Transmitting / receiving call device, transmitting / receiving call program, transmitting / receiving call method, and audio communication system

Abstract

To provide a transmitting / receiving call device capable of effectively and efficiently performing echo cancellation processing in a remote device.SOLUTION: The present invention is a transmitting / receiving call device connected to a subscriber terminal and having a two-wire / four-wire conversion circuit and connected to an echo canceller device via a first network. The transmitting / receiving call device includes storage means for storing a downstream voice signal received from an opposing transmitting / receiving call device opposite the transmitting / receiving call device, and transmitting means for transmitting the downstream voice signal stored in the storage means to the echo canceller device together with an upstream voice signal received from the subscriber terminal and converted by the two-wire / four-wire conversion circuit.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The present invention relates to a handset, a program, a method, and an audio communication system. [Background technology]

[0002] For example, in a conventional voice communication system, a transmitter / receiver incorporating a hybrid circuit is provided with an echo canceller for removing hybrid echo.

[0003] In recent years, in order to meet the demands of miniaturization and power saving for handset devices connected to telephone terminals, as well as considerations of EOL (End Of Life), there is a demand for remote devices on the network to remove echo components from signals sent to the other party, eliminating the echo cancellation resources that have traditionally been built into handset devices.

[0004] Patent Document 1 discloses that a remote device on a network performs echo cancellation processing. Specifically, the remote device (equipped with an echo canceller) described in Patent Document 1 includes a propagation delay time estimation unit, and measures a round-trip propagation delay time between a near-end transmitter / receiver and an echo canceller device at the start of voice communication, and estimates the delay time (propagation delay time) of an echo signal reflected and returning from the near-end transmitter / receiver using the round-trip propagation delay time. The remote device then performs echo cancellation processing using the estimated delay time. [Prior art documents] [Patent documents]

[0005] [Patent Document 1] Patent Publication No. 2021-036634 [Patent Document 2] JP 2009-033344 A Summary of the Invention [Problem to be solved by the invention]

[0006] However, the technology described in Patent Document 1 has the following main problems.

[0007] First, the remote device described in Patent Document 1 has a problem that since it performs estimation processing (estimation of propagation delay time), some error occurs.

[0008] Secondly, since the propagation delay time cannot be calculated at the start of a call, an appropriate amount of delay cannot be added until the propagation delay time is calculated.

[0009] Thirdly, the remote device described in Patent Document 1 transmits a voice characteristic signal capable of detecting echo components to the near-end handset in order to measure the round-trip propagation delay time, but there is a possibility that the signal may be heard by the subscriber.

[0010] Therefore, there is a demand for a handset, a program, a method, and an audio communication system that are capable of effectively and efficiently performing echo cancellation processing in a remote device. [Means for solving the problem]

[0011] The first invention is a handset having a two-wire / four-wire conversion circuit connected to a subscriber terminal and connected to an echo canceller device via a first network, characterized in having (1) a memory means for storing a downstream voice signal received from an opposing handset opposite the handset, and (2) a transmission means for transmitting the downstream voice signal stored in the memory means to the echo canceller device together with an upstream voice signal received from the subscriber terminal and converted by the two-wire / four-wire conversion circuit.

[0012] The second transceiver program of the present invention is characterized in that it causes a computer mounted on a transceiver device connected to a subscriber terminal and having a two-wire / four-wire conversion circuit connected to an echo canceller device via a first network to function as (1) a memory means for storing a downstream voice signal received from an opposing transceiver device opposite the transceiver device, and (2) a transmission means for transmitting the downstream voice signal stored in the memory means to the echo canceller device together with an upstream voice signal received from the subscriber terminal and converted by the two-wire / four-wire conversion circuit.

[0013] The third invention is a transmitting / receiving method used for a handset device having a two-wire / four-wire conversion circuit connected to a subscriber terminal and connected to an echo canceller device via a first network, the method having a storage means and a transmission means, characterized in that (1) the storage means stores a downstream voice signal received from an opposing handset device opposite the handset device, and (2) the transmission means transmits the downstream voice signal stored in the storage means to the echo canceller device together with an upstream voice signal received from the subscriber terminal and converted by the two-wire / four-wire conversion circuit.

[0014] The fourth aspect of the present invention is a voice communication system having a handset and an echo canceller, (1) the handset is connected to a subscriber terminal and connected to the echo canceller via a first network, and has a two-wire to four-wire conversion circuit, (1-1) a storage means for storing a downstream voice signal received from an opposite handset opposite to the handset, and (1-2) a transmission means for transmitting the downstream voice signal stored in the storage means to the echo canceller together with an upstream voice signal received from the subscriber terminal and converted by the two-wire to four-wire conversion circuit, (2) The echo canceller device is connected to the handset via the first network and to an opposing handset via a second network, and includes (2-1) an echo canceller that, when a transmission input signal and a reception input signal are input, outputs a transmission output signal that cancels the echo signal generated in the 2-wire / 4-wire conversion circuit, and (2-2) the transmission input signal input to the echo canceller is an upstream voice signal received from the handset, and the reception input signal is a downstream voice signal received from the handset. Effect of the Invention

[0015] According to the present invention, echo cancellation processing in a remote device can be performed effectively and efficiently. [Brief description of the drawings]

[0016] [Figure 1] 1 is an overall configuration diagram showing the overall configuration of an audio communication system according to an embodiment; [Diagram 2] FIG. 2 is an explanatory diagram illustrating an example of an echo canceller according to the embodiment; [Diagram 3] FIG. 4 is an explanatory diagram showing an example of echo path characteristics. [Figure 4] FIG. 13 is a diagram showing the overall configuration of a voice communication system according to a modified embodiment. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0017] (A) Main embodiment DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, an embodiment of a handset, a handset program, a handset method, and an audio communication system according to the present invention will be described in detail with reference to the drawings.

[0018] (A-1) Configuration of the embodiment (A-1-1) Overall composition FIG. 1 is a diagram showing the overall configuration of a voice communication system according to an embodiment.

[0019] In FIG. 1, the voice communication system 1 includes a transceiver device 2, an echo canceller device 3, a subscriber terminal 4, a near-end network (hereinafter also referred to as the "first network") N, and a far-end network (hereinafter also referred to as the "second network") M.

[0020] In the following description, it is assumed that a line echo occurs in the hybrid circuit 21 of the transceiver 2. The transceiver 2 connected to the echo path (echo route) will be referred to as the "near-end transceiver," and the opposing transceiver will be referred to as the "far-end transceiver" (not shown).

[0021] In the voice communication system 1, a near-end transmitting / receiving device 2 is connected to a near-end network N. An echo canceller device 3 is connected to the near-end network N and a far-end network M. Voice signals (voice packets) communicated between the near-end transmitting / receiving device 2 and a far-end transmitting / receiving device (not shown) are transmitted and received via the near-end network N, the echo canceller device 3, and the far-end network M.

[0022] The near-end network N may be, for example, a fixed-line telephone network (PSTN), a legacy IP network already installed by a carrier, or a combination of these. In this embodiment, an IP network is used as an example. The far-end network M is, for example, an IP network such as the Internet.

[0023] In this embodiment, a case will be described in which voice communication is performed with an echo canceller device 3 as a remote device interposed between a near-end network N as an IP network and a far-end network M as an IP network.

[0024] The subscriber terminal 4 is used by a near-end talker, and is connected to the near-end transmitter / receiver 2. The subscriber terminal 4 may be, for example, a fixed telephone set.

[0025] The near-end transmitter / receiver 2 has a hybrid circuit (two-wire-four-wire conversion circuit) 21 , a first buffer 22 , a second buffer 23 , a third buffer 24 , and a transmission unit 25 .

[0026] The near-end transmitting / receiving device 2 may be realized by installing a program (the transmitting / receiving program according to the embodiment) in a computer having a processor, memory, etc., but even in this case, the near-end transmitting / receiving device 2 can be functionally shown in Fig. 1. Note that the near-end transmitting / receiving device 2 may be realized in part or in whole by hardware.

[0027] The near-end transmitter / receiver 2 is connectable to a subscriber terminal 4 and a near-end network N, and performs voice communication with the subscriber terminal 4.

[0028] The hybrid circuit 21 converts between a subscriber line (two-wire type) and a trunk line (four-wire type) connected to the subscriber terminal 4, and outputs the signal as an audio signal (upstream audio signal). When a signal (downstream audio signal) from the far end reaches the subscriber terminal 4 via the hybrid circuit 21, a part of the signal is reflected by the hybrid circuit 21 due to impedance mismatch of the hybrid circuit 21, which generates a loop echo (upstream audio signal).

[0029] The first buffer 22 buffers the downstream audio signal for a certain period of time using the FIFO method.

[0030] The second buffer 23 buffers the downstream audio signal from the first buffer 22 for a certain period of time.

[0031] The third buffer 24 buffers the upstream audio signal from the hybrid circuit 21 for a certain period of time.

[0032] The transmitting unit 25 performs a process of adding a sequence number capable of identifying a pair of the downstream voice signal buffered in the second buffer 23 and the upstream voice signal buffered in the third buffer 24. The transmitting unit 25 transmits a transmission packet P including the sequence number, the upstream voice signal, and the downstream voice signal to the near-end network N.

[0033] The echo canceller device 3 includes an echo canceller 30 and a jitter buffer 35 .

[0034] The echo canceller device 3 may be realized by installing a program in a computer having a processor, memory, etc., but even in this case, the echo canceller device 3 can be functionally shown in Fig. 1. Note that the echo canceller device 3 may be realized in part or in whole by hardware.

[0035] The echo canceller device 3 is connectable to the near-end network N and the far-end network M, and transmits and receives signals between the near-end network N and the far-end network M.

[0036] The echo canceller 30 receives an upstream signal contained in a transmission packet P received via the near-end network N, and removes an echo signal (echo component) contained in the input signal. The echo canceller 30 generates a pseudo echo using an adaptive filter or the like based on the downstream signal contained in the transmission packet P, and removes the echo signal using the generated pseudo echo.

[0037] In this embodiment, the near-end transmitting / receiving device 2 is not equipped with an echo canceller, and instead, an echo canceller 30 is provided in an echo canceller device 3 on the network to perform echo cancellation processing between the opposite party. Therefore, since the echo canceller of the near-end transmitting / receiving device 2 can be eliminated, the near-end transmitting / receiving device 2 can be made smaller and more power-saving.

[0038] The jitter buffer 35 buffers the downstream voice signal from the far-end side for a certain period of time. The echo canceller device 3 transmits the downstream voice signal in the jitter buffer 35 to the near-end side network N by using a communication means (not shown).

[0039] (A-1-2) Details of echo canceller 30 FIG. 2 is an explanatory diagram illustrating an example of an echo canceller according to an embodiment.

[0040] As shown in FIG. 2, the echo canceller 30 includes a pseudo echo generator 31 as an adaptive filter, a subtractor 32, a double-talk detector 33, and an NLP (non-linear processor) .

[0041] A receiving input signal x(n) is input to a receiving input terminal Rin of the echo canceller. In this embodiment, the receiving input signal x(n) is not a downstream voice signal directly received from the far end, but corresponds to a downstream voice signal included in a transmission packet P transmitted from the near end. This receiving input signal x(n) is input to a pseudo echo generator 31 and a double-talk detector 33.

[0042] The pseudo echo generator 31 is an adaptive filter with a learning function, which generates a pseudo echo signal yr(n) from a receiving input signal x(n) inputted from a receiving input terminal Rin. The pseudo echo generator 31 regards the residual signal e(n) outputted from a residual output terminal RES as an error generated by the difference between the adaptive filter characteristic and the echo path transfer characteristic, and updates the coefficients of the adaptive filter so that the adaptive filter characteristic converges to the echo path transfer characteristic.

[0043] A transmission input signal d(n) is input to a transmission input terminal Sin of the echo canceller. In this embodiment, the transmission input signal d(n) corresponds to an upstream voice signal included in a transmission packet P transmitted from the near end.

[0044] The transmission input signal d(n) is a signal obtained by adding the echo y(n) in the hybrid circuit (echo path) 21 and the near-end transmission signal t(n) sent from the near-end talker, as shown in the following equation (1). d(n)=y(n)+t(n) …(1)

[0045] The subtractor 32 subtracts the pseudo echo signal yr(n) from the transmission input signal d(n) as shown in the following equation (2), and outputs a residual signal e(n). e(n) = d(n) - yr(n) … (2)

[0046] The residual signal e(n) output from the subtractor 32 contains a residual echo Δy(n) generated by the difference between the filter characteristic and the echo path transfer characteristic shown in the following equation (3), and the near-end transmission signal t(n). Δy(n)=y(n)-yr(n) …(3)

[0047] Substituting equation (1) into equation (2), we obtain equation (4). e(n)=d(n)-yr(n)=y(n)+t(n)-yr(n) =(y(n)-yr(n))+t(n) …(4)

[0048] Equation (5) is established from equations (3) and (4). e(n) = Δy(n) + t(n) … (5)

[0049] When the received input signal x(n) is null (x(n) = 0), the only signal component of the residual signal e(n) is the near-end transmission signal t(n). Therefore, it is necessary to control the adaptive filter coefficients so that they do not diverge by disabling the update of the adaptive filter coefficients.

[0050] When a voice signal is present in the received input signal x(n) and the near-end transmission signal t(n) is absent (t(n)=0), that is, in the single-talk state, the signal component of the residual signal e(n) is the residual echo Δy(n) only, so if the coefficients of the adaptive filter are updated by regarding this residual signal e(n) as an error caused by the difference between the adaptive filter characteristics and the echo path transfer characteristics, the characteristics of the adaptive filter will converge to the characteristics of the echo path. Therefore, in the single-talk state, it is sufficient to enable the coefficient update of the adaptive filter.

[0051] When a voice signal exists in both the received input signal x(n) and the near-end transmission signal t(n), that is, in a double-talk state, the residual signal e(n) contains the near-end transmission signal t(n) and the residual echo Δy(n) generated by the difference between the adaptive filter characteristics and the echo path transfer characteristics. Therefore, it is necessary to control the adaptive filter coefficients so that they do not diverge by disabling the update of the adaptive filter coefficients.

[0052] Therefore, the double-talk detector 33 monitors the received input signal x(n), the transmitted input signal d(n) and the residual signal e(n) to determine the communication state, and performs control to disable the update of the adaptive filter coefficient when there is no received input signal or when the communication state is a double-talk state, and to enable the update of the adaptive filter coefficient when the communication state is a single-talk state. The specific detection method used by the double-talk detector 33 is not particularly limited, but for example, the method described in Patent Document 2 can be applied.

[0053] The double-talk detector 33 also controls the enable / disable of the NLP 34 based on the call state. That is, in the single-talk state, the processing of the NLP 34 is enabled, and in the double-talk state, the processing of the NLP 34 is disabled.

[0054] The NLP 34 is enabled by the double-talk detector 33, and if the level of the residual signal e(n), which is the output of the subtractor 32, is below a predetermined level, it executes processing (nonlinear processing) such as forcibly setting the signal to 0 (silence) or attenuating it using a center clipper or the like.

[0055] (A-2) Operation of the embodiment Next, the operation of the audio communication system 1 according to the embodiment having the above configuration will be described.

[0056] 1, it is assumed that mutual communication of voice signals (voice data) is started between the near-end transceiver 2 and the far-end network M (far-end transceiver, not shown). That is, the echo canceller device 3 transmits the far-end downstream voice signal received from the far-end network M to the near-end transceiver 2 via the jitter buffer 35.

[0057] (A-2-1) Operation in the near-end transmitter / receiver device 2 The near-end transmitter / receiver 2 accumulates in a first buffer 22 a downstream voice signal received by a receiving means (not shown).

[0058] Next, the near-end transmitter / receiver 2 extracts downstream voice signals from the first buffer 22 in accordance with the FIFO method, and sends the downstream voice signals to the hybrid circuit 21 and the second buffer 23, respectively.

[0059] When a downstream voice signal is input, the hybrid circuit 21 converts between a trunk line (four-wire type) and a subscriber line (two-wire type) connected to the subscriber terminal 4, and provides the converted voice signal (downstream voice signal) to the subscriber terminal 4 (voice is output from the subscriber terminal 4).

[0060] At that time, a part of the downstream voice signal is reflected in the hybrid circuit 21, and an echo generated in the echo path is mixed with the voice signal (upstream voice signal) input from the subscriber terminal 4. The hybrid circuit 21 sends the upstream voice signal to the third buffer 24.

[0061] The transmitter 25 retrieves the downstream voice signal from the second buffer 23 and the upstream voice signal from the third buffer 24, and transmits a transmission packet P including both voice signals and a sequence number to the near-end network N.

[0062] (A-2-2) Operation within the echo canceller device 3 (echo canceller 30) A downstream voice signal of a transmission packet P received from the near-end transceiver 2 is input as a reception input signal x(n) to a reception input terminal Rin of the echo canceller 30. The reception input signal x(n) is input to a pseudo echo generator 31 and a double-talk detector 33.

[0063] The pseudo echo generator 31 generates a pseudo echo signal yr(n) from the received input signal x(n) and outputs the generated pseudo echo signal yr(n) to the subtractor 32 and the double-talk detector 33 .

[0064] On the other hand, an upstream voice signal of a transmission packet P is input as a transmission input signal d(n) to a transmission input terminal Sin of the echo canceller 30. The transmission input signal d(n) is input to a subtractor 32 and a double-talk detector 33.

[0065] The subtractor 32 subtracts the pseudo echo signal yr(n) from the input transmission input signal d(n) to calculate a residual signal e(n), and outputs the calculated residual signal e(n) to a residual output terminal RES and a double-talk detector 33.

[0066] The double-talk detector 33 detects (determines) a single-talk state (a state in which only the far-end reception input signal x(n) is significant) based on the input transmission input signal d(n), reception input signal x(n), and residual signal e(n). If the double-talk detector 33 detects a single-talk state, it enables the pseudo echo generator 31 (filter coefficient learning) and the NLP 34, and if it does not detect a single-talk state, it disables the pseudo echo generator 31 (filter coefficient learning) and the NLP 34.

[0067] The residual output terminal RES outputs the residual signal e(n) to the pseudo echo generator 31 and the NLP .

[0068] The NLP 34 is enabled by the double-talk detector 33, and when a predetermined condition is met, it performs processing such as center clipping on the residual signal e. The processed signal is output from the transmission output terminal Sout. On the other hand, when the NLP 34 is disabled by the double-talk detector 33, the residual signal e is output as is from the transmission output terminal Sout. A packet including the signal output from the transmission output terminal Sout (audio signal after echo cancellation) is transmitted to the far-end network M by communication means not shown.

[0069] Furthermore, when the pseudo echo generator 31 is enabled by the double-talk detector 33, the pseudo echo generator 31 learns the filter coefficients.

[0070] (A-3) Effects of the embodiment According to this embodiment, the following effects are obtained.

[0071] The downstream voice signal from the far end to the near end is returned in the upstream direction as close as possible to the echo generation point (the hybrid circuit 21 of the near end transmitter / receiver 2) and transmitted as one packet together with the upstream voice signal from the near end to the far end to the echo canceller device 3 (echo canceller 30) arranged in the network. The downstream voice signal in the packet is input to the receiving input terminal Rin of the echo canceller 30 as the receiving input signal x(n), and the upstream voice signal in the packet is input to the transmitting input terminal Sin as the transmitting input signal d(n).

[0072] This makes it possible to reduce the net delay in the echo path, which is a problem with echo cancellation on a conventional network, and thus improve the convergence speed and stability of the echo canceller, thereby improving the speech quality.

[0073] The advantage of being able to reduce the net delay will be further described below with reference to examples of echo path characteristics.

[0074] Fig. 3 is an explanatory diagram showing an example of echo path characteristics. In the example of Fig. 3, the net delay is 16 ms. If the net delay is known, then the echo canceller 30 simply inputs a signal delayed by the known net delay time to the adaptive filter (pseudo echo generator 31) via the receiving input terminal Rin.

[0075] However, since the net delay is usually unknown, it is common to increase the filter length of the adaptive filter to cover the maximum value of "net delay + dispersion time." The larger the filter length of the adaptive filter is, the more the performance of the echo canceller deteriorates due to the following problems (1) and (2).

[0076] (1) The convergence speed of the adaptive filter becomes slower The adaptive filter learns the filter coefficients by feeding back the echo (residual signal e(n)) that could not be completely eliminated due to the error between the actual echo and the pseudo echo. The larger the filter length of the adaptive filter, the smaller the amount of feedback to each coefficient becomes, and the slower the convergence speed becomes.

[0077] For example, a known algorithm for updating coefficients of an adaptive filter is the normalized LMS (Least Mean Square) algorithm. The following equation (6) is an example of coefficient updating using the normalized LMS algorithm.

number

[0078] where α is the step size parameter, h k (n) is the coefficient sequence before updating the filter coefficients, h k (n+1) is the updated coefficient sequence of the filter coefficients, and M is the tap length of the filter. β is usually a constant greater than 0, but β=0 may be used.

[0079] Also, ∥x(n)∥2 is the square norm of the received input signal x(n), and ∥x(n)∥ 2 =x 2 (n)+x 2 (n-1)+…+x 2 (n-M+1).

[0080] Here, if a variable called step size μ is defined as in equation (7), equation (6) can be expressed as equation (8) using μ.

[0081]

number

[0082]

number

[0083] The above equation (7) is obtained by extracting a part of the above equation (6) as a variable called step size μ, which indicates the magnitude of the update amount of the filter coefficients.

[0084] For example, if the sampling process is 8 kHz and the filter length is 16 ms, then M = 8 kHz × 16 ms = 128. If the filter length is increased, M will also increase, and ∥x(n)∥ 2 As a result, the denominator of equation (6) becomes larger and the coefficient update amount becomes smaller, so that the feedback amount to each coefficient becomes smaller and the convergence speed decreases.

[0085] (2) Decreased stability The larger the filter length, the lower the accuracy of single talk determination. For example, if the net delay is extremely long, such as 1 second, it is easy to imagine that the accuracy of the determination will decrease. If it is mistakenly determined to be single talk when it is not, and filter coefficient learning is performed, the coefficient will diverge. In other words, stability will decrease.

[0086] As described above, in this embodiment, since the net delay can be reduced as described above, the problems described in (1) and (2) above can be solved.

[0087] (B) Other embodiments Although various modified embodiments have been described in the above embodiment, the present invention can also be applied to the following modified embodiments.

[0088] (B-1) In the voice communication system 1 described in the above embodiment, the transceiver device 2 transmits a transmission packet P including a downstream voice signal as well as an upstream voice signal to the echo canceller device 3. However, as a variant, a sequence number identifying the downstream voice signal may be transmitted instead of the downstream voice signal.

[0089] FIG. 4 is a diagram showing the overall configuration of a voice communication system according to a modified embodiment.

[0090] Among the components in Fig. 4, those having the same or corresponding configurations as those in Fig. 1 have the same configurations and will not be described below. The unique features of the voice communication system 1A of the modified example will be described below.

[0091] The echo canceller device 3A includes a sequence processor 36, a downstream voice signal storage unit 37, and a receiver 38 in addition to the above-mentioned echo canceller 30 and jitter buffer 35.

[0092] The sequence processing unit 36 ​​assigns a sequence number to the downstream audio signal from the jitter buffer 35 so that the audio signal can be identified, and then stores the downstream audio signal in the downstream audio signal storage unit 37 in association with information on the assigned sequence number.

[0093] The downstream tone signal and the sequence number are transmitted as a transmission packet PA to the near-end network N by a communication means (not shown).

[0094] The downstream voice signal storage unit 37 is a storage unit that temporarily stores information on the downstream voice signal and the sequence number.

[0095] When the receiving unit 38 receives a transmission packet PB described later, it searches for downstream voice signals stored in the downstream voice signal storage unit 37 using the sequence number included in the transmission packet PB as a key, and inputs the hit downstream voice signal to the receiving input terminal Rin of the echo canceller 30. The receiving unit 38 also inputs the upstream voice signal included in the received transmission packet PB to the transmitting input terminal Sin.

[0096] In the transmitter / receiver device 2A, instead of the above-mentioned first buffer 22, second buffer 23, and transmission unit 25, a first buffer 22A, a second buffer 23A, and a transmission unit 25A are applied.

[0097] The first buffer 22A is the same as the above-mentioned first buffer 22, except that it holds sequence numbers in addition to downstream voice signals.

[0098] The second buffer 23A is the same as the above-mentioned second buffer 23, except that the second buffer 23A holds a sequence number capable of identifying a downstream voice signal instead of the downstream voice signal.

[0099] The transmitting unit 25A is the same as the transmitting unit 25 described above, except that instead of the transmitting packet P, the transmitting unit 25A transmits a transmitting packet PB including an upstream signal and the sequence number stored in the second buffer 23A to the near-end network N.

[0100] According to the above-described modified voice communication system 1A, there is no longer a need to return the downstream voice signal from the near-end side to the far-end side, so that in addition to the effects described in the above-described embodiment, it is possible to obtain further effects such as reducing the amount of communication in the near-end network N.

[0101] (B-2) In the voice communication system 1 described in the above embodiment, the transmitting / receiving device 2 transmits the upstream voice signal and the downstream voice signal together with a sequence number, but this sequence number may be omitted. [Explanation of symbols]

[0102] 1, 1A...voice communication system, 2, 2A...transmitter / receiver device, 3, 3A...echo canceller device, 4...subscriber terminal, 21...hybrid circuit, 22, 22A...first buffer, 23, 23A...second buffer, 24...third buffer, 25, 25A...transmitter, 30...echo canceller, 31...pseudo echo generator, 32...subtractor, 33...double talk detector, 34...NLP, 35...jitter buffer, 36...sequence processor, 37...downstream voice signal storage unit, 38...receiver, M...far-end network, N...near-end network, P, PA, PB...transmitted packet.

Claims

1. A handset connected to a subscriber terminal and having a two-wire to four-wire conversion circuit connected to an echo canceller via a first network, a storage means for storing a downstream voice signal received from an opposite transceiver device opposite to said transceiver device; a transmitting means for transmitting the downstream voice signal stored in the memory means to the echo canceller device together with an upstream voice signal received from the subscriber terminal and converted by the two-wire to four-wire conversion circuit; A transmitting / receiving device comprising:

2. A computer installed in a handset having a two-wire to four-wire conversion circuit connected to a subscriber terminal and connected to an echo canceller via a first network, a storage means for storing a downstream voice signal received from an opposite transceiver device opposite to said transceiver device; a transmitting means for transmitting the downstream voice signal stored in the memory means to the echo canceller device together with an upstream voice signal received from the subscriber terminal and converted by the two-wire to four-wire conversion circuit; A call transmission / reception program characterized by causing the call transmission / reception program to function as follows.

3. 1. A method for use in a handset device having a two-wire to four-wire conversion circuit connected to a subscriber terminal and connected to an echo canceller device via a first network, comprising: A storage means and a transmission means are provided, The storage means stores a downstream voice signal received from an opposite transceiver device opposite to the transceiver device, The transmission means transmits the downstream voice signal stored in the storage means to the echo canceller device together with the upstream voice signal received from the subscriber terminal and converted by the two-wire to four-wire conversion circuit. A method for transmitting and receiving a call, comprising:

4. In a voice communication system having a transmitter / receiver and an echo canceller, The handset is connected to a subscriber terminal and to the echo canceller via a first network, and includes a two-wire to four-wire conversion circuit; a storage means for storing a downstream voice signal received from an opposite transceiver device opposite to the transceiver device; a transmission means for transmitting the downstream voice signal stored in the storage means to the echo canceller device together with an upstream voice signal received from the subscriber terminal and converted by the two-wire to four-wire conversion circuit, the echo canceller device is connected to the transceiver device via the first network and is connected to an opposing transceiver device via a second network; an echo canceller for receiving a transmission input signal and a reception input signal, and outputting a transmission output signal obtained by canceling an echo signal generated in the two-wire to four-wire conversion circuit; The transmission input signal input to the echo canceller is an upstream voice signal received from the handset, and the reception input signal is a downstream voice signal received from the handset.

1. A voice communication system comprising:

Images