JP5367554B2 - Network equipment - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To remove an echo component in one's opposing party over a digital network. <P>SOLUTION: A multi-interface unit 12 is connected to an IP door terminal unit 16 via the digital network 20. At the side of the terminal unit 16, output sound of a speaker 24 is input into a microphone 26, and hence an echo component is generated. In this case, a sound signal x or y transmitted and received between the unit 12 and the terminal unit 16 is delayed by six buffers 30, 50, 52, 54, 34 and 36. An adaptive filter 32 performs adaptive operation, based on only a specified part after a lapse of prescribed time when delay time by the buffers 30, 50, 52, 54, 34 and 36 is considered, in the sound signal x input to the adaptive filter 32. Thus, echoes are canceled over the digital network 20 from the unit 12 to the terminal unit 16. <P>COPYRIGHT: (C)2011,JPO&amp;INPIT

Description

The present invention relates to a network device, in particular, a sound information output means for outputting a first sound information corresponding to the first sound signal, exists between those spatial provided in the same space as the sound information output means A network in which a plurality of counterpart devices each having sound information detecting means for detecting second sound information and generating a second sound signal corresponding to the detected second sound information are connected via a digital network Relates to the device.

Conventionally, as this type of network device, for example, there is a multi-interface unit (N-8000MI) disclosed in Non-Patent Document 1. This multi-interface unit is used in a station communication system 100 as shown in FIG. 5 , for example. That is, in the station communication system 100, the multi-interface unit 102 is installed in a large-scale station such as a main station. The multi-interface unit 102 includes a portable terminal device 104 such as a PHS (Personal Handy-phone System) terminal device used exclusively in the same large-scale station premises and a small station such as another station, particularly an unmanned station. IP (Internet) as a kind of intercom slave installed at the station
Protocol) Controls bidirectional communication with the door terminal device (N-8540DS) 106. For this reason, the multi-interface unit 102 is connected to a PBX (Private Branch eXchange) 108 installed in the same large-scale station premises, and at the same time, an IP door terminal device 106 via a dedicated digital network 110. It is connected to the. In general, since a large-scale station has a plurality of attendants, a plurality of portable terminal devices 104, 104,... According to the number of persons are prepared. In addition, when there are a plurality of small-scale stations, a plurality of IP door terminal devices 106, 106,.

  Here, when the call button 112 of the IP door terminal device 106 installed in a certain small station is pressed, a call signal corresponding to this is sent from the IP door terminal device 106 via the digital network 110 to the large-scale station. Sent to. The call signal transmitted to the large-scale station is transmitted to the PBX 108 via the multi-interface unit 102 and further transmitted to each (or a specific) mobile terminal device 104 via the PBX 108. Each mobile terminal device 104 that has received this calling signal outputs a predetermined ringing sound, and when a predetermined operation is performed by any of the mobile terminal devices 104 in response to this, from the mobile terminal device 104 A response signal is transmitted to the caller IP door terminal device 106 through a route opposite to the call signal. When the caller IP door terminal device 106 receives this response signal, bidirectional communication between the IP door terminal device 106 and the mobile terminal device 104 of the response source is possible, that is, a call is possible. Note that the IP door terminal device 106 is capable of so-called hands-free calling.

By the way, in the IP door terminal device 106 capable of hands-free calling as described above, the output sound of the speaker 114 is input to the microphone 116 , so that the output signal of the microphone 116 , that is, the mobile phone device 104 of the other party is transmitted. Echo component appears in the audio signal. Therefore, in order to realize a good call, it is necessary to remove this echo component. Therefore, each IP door terminal device 106 is equipped with an echo cancellation function (echo canceller) for removing the echo component.

“Packet Intercom System / N-8000 Series” catalog, catalog No. B-083 TH03, February 2006, manufactured by TIO Co., Ltd.

  However, when the IP door terminal device 106 is provided with an echo canceling function as described above, the configuration of the IP door terminal device 106 is naturally complicated. As a result, the price of the IP door terminal device 106 increases, and as a result, the entire station communication system 100 including the IP door terminal device 106 increases in price. This becomes more significant as the number of IP door terminal devices 106, that is, the number of small-scale stations increases.

  In order to solve this problem, for example, the multi-interface unit 102 is equipped with an echo canceling function, and the echo canceling function of the multi-interface unit 102 centralizes the echo components related to all the IP door terminal devices 106, 106,. However, in this case, since the echo component is removed through the digital network 110 for the multi-interface unit 102, another problem occurs. That is, since the signal to be transmitted is transmitted in a packetized (encoded) state between the multi-interface unit 102 and each IP door terminal device 106, the processing time required for this packetization In addition, various delay times including a processing time necessary for returning the packetized signal to the original state (decoding) occur. The delay time also varies depending on the transmission path capacity (transmission speed) between the both 102 and 106, particularly the transmission path capacity of the branch line connecting the digital network 110 and each IP door terminal device 106. For example, the digital network 110 as a main line is usually a broadband transmission line such as an optical line including a line connecting the multi-interface unit 102 and the digital network 110. The branch line connecting to the IP door terminal device 106 may be a narrow band transmission line such as a general analog telephone line or a medium band transmission line such as an ISDN (Integrated Services Digital Network) line. With respect to the IP door terminal device 106 connected to such a narrow-band or medium-band branch line, packet communication is performed with the multi-interface unit 102 according to the transmission line capacity on the branch line side having a small transmission line capacity. The delay time here becomes longer. On the other hand, since echo cancellation is a kind of system identification, in a system including such a delay time, the delay time is also an identification target. Therefore, the system including the delay time mentioned here cannot be properly identified simply by applying the conventional echo cancellation function such as that mounted on the IP door terminal device 106 to the multi-interface unit 102 as it is. . Strictly speaking, in order to properly identify a system including the delay time, an adaptive filter (DSP: Digital Signal Processor) having a correspondingly high performance (for example, an extremely large tap length and an extremely high processing speed). However, such an adaptive filter does not exist in practical use.

  Accordingly, an object of the present invention is to provide a network device capable of system identification including echo cancellation over a digital network.

To this end, the network device of the present invention, a sound information output means for outputting a first sound information corresponding to the first sound signal, those wherein is provided in the same space as the sound information output means a plurality of opposite device having a sound information detecting means, respectively for generating a second sound signal corresponding to the second sound information the detected by detecting the second sound information present between the sky via a digital network It is assumed that they are connected. Under this premise, the transmission means for transmitting the first sound signal to the sound information output means of the counterpart device that is communicating via the digital network, and the counterpart device that is communicating the second sound signal via the digital network Receiving means for receiving from the sound information detecting means . Furthermore, generated based simulation signal to the echo component simulates a wraparound component to sound information detection means of the first sound information in the other party device in communication with the first sound signal is input to the first sound signal Adaptive filter means is provided. In addition to this, delay means for delaying one or both of the first sound signal transmitted from the transmitting means to the communicating counterpart device and the second sound signal received from the communicating counterpart device by the receiving means, It has . In addition, the adaptive filter means excludes the portion of the first sound signal input to itself from the current time to a time point that is traced back by a predetermined period considering the delay time by the delay means. and then, based on the specific parts up to the point of going back to the predetermined period only just past period, further accordance with the tap length of itself from the timing predated past, performing adaptive processing, is that.

According to this configuration, the network device of the present invention is connected to a plurality of counterpart devices via a digital network . Each counterpart device has sound information output means and sound information detection means existing in the same space . Then, communication with any of the other party devices becomes possible. For example , the first sound signal is transmitted from the transmission unit constituting the network device to the sound information output unit of the other party device that is communicating via the digital network. When transmitted, first sound information corresponding to the first sound signal is output from the sound information output means in the counterpart device in communication . At this time, the second sound information existing in the space in which the first sound information is output is detected by the sound information detecting means of the counterpart device in communication . The sound information detecting means outputs a second sound signal corresponding to the detected second sound information, and the second sound signal is received by the receiving means via the digital network. Here, the adaptive filter means , based on the first sound signal input to itself, a simulated signal simulating an echo component that is a wraparound component of the first sound information to the sound information detecting means in the communicating counterpart device. Generate. In short, the adaptive filter means identifies the transfer function (acoustic characteristic) of the space from the sound information output means to the sound information detection means of the counterpart device in communication through the digital network. However, in the transmission of the first sound signal and the second sound signal to / from the counterpart device that is communicating via the digital network, the delay time is generated as described above, so that the adaptive filter means has this delay time. The identification is performed in consideration. That is, in the present invention, in addition to the delay time, one or both of the first sound signal and the second sound signal are further delayed (intentionally) by the delay means. Then, the adaptive filter means excludes the portion of the first sound signal input to itself from the present time to the time point that has been traced back by a predetermined period considering the delay time by the delay means. Then , the adaptive processing is performed based on a specific part from the time point that goes back in the past by the predetermined period to the time point that goes back in the past by a period corresponding to its own tap length . That is, the portion to be subjected to adaptive processing is narrowed down. This enables practical system identification over digital networks.

  In the present invention, the delay means considers the balance of the transmission of the first sound signal and the second sound signal via the digital network and applies both the first sound signal and the second sound signal for the same time. It is desirable to delay.

In addition, the first sound signal and the second sound signal transmitted via a digital network are generally affected by fluctuations in a transmission path including the digital network. Then, due to the influence of the fluctuation, the position of the specific portion to be subjected to the adaptive operation by the adaptive filter means in the first sound signal changes, that is, the optimum value for the predetermined period described above changes. Therefore, in order to always optimal value of this predetermined period, the predetermined period may be arbitrarily be changed.

Furthermore, the delay time of the first sound signal and the second sound signal transmitted through the transmission path varies depending on the capacity of the transmission path with the counterpart device during communication via the digital network, and accordingly The optimum value of the delay time by the above-mentioned delay means changes, and as a result , the optimum value of a predetermined period that is not subject to the adaptive processing by the adaptive filter means also changes. In addition, the optimum value of the period to be subjected to the adaptive process, that is, the period of the specific portion according to the tap length of the adaptive filter means also changes. Accordingly, parameter setting means for setting various parameters including the delay time by the delay means, the predetermined period, and the period of the specific portion in accordance with the transmission path capacity with the counterpart apparatus in communication may be further provided. The parameter setting means may recognize in advance the transmission path capacity with each of all the counterpart devices, and may set various parameters for the counterpart device in communication based on this recognition.

  In particular, in order to realize echo cancellation, it is sufficient to further include subtracting means for subtracting the simulation signal from the second sound signal.

  As described above, according to the present invention, practical system identification over a digital network is possible. Therefore, by applying the present invention, echo cancellation, which is a kind of system identification, can be realized through a digital network.

It is the schematic which shows the whole structure of the station communication system to which this invention was applied. It is an illustration figure which shows notionally the structure of the part relevant to the echo cancellation in the same embodiment. It is an illustration figure which shows the aspect of the transmission signal in the same embodiment, and a received signal. It is the list which summarized each communication condition in the embodiment. It is the schematic which shows the whole structure of the conventional station communication system.

An embodiment of the present invention will be described with reference to FIGS.

FIG. 1 is an illustrative view showing an overall configuration of a station communication system 10 to which the present invention is applied. As can be seen from FIG. 1, the station communication system 10 according to the present embodiment includes a multi-interface unit 12 as a network device similar to that shown in FIG. 5 , a plurality of portable terminal devices 14, 14,. .., A PBX 18, and a digital network 20. Each IP door terminal device 16 includes a call button 22, a speaker 24 as sound information output means, and a microphone 26 as sound information detection means. However, in this embodiment, each IP door terminal device 16 is not equipped with an echo cancellation function, but the multi-interface unit 12 is equipped with an echo cancellation function. Then, the echo cancellation function of the multi-interface unit 12 performs echo cancellation on all the IP door terminal devices 16, 16,. In other words, echo cancellation through the network 20 is realized from the multi-interface unit 12 to each of the IP door terminal devices 16, 16,.

  In order to realize the echo cancellation over the network 20, the multi-interface unit 12 in the present embodiment includes a transmission buffer 30 as shown in FIG. FIG. 2 shows echo cancellation between a certain IP door terminal device 16 and the multi-interface unit 12 when a call is made between the certain IP door terminal device 16 and a certain portable terminal device 14. It is an illustration figure which shows the part which is related notionally. Although not shown in detail, the transmission buffer 30 constitutes a transmission circuit as transmission means.

The transmission circuit including the transmission buffer 30 encodes the audio signal x as the first sound signal to be transmitted to the other party's IP door terminal device 16 into, for example, a UDP (User Datagram Protocol) packet, so-called packetized. The packetized voice signal x is temporarily stored in the transmission buffer 30 and then transmitted to the counterpart IP door terminal device 16 via the digital network 20. Note that the transmission buffer 30 has a storage capacity of, for example, two packets, and the audio signal x for two packets stored in the transmission buffer 30 is sequentially one packet at a time according to the stored order. Read and send. The multi-interface unit 12 is, for example, an FIR (Finite).
Impulse Response) type adaptive filter 32 is provided, and this adaptive filter 32 is also inputted with the same audio signal x (that is, before being packetized) that is inputted to the transmission circuit including the transmission buffer 30. Is done.

  The counterpart IP door terminal device 16 has an adjustment buffer 50, and the audio signal x transmitted from the multi-interface unit 12 through the digital network 20 to the adjustment buffer 50 sequentially for each packet. Input and memorize. This adjustment buffer 50 is for absorbing the influence of fluctuations in the transmission path including the digital network 20, and has a storage capacity of, for example, 5 packets. Then, the audio signals x for 5 packets stored in the adjustment buffer 50 are sequentially read out one packet at a time in accordance with the stored order, and further input to the reception buffer 52. Specifically, the reception buffer 52 Are input to the receiving circuit. The reception circuit including the reception buffer 52 decodes the audio signal x input thereto into the original state while storing it in the reception buffer 52. The reception buffer 52 has a storage capacity of, for example, 3 packets, and the audio signal x for 3 packets stored in the reception buffer 52 is sequentially one packet at a time according to the stored order. Read and decrypt. The decoded audio signal x is input to the speaker 24. Thereby, the sound as the first sound information corresponding to the audio signal x is output from the speaker 24.

  At this time, as indicated by a broken-line arrow 60 in FIG. 2, an echo component appears in the audio signal y output from the microphone 26 when a part of the output sound of the speaker 24 is input to the microphone 26. Then, the audio signal y including the echo component is input to a transmission buffer 54 provided in the IP door terminal device 16, and specifically input to a transmission circuit including the transmission buffer 54. The transmission circuit including the transmission buffer 54 is the same as that provided in the multi-interface unit 12, and encodes the input audio signal y into a UDP packet, that is, packetizes it. The transmission packet 54 has a storage capacity for two packets. The audio signal y packetized by the transmission circuit including the transmission buffer 54 is sequentially transmitted to the multi-interface unit 12 via the digital network 20 one packet at a time.

  The multi-interface unit 12 has an adjustment buffer 34 as delay means, and the audio signal y transmitted from the IP door terminal device 16 via the digital network 20 to the adjustment buffer 34 one packet at a time. Sequentially input and stored. This adjustment buffer 34 also has a storage capacity of 5 packets, similar to that provided in the IP door terminal device 16, and acts to absorb the influence of fluctuations in the transmission path including the digital network 20. Play. Then, the audio signal y for 5 packets stored in the adjustment buffer 34 is sequentially read out one packet at a time in accordance with the stored order, and further input to the reception buffer 36. Specifically, the reception buffer 36 Is input to a receiving circuit as receiving means.

  The receiving circuit including the receiving buffer 36 decodes the audio signal y input thereto to the original state while storing the audio signal y in the receiving buffer 36 in the same manner as that provided in the IP door terminal device 16. The reception buffer 36 has a storage capacity for three packets, and the audio signal y stored in the reception buffer 36 is sequentially read out and decoded one packet at a time. The decoded audio signal y is input to an adder 38 as subtracting means.

  The adder 38 also receives the signal u after processing by the adaptive filter 32, and the adder 38 subtracts the signal u after processing from the audio signal y. Then, the signal after subtraction by the adder 38, that is, the error signal e indicating the difference between the processed signal u and the audio signal y is sent to the PBX 18 via various circuits (not shown), and eventually the IP door terminal device. It is sent to 16 mobile terminal devices 14 of the other party.

  In addition, the error signal e is fed back to the adaptive filter 32. The adaptive filter 32 appropriately sets its own filter coefficient W so that the error signal e is minimized, in other words, the processed signal u by the adaptive filter 32 is equivalent to the audio signal y (u = y). Update, so-called adaptive operation.

However, the audio signal x input to the adaptive filter 32, that is, the original transmission signal x, and the audio signal y to be subtracted from the processed signal u by the adaptive filter 32, that is, the reception signal y containing an echo component in short. There is a phase difference Td as shown in FIG. This phase difference Td is mainly caused by the delay time due to each of the buffers 30, 50, 52, 54, 34, and 36 described above, that is, corresponds to a total time of 20 packets. Therefore, in order to realize echo cancellation through the digital network 20 from the multi-interface unit 12 to the IP door terminal device 16, it is necessary to consider this delay time (phase difference) Td.

Therefore, the adaptive filter 32 in the present embodiment, among the transmission signals x inputted to themselves, portions up to the point of going back only in the past predetermined period Ti content in consideration of whether delay time Td present time point, The adaptive process is performed only on the part from the time point that has been traced back to the past by the predetermined period T i to the time point that has been traced back by the period Ta corresponding to its own tap length. That is, the portion to be subjected to adaptive processing is narrowed down. Thus, the adaptive filter 32 will have a tap length of the period (= Ti + Ta) content of the sum of the adaptive non-target period Ti and the adaptation period Ta speak to pseudo, or present time point to the pseudo As a result, an equivalent result can be obtained by adaptively processing the transmission signal x for the period that goes back in the past by a period (= Ti + Ta> Td) longer than the delay time Td. Therefore, practical echo cancellation can be performed from the multi-interface unit 12 to the IP door terminal device 16 even through the digital network 20.

Note that, in the station a communication system 10 having such a digital network 20, by the above-described fluctuations, the relative phase changes to the transmitted signal x received signal y. Therefore, the adaptation target period Ta (tap length of the adaptive filter 34) needs to have some allowance in consideration of the phase change due to the fluctuation. Further, before a period corresponding to the delay time Td elapses from a certain time point due to fluctuations ( for example , to the left side in FIG. 3B ), transmission is performed at the certain time point (to the transmission circuit including the transmission buffer 30 ). because when the received signal y comprising an echo component of the transmission signal x input) also appears, for adaptive non-target period Ti, the Ru is set to small (Ti <Td) than the delay time Td is desirable . Specifically, it is as follows.

  That is, the digital network 20 as the main line in the present embodiment is a broadband transmission line such as an optical line, and has a capacity exceeding 128 kbps (kilobits per second), for example. A line connecting the digital network 20 and the multi-interface unit 12 is also a similar broadband transmission line. On the other hand, the branch line connecting the digital network 20 and each IP door terminal device 16 may be a broadband transmission line depending on the installation environment of each IP door terminal device 16, or the capacity may be 64 kbps. Some of them are narrow-band transmission paths of less than or equal to 64 kbps and 128 kbps or less. The multi-interface unit 12 performs packet communication according to the transmission path capacity of the counterpart IP door terminal device 16 (branch line). For example, when the transmission path capacity of the counterpart IP door terminal device 16 is wide band, packet communication is performed in a wide band, and when the transmission path capacity of the counterpart IP door terminal apparatus 16 is medium band, Packet communication is performed in the bandwidth. When the transmission path capacity of the counterpart IP door terminal device 16 is a narrow band, packet communication is performed in the narrow band. Since the transmission path capacity of each IP door terminal device 16 is set (recognized) in advance on the multi-interface unit 12 side, it is possible to automatically switch such transmission path capacity.

Here, for example, if the wideband packet communication is performed, as shown in FIG. 4, the sampling frequency of the transmission signal x and the received signal y is set to 16 kHz. The data amount per packet is set to be equivalent to 8 ms. Therefore, the delay time Td is 160 ms for 20 packets. Of these, the adjustment delay due to the adjustment buffer 50 on the IP door terminal device 16 side and the adjustment buffer 34 on the multi-interface unit 12 side is 80 ms for 10 packets. The application target period Ta is determined by the tap length of the adaptive filter 34 and is, for example, 32 ms. On the other hand, the non-adaptive period Ti is determined in consideration of the application target period Ta, the delay time Td, and the above-described fluctuation, and is set to 134.375 ms, for example, based on experience and experimental results. The non-applicable period Ti can be arbitrarily changed according to various situations including the degree of fluctuation.

  When packet communication is performed in the middle band, the sampling frequency is set to 8 kHz, which is half of that in the wide band, for example. On the other hand, the amount of data per packet is equivalent to 8 ms, which is the same as that in the wideband, and therefore the delay time Td is also 160 ms, which is the same as in the wideband. Of these, the adjustment delay is 80 ms. Furthermore, since the adaptation target period Ta is determined by the tap length of the adaptive filter 34 as described above, the sampling frequency is 8 kHz and is 128 ms. Accordingly, the non-adaptive period Ti is set to 112.5 ms, for example.

  When packet communication is performed in a narrow band, the sampling frequency is set to 8 kHz, which is the same as that in the middle band, for example. On the other hand, the amount of data per packet is equivalent to 32 ms, which is four times that of the middle band. Therefore, the delay time Td is 640 ms for 20 packets, and the adjustment delay is 320 ms for 10 packets. Further, the adaptation target period Ta is 128 ms, which is the same as that in the middle band, and the non-adaptation period Ti is, for example, 576 ms.

  The parameters including these sampling frequencies are merely examples, and are not limited thereto. That is, it is desirable that it is determined appropriately according to the situation.

  As described above, according to the present embodiment, echo cancellation through the network 20 is realized from the multi-interface unit 12 to all the IP door terminal devices 16, 16,. Therefore, it is not necessary to mount an echo cancellation function for each IP door terminal device 16, 16,. Therefore, the overall configuration of the station communication system 10 including the IP door terminal devices 16, 16,... Can be simplified, and the price thereof can be reduced. This becomes more noticeable as the number of IP door terminal devices 16 increases.

  In this embodiment, the storage capacities of the transmission buffer 30, adjustment buffer 34, and reception buffer 36 on the multi-interface unit 12 side are set to 2 packets, 5 packets, and 3 packets, respectively. Not limited to this. These numbers of packets may be appropriately set within a range in which a good call is realized including an appropriate operation of the echo cancellation function (for example, within a range in which excessive delay or the like does not occur). The same applies to the buffers 50, 52 and 54 on the IP door terminal device 16 side. Also, the total number of packets in the transmission side (outward side) buffers 30, 50 and 52 through which the transmission signal x passes, and the total number of packets in the reception side (return side) buffers 54, 34 and 36 through which the reception signal y passes. The numbers are the same as each other, but may be different. However, considering the communication balance between transmission and reception, it is desirable that both are the same. Further, as a result, the adjustment buffer 50 on the IP door terminal device 16 side acts as a delay means similar to the adjustment buffer 34 on the multi-interface unit 12 side, but another adjustment purpose that acts as this delay means. A buffer may be provided on the transmission side of the multi-interface unit 12. In any case, it is important to set each condition appropriately so as to realize a good call including proper operation of the echo cancellation function.

  In the present embodiment, the case where the present invention is applied to the multi-interface unit 12 constituting the station communication system 10 has been described as an example, but the present invention is not limited thereto. That is, the multi-interface unit 12 can be applied to uses other than the station communication system 10, and the present invention can be applied to apparatuses other than the multi-interface unit 12. That is, the present invention can be applied to any device in which sound information output means and sound information detection means provided in the same space are connected via a digital network. The digital network is not limited to wired communication but may be wireless.

  Further, since the echo cancellation exemplified in the present embodiment is a kind of system identification, the present invention is not limited to the echo cancellation, and the present invention is applied to simply identify the transfer function of the space on the IP door terminal device 16 side, for example. May be. If this is applied, a transfer function of an arbitrary space such as a concert hall in a remote place can be identified through a digital network, and an equalizing operation of the arbitrary space can be performed through the digital network.

DESCRIPTION OF SYMBOLS 10 Station communication system 12 Multi-interface unit 16 IP door terminal device 20 Digital network 24 Speaker 26 Microphone 32 Adaptive filter 34 Adjustment buffer

Claims (5)

And sound information output means for outputting a first sound information corresponding to the first sound signal, second detecting the second sound information present in the provided by our Ri該 space in the same space as the sound information output means A plurality of counterpart devices each having sound information detecting means for generating a second sound signal according to sound information, are network devices connected via a digital network,
Transmitting means for transmitting the first sound signal to the sound information output means of the counterpart device in communication via the digital network;
Receiving means for receiving the second sound signal from the sound information detecting means of the counterpart device in communication via the digital network;
Based on the first sound signal, a simulated signal simulating an echo component that is a wraparound component of the first sound information to the sound information detecting means in the counterpart device in communication with the first sound signal is input . Adaptive filter means to generate ,
Delay means for delaying one or both of the second sound signal received from the counterpart device in the communication by said first sound signal and said receiving means is transmitted to the opposite device in the communication from the transmission means,
Comprising
The adaptive filter means excludes the portion of the first sound signal input to itself from the present time to a time point that has been traced back by a predetermined period considering the delay time by the delay means. , based on the specific parts up to the point of going back to the past by period further in accordance with the tap length of itself from the timing predated only past the predetermined period component performs the adaptation process,
According to the delay time, the predetermined period, and the tap length according to the capacity of the transmission path of the first sound signal and the second sound signal with the counterpart device in communication via the digital network A parameter setting means for setting a parameter including a period;
Network device. The parameter setting means recognizes in advance the capacity of the transmission path between each of the plurality of counterpart devices, and sets the parameters for the counterpart device in communication based on this recognition.
The network device according to claim 1. The delay means delays both the first sound signal and the second sound signal over the same time;
The network device according to claim 1 or 2 . The predetermined period can be arbitrarily changed.
The network device according to claim 1. Subtracting means for subtracting the simulation signal from the second sound signal;
The network device according to claim 1.

Images