showing a speakerphone apparatus according to a second embodiment of the present invention. The speakerphone device of the second embodiment is different from the speakerphone device of the first embodiment in that the echo suppressor 30 is combined with the echo canceller 20 to provide an amount of attenuation in the transmission signal path, and other aspects are the same as the first embodiment. The same is true. Therefore, the same components are denoted by the same reference numerals, and the description thereof will be omitted.

In this embodiment, in addition to the adaptive filter 21 and the subtractor 22, the echo canceller 20 includes a double talk detector 23. The double talk detector 23 detects double talk, that is, the near speaker and the far end speaker, based on whether the signal output from the microphone amplifier G1 includes a signal interference convergence of a certain level of the adaptive filter 21. The state of speaking to each other at substantially the same time. When the double talk detector 23 detects a double talk, the adaptive filter 21 does not update the filter coefficients and maintains the previous value.

Next, the operation of the echo suppressor 30 will be explained with reference to the flowchart shown in FIG. The echo suppressor 30 continues to monitor the double talk detector 23 of the echo suppressor 20 (step 21). When the double talk detector 23 does not detect the double talk, the echo suppressor 30 determines that there is a residual echo in the to-be-removed or a voice signal to be transmitted in the transmission signal path, and by making the input signal and the specific attenuation coefficient Multiply it to attenuate and output it (step 22).

In addition, when the double talk detector 23 detects the double talk, the echo suppressor 30 determines that there is a voice signal to be transmitted, and does not attenuate the input signal by not multiplying the input signal by a specific attenuation coefficient. Ground output (step 23).

With the second embodiment, even if a transmission delay occurs in the voice transmitted between the plurality of communication devices, the echo suppressor can effectively attenuate the residual echo caused by the transmission delay in the transmission signal path, thereby achieving a full double Communication.

In this embodiment, when the double talk detector 23 is detecting a double talk, the echo suppressor 30 does not insert an attenuation amount in the transmission signal path, and when the double talk detector 23 does not detect the double end When talking, the echo suppressor 30 inserts an attenuation amount in the transmission signal path. Therefore, it is possible to prevent the voice from the speaker at the near end side from being erroneously attenuated, and to prevent the inflection in the voice. In addition, since an unpleasant echo can be attenuated during communication, a pleasant environment can be provided for communication.

Furthermore, since the double talk detection component (double talk detector 23) provided in the echo canceller 20 can be used in this embodiment, simplification of the configuration can be achieved (for example, the amount of execution in the DSP) Reduction, lower specifications in the DSP, reduction in ROM capacity for storing programs, etc., and cost reduction.

Third embodiment

Figure 6 is a block diagram showing a speakerphone apparatus in accordance with a third embodiment of the present invention. The basic configuration of this embodiment is the same as that of the first embodiment and the second embodiment, except that the echo suppressor 30 incorporates both the voice switcher 10 and the echo canceller 20 to insert an attenuation amount in the transmission signal path. Therefore, the same components are denoted by the same reference numerals, and the description thereof will be omitted.

The operation of the echo suppressor 30 will be explained hereinafter with reference to FIG. The echo suppressor 30 constantly monitors the state of the voice switcher 10 (step 31), and when the voice switcher 10 is in the receiving state, the echo suppressor 30 determines whether the double talk detector 23 detects a double talk (step 32). When the double talk detector does not detect the double talk, the echo suppressor 30 considers that there is a residual echo to be cancelled or a voice signal to be transmitted, and by multiplying the input signal by a specific attenuation coefficient. It is attenuated and outputted (step 33).

When the voice switcher 10 is not in the receiving state, or when the voice switcher 10 is in the receiving state, when the double talk detector 23 detects the double talk, the echo suppressor 30 determines that there is no residual echo or voice. The signal is to be transmitted, and the input signal is output as it is without attenuation (step 34).

With this embodiment, it is possible to prevent the voice from the speaker at the near end side from being erroneously attenuated and to prevent the inflection in the voice. Furthermore, the echo suppressor 30 can only attenuate unpleasant echoes without unpleasant residual echoes (because the echo suppressor 30 does not accidentally insert an amount of attenuation), thereby providing a pleasing communication for communication. surroundings.

Here, the echo suppressor 30 can adjust the amount of attenuation inserted in the transmission signal path based on the estimate α' of the acoustic feedback gain α, which is estimated by the total loss amount calculator 14. Next, an operation of the echo suppressor 30 when the echo suppressor 30 adjusts the amount of attenuation inserted in the transmission signal path will be described with reference to FIG.

The echo suppressor 30 constantly monitors the state of the voice switcher 10 (step 41). When the voice switcher 10 is in the receiving state, the echo suppressor 30 determines whether the double talk detector 23 detects a double talk (step 42). When the voice switcher 10 is not in the receiving state, or when the voice switcher 10 is in the receiving state, when the double talk detector detects a double talk, the echo suppressor 30 considers that there is no residual echo or voice signal. To be transmitted, and the input signal is output as it is without attenuation (step 47).

On the other hand, when the voice switcher 10 is in the receiving state and the double talk detector 23 does not detect the double talk, the echo suppressor 30 sets an attenuation coefficient, and multiplies the attenuation coefficient by the input signal. The standard value SUP_MIN is reached (step 43). Thereafter, when the sum of the estimated α' of the acoustic feedback gain α estimated by the voice switcher 10 and the total loss amount Lt(n) is the threshold value TH or more (step 44), the echo suppressor 30 will attenuate the coefficient. The standard value SUP_MIN is changed to be greater than the upper limit SUP_MAX of the standard value (step 45).

Further, when the sum of the estimated α' of the acoustic feedback gain α estimated by the voice switcher 10 and the total loss amount Lt(n) is smaller than the threshold value TH (step 44), the echo suppressor 30 maintains the standard value without change. SUP_MIN is used as the attenuation coefficient. Finally, the echo suppressor 30 attenuates the input signal by multiplying the input signal by the determined attenuation coefficient (standard value SUP_MIN or upper limit SUP_MAX) and outputs the result (step 46).

With this configuration, even if the echo suppressor 30 erroneously attenuates the speech from the speaker at the near end side, the voice inflection can be prevented.

Meanwhile, if the speaker is in an environment where the ambient noise is generally high, the adaptive filter 21 of the echo canceller 20 does not update the filter coefficients for a long period of time after the start of communication. For the reason described, the echo canceller 20 does not converge, and the double talk detector 23 cannot correctly detect the double talk. In this case, even if the voice switcher 10 is in the receiving state and the double talk detector 23 does not detect the double talk, the echo suppressor 30 can output the input signal as it is without attenuation, which may cause Uncomfortable residual echo.

In this regard, although the echo canceller 20 does not sufficiently converge after starting communication via the communication terminal of the other party, the echo canceller 30 is in the transmission signal path when the insertion loss amount distribution processor 15 estimates that the voice switcher 10 is in the reception state. Insert a specific amount of attenuation. That is, the echo suppressor 30 attenuates the uncomfortable residual echo, and the echo canceller 20 does not sufficiently converge after the start of communication, thereby achieving a pleasant environment for speakerphone communication.

On the other hand, if the total loss amount calculator 14 is operating in the fixed mode (e.g., immediately after the start of communication) while inputting different voices from the near-end side and the far-end side, the voice switcher 10 is in the receiving state and the transmitting state. Alternate between. For the reason described, the echo suppressor 30 does not operate properly, and thus an inflection may occur in the transmitted speech, where the level of the transmitted speech fluctuates greatly.

In this regard, when the insertion loss amount distribution processor 15 estimates that the voice switcher 10 is not in the reception state or when the specific amount of attenuation is being inserted in the transmission signal path, the double talk detector 23 detects the double talk, the echo The suppressor 30 immediately stops inserting a specific amount of attenuation into the transmission signal path. This makes it possible to prevent the inflection due to the erroneous attenuation of the transmission signal, thereby providing a pleasant condition for the speakerphone communication.

Fourth embodiment

Figure 9 is a block diagram showing a speakerphone apparatus in accordance with a fourth embodiment of the present invention. In this embodiment, it is characterized in that the echo suppressor 30 adjusts the amount of attenuation by using the signal level average value of the received signal, and the other basic configuration of the embodiment is the same as that of the third embodiment. Therefore, the same components are denoted by the same reference numerals, and the description thereof will be omitted.

In the present embodiment, the adaptive filter 21 recognizes the impulse response of the acoustic echo path Hac in an adaptive manner and based on the signal y(n) input from the far-end side (i.e., the received signal input to the speaker amplifier G2) To estimate the echo component (acoustic echo) g(n). To estimate the echo component g(n), update the filter coefficient hp(n) by using the following formula:

Hp(n+1)=hp(n)+K‧F(Ey(n))

Here, p is the number of taps, n is the sampling period, K is the step gain, and F(n) is the coefficient update function. Further, Ey(n) is a signal level average of the input signal (received signal) y(n) by using a signal level average (SLA) provided in the adaptive filter 21. The following formula in the arithmetic unit 24 is calculated. Here, SPAN is a period when the size of the input signal y(n) is abs[y(n)].

In addition, the echo component (estimation) g(n) can be calculated by using the following formula:

In the present embodiment, the echo suppressor 30 obtains the signal level average Ey(n) from the SLA operation unit provided in the adaptive filter 21.

Hereinafter, the operation of the echo suppressor 30 will be explained with reference to the flowchart shown in FIG.

The echo suppressor 30 constantly monitors the state of the voice switcher 10 (step 51). When the voice switcher 10 is in the receiving state, the echo suppressor 30 determines whether the double talk detector 23 of the echo canceller 20 detects a double talk (step 52). When the voice switcher 10 is not in the receiving state, or when the double talk detector 23 detects the double talk and the voice switcher 10 is in the receiving state, the echo suppressor 30 considers that there is no residual echo or voice signal. To be transmitted, and the input signal is output as it is without attenuation (step 58).

On the other hand, when the voice switcher 10 is in the receiving state and the double talk detector 23 does not detect the double talk, the attenuation coefficient multiplied by the input signal by the echo suppressor 30 depends on the SLA. The signal level average value RS_AVE(n) (=Ey(n)) calculated by the arithmetic unit 24. More specifically, when RS_AVE(n)>TH, the attenuation coefficient can be determined according to the following formula (1), and when RS_AVE(n)<TH, the attenuation coefficient can be determined according to the following formula (2):

Here, SUP_MAX and SUP_MIN are respectively suppressing a signal level corresponding to a loud voice (for example, OOdB sound pressure or more) and a signal level corresponding to a standard volume (for example, OO to OOdB sound pressure). The required attenuation factor, and RS_AVEmax is the loud signal level average RS_AVE(n). Regarding formula (2), when each sampling period RS_AVE(n)□TH, the attenuation coefficient is multiplied by a specific transition coefficient and gradually decreased from SUP_MAX×RS_AVE(n)/RS_AVEmax to SUP_MIN, where SUP_MIN is set to The lower limit of the attenuation factor.

More specifically, the echo suppressor 30 first determines whether the attenuation coefficient is set to SUP_MIN (step 53). When the attenuation coefficient is not set to SUP_MIN, the echo suppressor 30 multiplies it by the conversion coefficient, thereby setting a new attenuation coefficient (step 54). Otherwise, SUP_MIN is used as the attenuation coefficient because it is the lower limit of the attenuation coefficient.

Next, if RS_AVE(n) > TH, that is, the condition of the formula (1) is satisfied (step 55), the echo suppressor 30 sets the attenuation coefficient to SUP_MAX × RS_AVE(n) / RS_AVEmax (step 56), otherwise, Use the attenuation factor set in the previous step without changing the attenuation factor. Finally, the echo suppressor 30 attenuates the input signal for output by multiplying the input signal by the set attenuation coefficient (step 57).

With the present embodiment, even if the echo suppressor 30 accidentally attenuates the speech from the speaker at the near end side, the occurrence of the speech inflection can be suppressed. Furthermore, since the signal level average value RS_AVE(n) is calculated by using the signal level average operation unit 24 provided in the echo canceller 20, simplification of the configuration can be achieved (for example, the amount of program executed in the DSP) Reduction, lower specifications in the DSP, reduction in ROM capacity for storing programs, etc., and cost reduction.

Fifth embodiment

Next, a fifth embodiment of the present invention will be explained with reference to the drawings. In the present embodiment, the echo suppressor 30 combines the signal level average (SLA) provided in the adaptive filter 21 of the voice switcher 10, the double talk detector 23 of the echo canceller 20, and the echo canceller 20. The arithmetic unit 24 inserts an attenuation amount in the transmission signal path. Other configurations of this embodiment are the same as those of the fourth embodiment. Therefore, the same components are denoted by the same reference numerals, and the description thereof will be omitted.

The operation of the echo suppressor 30 will be explained in detail with reference to the flowchart shown in FIG. Steps 61 to 67 in Fig. 11 are the same as steps 51 to 57 in Fig. 10 regarding the fourth embodiment. Therefore, the description thereof will be omitted.

When the voice switcher 10 is not in the receiving state, or when the voice switcher 10 is in the receiving state, when the double talk detector 23 detects the double talk, the echo suppressor 30 further determines whether the voice switcher 10 is In the transmission state (step 68). When the voice switcher 10 is in the transmission state, the echo suppressor 30 determines that there is no residual echo or has a voice signal to be transmitted, and outputs the input signal as it is without attenuation (step 611).

If the communication is performed between the speakerphone device according to the present invention and a communication device using the earpiece (hereinafter referred to as "handset communication device"), and the surrounding noise is transmitted from the transmission signal transmitted from the speakerphone device to the earpiece communication device The level is quite high, and the ambient noise is temporarily attenuated when the echo canceller 30 in the speakerphone device attenuates the residual echo. This attenuation of ambient noise is heard by the handset after a delay, and thus the speaker at the handset communication device may experience an uncomfortable disconnection sensation in speech during communication.

In this regard, in the present embodiment, when the voice switcher 10 is not in the transmission state in step 68 (for example, the insertion loss amount distribution processor 15 inserts the same loss amount in the transmission signal attenuator 11 and the reception signal attenuator 12 (hereinafter, this state is referred to as "intermediate state")), and the echo suppressor 30 monotonically attenuates the amount of attenuation inserted in the transmission signal path.

More specifically, when the voice switcher 10 is not in the transmission state in step 68, the echo suppressor 30 further determines whether the attenuation coefficient is zero (step 69). If the attenuation coefficient is zero, the process returns to step 61 without changing the attenuation coefficient. Otherwise, the echo suppressor 30 sets a new attenuation coefficient by multiplying the attenuation coefficient by the conversion coefficient (step 610). That is, the echo suppressor 30 gradually reduces the attenuation coefficient (attenuation amount) by repeating the processing of step 61 or 62->step 68->step 69->step 610->step 61 (as shown in the line in FIG. 12). A shows monotonously reduced).

In the present embodiment, since the echo suppressor 30 monotonically reduces the amount of attenuation inserted in the transmission signal path, the surrounding noise is not temporarily attenuated. This makes it difficult to make the speaker at the earpiece communication device feel an uncomfortable disconnection feeling.

Meanwhile, when the attenuation coefficient (attenuation amount) is monotonically decreasing by repeating the processing of step 61 or 62->step 68->step 69->step 610->step 61, the communication state of the voice switcher 10 is changed to transmission. In the state, the echo suppressor 30 immediately reduces the attenuation coefficient (attenuation amount) to zero (see, for example, the broken line B in Fig. 12). Therefore, it is possible to prevent the inflection in the speech heard by the speaker at the counterpart communication device because the speech (transmission voice) from the speaker at the speakerphone device is accidentally attenuated by the echo suppressor 30.

Further, when the attenuation coefficient (attenuation amount) is monotonically decreasing by repeating the processing of step 61 or 62->step 68->step 69->step 610->step 61, the communication state of the voice switcher 10 is changed to transmission. When the state and the double talk detector 23 does not detect the double talk, the echo suppressor 30 inserts a signal in the transmission signal path by performing the processing of steps 63 to 67 (see, for example, the broken line C in FIG. 12). A specific amount of attenuation. This makes it possible to prevent the echo suppressor 30 from accidentally inserting an amount of attenuation and preventing an uncomfortable residual echo from occurring, and accurately attenuating only an unpleasant echo during communication.

In the present embodiment, the time ratio (reduction ratio) in which the echo suppressor 30 monotonically decreases the amount of attenuation is constant (see, for example, line A in Fig. 12). When the attenuation amount is linearly attenuated as described above, the period (transition period) required to decrease to zero can be shortened as compared with the case where the attenuation amount is gradually attenuated.

Sixth embodiment

Next, a sixth embodiment according to the present invention will be explained with reference to Figs. 13 to 15 . As shown in FIG. 13, the sixth embodiment differs from the fifth embodiment in that a step 712 of controlling the amount of attenuation based on ambient noise levels is provided between steps 77 and 78. Steps 71 to 79 and steps 710 and 711 are the same as steps 61 to 69 and steps 610 and 611 explained in the fifth embodiment, and a description thereof will be omitted.

In the present embodiment, the echo suppressor 30 further controls the amount of attenuation in conjunction with ambient noise levels. The process of controlling the amount of attenuation in conjunction with the surrounding noise level is illustrated in the flow chart shown in FIG.

That is, as the surrounding noise level on the side of the speakerphone device becomes higher, the amount of attenuation inserted by the echo suppressor 30 in the transmission signal path is reduced. Therefore, the instantaneous attenuation of the surrounding noise can be suppressed, thereby making it difficult to feel the feeling of disconnection. In this case, if the echo suppressor reduces the amount of attenuation, the attenuation effect of the residual echo is lowered. However, in actual use, the residual echo is covered by ambient noise and does not cause trouble in communication.

Hereinafter, a process of controlling the amount of attenuation in conjunction with the surrounding noise level will be described with reference to the flowchart shown in FIG.

The echo suppressor detects (estimates) the level of ambient noise (surrounding noise level) LEVEL_NOISE continuously present in the transmission signal by using the long-term average value of the transmission signal input from the input point Tin of the transmission signal attenuator 11, The estimated ambient noise level LEVEL_NOISE is compared to the threshold Nth (step 821). When the ambient noise level LEVEL_NOISE is the threshold Nth or less, the echo suppressor 30 calculates the standard attenuation coefficient ATT by using the following formula (step 822):

ATT=α×LEVEL_NOISE+β, α<0

Further, when the ambient noise level LEVEL_NOISE exceeds the threshold Nth, the echo suppressor 30 sets the standard attenuation coefficient ATT to a constant value ATTmin regardless of the surrounding noise level LEVEL_NOISE (step 823).

Subsequently, the echo suppressor 30 calculates the attenuation coefficient (corresponding to the attenuation amount of the back volume estimated from the received signal level) by the processing of steps 73 to 76 and the standard attenuation calculated by the processing of steps 821 to 823. The coefficients ATT are compared (step 824). When the attenuation coefficient is the standard attenuation coefficient ATT or less, the echo suppressor 30 updates the attenuation coefficient with the standard attenuation coefficient ATT (step 825), thereby proceeding to step 77. Further, when the attenuation coefficient is greater than the standard attenuation coefficient ATT, the echo suppressor 30 proceeds to step 77 without updating the attenuation coefficient.

That is, in steps 824 and 825, the echo suppressor 30 will correspond to the ambient noise based on the amount of attenuation of the back volume estimated from the level of the speech signal in the received signal path (see, for example, curve D in Figure 15). The amount of attenuation determined by the level (see, for example, curve E in Fig. 15) is compared, and then an attenuation amount having a smaller value among the attenuation amounts is inserted in the transmission signal path. Further, when the ambient noise level LEVEL_NOISE exceeds the threshold value Nth, the echo suppressor 30 sets the attenuation amount (attenuation coefficient) to a constant value ATT based on the surrounding noise level LEVEL_NOISE. This prevents the residual echo from being attenuated when the ambient noise level is very high.

While the invention has been shown and described with respect to the embodiments of the embodiments of the present invention, it will be understood that various changes and modifications can be made without departing from the scope of the invention as defined in the appended claims.

1. . . microphone

10. . . Voice switcher

11. . . Transmission signal attenuator

12. . . Receive signal attenuator

13. . . Insert loss controller

14. . . Total loss calculator

15. . . Insert loss allocation processor

2. . . speaker

20. . . Echo canceller

twenty one. . . Adaptive filter

twenty two. . . Subtractor

twenty three. . . Double talk detector

twenty four. . . Signal level average arithmetic unit

30. . . Echo suppressor

A. . . Receive signal path

B. . . Transmission signal path

C. . . Acoustic feedback path

D, E. . . curve

G1. . . Microphone amplifier

G2. . . Speaker amplifier

G3. . . Amplifier

α. . . Acoustic feedback gain

''. . . Acoustic feedback gain estimation

The objects and features of the present invention will become apparent from the following description of the preferred embodiments illustrated in the appended claims

1 is a block diagram of a first embodiment of the present invention.

Figure 2 presents a flow chart for explaining the operation of the voice switcher shown in Figure 1.

Fig. 3 shows a flow chart for explaining the operation of the echo suppressor shown in Fig. 1.

Figure 4 illustrates a block diagram in accordance with a second embodiment of the present invention.

Figure 5 is a flow chart for explaining the echo suppressor shown in Figure 4.

Figure 6 is a block diagram showing a third embodiment of the present invention.

Figure 7 presents a flow chart for explaining the operation of the echo suppressor depicted in Figure 6.

Fig. 8 is a flowchart for explaining the operation of the echo suppressor in the modified example of the third embodiment.

Figure 9 is a block diagram showing a fourth embodiment of the present invention.

Figure 10 depicts a flow chart for explaining the operation of the echo suppressor in the fourth embodiment.

Figure 11 presents a flow chart for explaining the operation of the echo suppressor in the fifth embodiment of the present invention.

Figure 12 is a diagram showing the amount of attenuation of the echo suppressor as a function of time in the fifth embodiment of the present invention.

Figure 13 is a flow chart for explaining the operation of the echo suppressor in the sixth embodiment of the present invention.

Figure 14 is a flow chart for explaining the operation of the echo suppressor in the sixth embodiment.

Fig. 15 is a view showing the amount of attenuation of the echo suppressor as a function of time in the sixth embodiment.

1. . . microphone

2. . . speaker

10. . . Voice switcher

11. . . Transmission signal attenuator

12. . . Receive signal attenuator

13. . . Insert loss controller

14. . . Total loss calculator

15. . . Insert loss allocation processor

20. . . Echo canceller

twenty one. . . Adaptive filter

twenty two. . . Subtractor

30. . . Echo suppressor

A. . . Receive signal path

B. . . Transmission signal path

C. . . Acoustic feedback path

G1. . . Microphone amplifier

G2. . . Speaker amplifier

G3. . . Amplifier

Claims (19)

A speakerphone device comprising: a microphone; a speaker; a voice switcher for switching between transmission and reception in a communication state by: providing loss in a received signal path and a transmission signal path, and communication from the other party a receiving signal of the terminal is transmitted to the speaker via the receiving signal path, and a transmission signal collected by the microphone is transmitted to the communication terminal of the other party via the transmission signal path; an echo canceller is used Acoustic echo caused by suppressing acoustic coupling between the microphone and the speaker; double talk detector for detecting double talk; and echo suppressor for combining the voice switch And/or the double talk detector to attenuate residual echoes by inserting a specific amount of attenuation in the transmission signal path, wherein the voice switcher includes: a transmission for inserting a loss in the transmission signal path a side loss insertion unit; a receiving side loss insertion unit for inserting a loss in the received signal path; and controlling by the a transmission loss loss insertion unit and a corresponding amount of the insertion loss amount control unit of the loss inserted by the receiving side loss insertion unit, the insertion loss amount control unit comprising: a total loss amount calculator for estimating a path of the self An output terminal of the receiving side loss insertion unit is fed back to an acoustic feedback gain of an input terminal of the transmission side loss insertion unit via an acoustic echo path, and the insertion in the path is calculated based on the estimation of the acoustic feedback gain a sum of the amount of loss; and an insertion loss amount allocation processor for monitoring the transmission signal and the received signal and estimating a communication state, and calculating based on the result of the estimation and by the total loss amount calculator And determining an allocation of the insertion loss amount to the transmission side loss insertion unit and the reception side loss insertion unit, and the total loss amount calculator has an update mode and a fixed mode, in the update mode, based on Calculating and updating the loss inserted in the path in an adaptive manner for a corresponding estimate of the feedback gain a sum of, in the fixed mode, the total loss amount is fixed to an initial value, and operates in the fixed mode when the echo canceller does not sufficiently converge after communication starts, and completely converges in the echo canceller Then operate in the update mode. The speakerphone device of claim 1, wherein if the insertion loss amount distribution processor estimates that the voice switcher is in a receiving state, the echo suppressor inserts in the transmission signal path The specific amount of attenuation is described, otherwise the specific amount of attenuation is not inserted. The speakerphone device of claim 1, wherein the echo suppressor does not insert the specific fading in the transmission signal path when the double talk detector is detecting a double talk Decrement, otherwise insert the specific amount of attenuation. The speakerphone device of claim 3, wherein if the double talk detector does not detect a double talk and the insertion loss amount distribution processor estimates that the voice switch is in the Receiving the state, the echo suppressor inserts the specific amount of attenuation in the transmission signal path. The speakerphone device of claim 4, wherein, when the echo canceller does not sufficiently converge after the start of communication, only the insertion loss amount distribution processor estimates that the voice switcher is located In the case of the reception state, the echo suppressor provides the specific amount of attenuation in the transmission signal path. The speakerphone device of claim 1, wherein the insertion loss amount distribution processor estimates that the voice switcher is not in the period of insertion of the specific attenuation amount in the transmission signal path When the receiving state or the double talk detector detects a double talk, the echo suppressor immediately stops inserting the attenuation amount into the transmission signal path. The speakerphone device of claim 1, wherein the echo canceller comprises: an adaptive filter for identifying a characteristic of an echo path in an adaptive manner; and a subtractor for transmitting the signal from the signal Subtracting the output of the adaptive filter; and the double talk detector for detecting a double talk, and when the double talk detector detects a double talk, the The echo canceller does not update the coefficients of the adaptive filter. The speakerphone device of claim 1, wherein the echo canceller, the voice switcher, and the echo suppressor are disposed in the transmission signal path in the order, and execute the same Corresponding processing. The speakerphone device of claim 1, wherein the echo suppressor adjusts the amount of attenuation based on the acoustic feedback gain estimated by the total loss amount calculator. The speakerphone device of claim 1, wherein the echo suppressor adjusts the amount of attenuation based on a signal level average of the received signal. The speakerphone device of claim 10, wherein the echo suppressor inserts the signal inserted in the transmission signal path when the signal level of the received signal is lower than a threshold value The amount of attenuation is set to a specific standard value, and when the signal level is the threshold or higher, the echo suppressor is in a range from an upper limit of the specific standard value to the specific standard value The amount of attenuation is adjusted internally. The speakerphone device of claim 10, wherein the echo canceller comprises: a signal level average (SLA) operation unit for calculating a signal level average of the received signal; a filter for identifying characteristics of the echo path in an adaptive manner; and a subtractor for subtracting an output of the adaptive filter from the transmission signal, and wherein the echo suppressor is used by The signal level average calculated by the signal level average calculation unit adjusts the attenuation amount. The speakerphone device of claim 1, wherein the double talk is not detected only by the double talk detector and the insertion loss amount distribution processor estimates that the voice switch is in the In the case of the reception state, the echo suppressor inserts a specific amount of attenuation in the transmission signal path, and when the insertion loss amount allocation processor estimates that the voice switcher is not in the receiving state, The echo suppressor when in the transmission state, or when the insertion loss amount distribution processor estimates that the voice switch is in the receiving state and the double talk detector detects a double talk The amount of attenuation inserted in the transmission signal path is monotonically reduced. The speakerphone device of claim 13, wherein when the communication state estimated by the insertion loss amount distribution processor becomes the transmission state when the attenuation amount is monotonically decreasing, The echo suppressor immediately reduces the amount of attenuation to zero. The speakerphone device of claim 13, wherein when the amount of attenuation inserted in the transmission signal path is monotonically decreasing, the double talk detector does not detect the double The end effector, when the communication state estimated by the insertion loss amount allocation processor is the reception state, the echo suppressor again inserts the specific attenuation amount in the transmission signal path. The speakerphone device according to any one of claims 13 to 15, wherein the communication state estimated by the insertion loss amount allocation processor is neither in the receiving state nor in the In the transmission state, or when the communication state is the receiving state and the double-talking detector detects a double-talking call, the echo suppressor makes the transmission signal path at a constant reduction ratio The amount of attenuation inserted in is reduced. The speakerphone device of claim 1, wherein the communication state estimated by the insertion loss amount allocation processor is only when the double talk detector does not detect a double talk. In the receiving state, the echo suppressor inserts the specific attenuation amount in the transmission signal path, and the echo suppressor estimates a level of ambient noise superimposed in the transmission signal, and is based on the An estimate of the surrounding noise level is described to reduce the amount of attenuation inserted in the transmission signal path when the ambient noise level becomes high. The speakerphone device of claim 17, wherein the echo suppressor combines a volume estimated by the signal level of the received signal with a surrounding noise level superimposed on the transmitted signal. The amount of attenuation to be compared is compared, and one of the return volume and the smaller of the attenuation amounts is inserted in the transmission signal path. The speakerphone device of claim 18, wherein, when the noise level exceeds a certain level, the echo suppressor adjusts the attenuation amount to a constant value based on the ambient noise level .