JP5572338B2 - Multipoint connection device, multipoint connection method - Google Patents

Description

  The present invention mixes audio signals at a plurality of points when transmitting / receiving digitalized audio signals such as voice and music (hereinafter referred to as audio signals in this specification) via a digital communication network. In addition, the present invention relates to a multipoint connection technology for delivering mixed signals to each point.

  The use (for example, a voice conference) of transmitting and receiving voice signals between a plurality of three or more points via a digital communication network such as a packet communication network is increasing. In order to build an audio conference system, a method is known in which a server device called a multipoint connection device is arranged on a communication network, audio signals sent from a plurality of points are mixed, and distributed to each point. Yes.

  FIG. 1 is a conceptual diagram showing that a multipoint connection device (MCU) 100 serves as a core to realize mutual communication between points when an audio conference is performed at three points. FIG. 2 shows a configuration example of the multipoint mixing unit 110 included in the multipoint connection apparatus 100. Each voice code sent from each of the points A, B, and C is decoded by a decoder 112a, 112b, and 112c, for example, into a digital voice signal in the PCM format. The PCM mixing unit 111 uses each decoded digital audio signal to create a mixed signal for each point. For example, the mixed signal for point A is a signal obtained by mixing points B and C, and the mixed signal for point B is a signal obtained by mixing points A and C. The reason why the local signal is not included in the mixed signal is to prevent the voice from returning as an echo. The mixed signal for each point is encoded by the encoders 113a, 113b, and 113c and distributed to each point.

  In such a multipoint connection apparatus 100, when many arithmetic processes are required for the decoders 112a, 112b, and 112c and the encoders 113a, 113b, and 113c, it becomes a problem that a large load is applied to the multipoint connection apparatus 100. For example, when ITU-T (International Telecommunication Union-Telecommunication Standardization Sector) G.711 (see Non-Patent Document 1) is used as a speech encoding method, decoders 112a, 112b, 112c, encoders 113a, 113b, 113c. In both cases, since the required arithmetic processing is small, the multipoint connection apparatus 100 is not subjected to a great load. However, for example, when ITU-T G.722 (see Non-Patent Document 2), which is a wideband speech coding system, is used as the speech coding system, it is more than when using G.711. As a result, arithmetic processing is required, which leads to a problem that the number of points that can be processed by one multipoint connection device 100 is greatly reduced.

  To deal with such problems, an embedded coding system (also called a scalable coding system) having G.711 as a core is used as a speech coding system, and selective mixing (partial mixing) is performed as processing of the multipoint connection apparatus 100. A method that uses mixing is also proposed. Details of this processing are described in Non-Patent Document 3 and Non-Patent Document 4.

  FIG. 3 is a functional configuration example included in the multipoint connection device 100 by selective mixing described in Non-Patent Document 3. The PCM mixing unit 111 is the same as the PCM mixing unit 111 shown in FIG. Voice packets sent from each point are separated into G.711 bits (first voice code), extension bits (second voice code), and control information by demultiplexing units (deMUX) 114a, 114b, and 114c, respectively. Is done. The G.711 bits corresponding to each point are decoded by the G.711 decoders 112a, 112b, and 112c, mixed by the PCM mixing unit 111, and then encoded by the G.711 encoders 113a, 113b, and 113c. A G.711 bit for the point is created.

  The extension bits separated from the voice packet at each point are input to the extension bit switching unit 117, respectively. The control information separated from the voice packet at each point is input to the point selection unit 116. The point selection unit 116 determines a main utterance point and a subordinate utterance point from time to time using the control information of each point, and outputs a control signal for controlling the extension bit switching unit 117. If the main utterance point at a certain point in time is A and the subordinate utterance point is B, the extension bit switching unit 117 sets the extension bit of the point A for the points B and C, and the point B for the point A. Operates to output extension bits. The G.711 bit and the extension bit for each point are respectively combined by multiplexing units (MUX) 115a, 115b, and 115c, and a voice packet for each point is transmitted to each point.

  In the configuration shown in FIG. 3, the G.711 encoding method, the G.711 decoders 112a, 112b, and 112c, and the G.711 encoders 113a, 113b, and 113c are other encoding methods and the corresponding encoders and decoders. However, in general, an encoding method with a small processing amount for encoding and decoding is desirable. Further, the extension bits correspond to speech codes for a low-frequency enhancement layer and a high-frequency enhancement layer excluding the G.711 core layer in the case of the G.711.1 speech coding scheme. The control information is multiplexed into the voice code and sent independently, as well as the G.711 bit, the PCM voice signal output from the G.711 decoder, or the extension bit. Information created by In general, as control information, power of a voice signal (including PCM voice output from a G.711 decoder), voice / non-voice section information (also referred to as VAD), voiced / unvoiced voice identification information, and the like are used.

ITU-T Recommendation G.711, "Pulse Code Modulation (PCM) of voice frequencies," ITU-T, 1988. ITU-T Recommendation G.722, "7kHz Audio-Coding within 64kbit / s," ITU-T, 1988. “A G.711 Embedded Wideband Speech Coding for VoIP Conferences”, IEICE TRANS. INF. & SYST., VOL.E89-D, No 9, 2006. "International Standard ITU-T G.711.1 for Wideband Speech Coding", NTT Technical Journal, Vol.20 No. 5, 2008.

  As shown in FIG. 3, in the case where only a part of voice codes included in a voice packet is mixed as usual, and other voice codes included in the voice packet are processed as selection formulas that are switched from moment to moment without being mixed. There is a possibility that the quality of the mixed sound reproduced at each point may deteriorate due to the processing algorithms of the point selection unit and the extension bit switching unit.

  Further, Non-Patent Document 3 exemplifies an algorithm that uses the VAD information to suppress switching of utterance points at a cycle of 200 ms or less. While this algorithm has the merit of not causing annoying noise due to frequent switching, the voice at the main speaking point is temporarily heard as a narrowband sound of 3.4 kHz band at the switching point of the speaking point, 7 kHz There was a problem that the feeling of switching between wideband audio in the band and narrowband audio in the 3.4 kHz band was conspicuous.

  Accordingly, an object of the present invention is to provide a multipoint connection technique with little deterioration in sound quality.

  The present invention divides voice packets sent from three or more points into time units (subframe length units) shorter than the packetization period, and outputs divided voice packets [time direction division processing]. Then, at least the first voice code and the second voice code are extracted from each of the divided voice packets corresponding to each point [demultiplexing processing], and the first voice code corresponding to each point is decoded to obtain the first voice signal. Output [decoding process], mix the first audio signal corresponding to each point to output a mixed audio signal for each point [mixing process], and encode and mix the mixed audio signal corresponding to each point A voice code is output [encoding process], an utterance point is determined from each point, and a control signal corresponding to the utterance point is output [point selection process] For each point determined according to the control signal, a second speech code determined according to the control signal among the second speech codes corresponding to each point is output [second speech code switching process], and The corresponding mixed speech code and the second speech code corresponding to each point output in the second speech code switching process are combined to output a unit speech packet in subframe length units [multiplexing process]. A plurality of unit voice packets corresponding to the points are combined to output a transmission voice packet having a time unit of a packetization period [time direction combining process].

  Or a point selection process respond | corresponds to each point with the acoustic attribute determination process which calculates | requires the acoustic attribute of the 1st audio | voice code corresponding to each point, the process which memorize | stores the acoustic attribute of the 1st audio | voice code corresponding to each point, and Based on the current acoustic attribute of the first speech code and the past acoustic attribute of the first speech code corresponding to each location, the speech location determination process for determining the current speech location from each location is determined. You may make it perform the control signal output process which outputs the control signal corresponding to an utterance point.

  Alternatively, in the point selection process, based on the acoustic attribute determination process for obtaining the acoustic attribute of the first speech code corresponding to each point and the acoustic attribute of the first speech code corresponding to each point, the current utterance from each point You may be made to perform the utterance point determination process which determines a point, the process which memorize | stores the information showing an utterance point, and the control signal output process which outputs the control signal corresponding to the determined utterance point.

  According to the present invention, processing is performed in a unit of time shorter than the packetization period, and then the voice packet having the packetization period is recovered, so that no annoying noise is generated and wideband and narrowband voices are not generated. It is possible to achieve both inconspicuous switching, and it is possible to realize a multipoint connection technology with a small amount of calculation and little deterioration in sound quality.

The conceptual diagram which shows that a multipoint connection apparatus (MCU) becomes a core and implement | achieves the mutual communication between each point, when performing a voice conference at three points. The figure which shows the function structural example of the multipoint mixing part contained in the conventional multipoint connection apparatus. The figure which shows the function structural example contained in the multipoint connection apparatus by the conventional selection mixing. The block diagram which shows the function structural example of the multipoint connection apparatus of 1st Embodiment. The figure which shows the processing flow in the multipoint connection apparatus of 1st Embodiment. The block diagram which shows the function structural example of the multipoint connection apparatus of 2nd Embodiment. The figure which shows the processing flow in the multipoint connection apparatus of 2nd Embodiment. The block diagram which shows the function structural example of the structural example 1 of the control information calculation and the spot selection part 116p. The figure which shows the processing flow in the structural example 1 of the control information calculation and the spot selection part 116p. The block diagram which shows the function structural example of the structural example 2 of the control information calculation and the spot selection part 116p. The figure which shows the processing flow in the structural example 2 of the control information calculation and the spot selection part 116p. The block diagram which shows the function structural example of the structural example 3 of the control information calculation and the spot selection part 116p. The figure which shows the processing flow in the structural example 3 of the control information calculation and the spot selection part 116p. The block diagram which shows the function structural example of the structural example 4 of the control information calculation and the spot selection part 116p. The figure which shows the processing flow in the structural example 4 of the control information calculation and the spot selection part 116p.

[First Embodiment]
For example, when voice packets are transmitted / received via a packet communication network, communication is performed by determining a time length of a voice signal corresponding to a voice code to be put in one voice packet, called a packetization period. In general, the packetization period is often 10 milliseconds or 20 milliseconds. On the other hand, each audio coding method has a time length called a frame as a minimum time unit for performing encoding / decoding processing. For example, the frame length of G.711.1 is 5 milliseconds. In addition, there is a speech coding scheme in which the frame length is 20 milliseconds, but a part of the internal information can be divided into shorter time units, for example, 5 milliseconds. A time unit shorter than such a frame is called a subframe.

  In FIG. 4, the structural example of the multipoint connection apparatus 200 of 1st Embodiment by this invention is shown. The multipoint connection device 200 has a configuration in which time direction division units 120a, 120b, 120c and time direction coupling units 121a, 121b, 121c are added to the multipoint connection device 100 shown in FIG. FIG. 5 shows a processing flow of the multipoint connection apparatus 200 of the first embodiment.

  In this multipoint connection apparatus 200, first, the time direction division units 120a, 120b, and 120c each of the voice packets received from each point are as short as possible in time units (hereinafter referred to as subframe length units). .) And output a divided voice packet (step S1). However, the processing target divided by the time direction dividing units 120a, 120b, and 120c is a payload of a voice packet (a group of data related to voice information). Specifically, the time direction division unit 120a divides the voice packet received from the point A into subframe length units shorter than the packetization period as much as possible, and outputs the divided voice packet. The time direction division unit 120b performs the same process as the time direction division unit 120a on the voice packet received from the point B, and the time direction division unit 120c performs the same process on the voice packet received from the point C. The input and output of the time direction division units 120a, 120b, and 120c do not need to be in a narrowly-defined packet format (including header information) for transmission / reception via the communication network. It may be a delimited code string.

  Then, the demultiplexing units (deMUX) 114a, 114b, and 114c convert the divided speech packets in units of subframe length corresponding to each point into G.711 bits (first speech code) and extension bits (second speech code), respectively. ) And separated into control information (step S2). Specifically, the demultiplexing unit 114a separates the divided voice packets in subframe length units corresponding to the point A into G.711 bits, extension bits, and control information, respectively. The demultiplexing unit 114b is for each sub-frame length unit divided speech packet corresponding to the B point, and the demultiplexing unit 114c is for each sub-frame length unit divided speech packet corresponding to the C point. The same processing as that of the demultiplexing unit 114a is performed.

  The G.711 bits corresponding to each point are decoded by the G.711 decoders 112a, 112b, and 112c (step S3) and mixed by the PCM mixing unit 111 (see FIG. 2) (step S4). Encoding is performed by the 711 encoders 113a, 113b, and 113c, and G.711 bits for each point are created (step S5).

  In other words, the G.711 decoders 112a, 112b, and 112c respectively convert the G.711 bits in subframe length units sent from the demultiplexing units 114a, 114b, and 114c, for example, digital audio signals in the PCM format (first Audio signal). Specifically, the G.711 decoder 112a decodes each G.711 bit in the subframe length unit sent from the demultiplexing unit 114a into a digital audio signal in the PCM format. For each G.711 bit in the subframe length unit sent from the demultiplexing unit 114b, the G.711 decoder 112c sends a subframe sent from the demultiplexing unit 114c. The same processing as the G.711 decoder 112a is performed on each G.711 bit in the long unit.

  Then, the PCM mixing unit 111 generates a mixed audio signal for each point using each decoded digital audio signal in subframe length units. In this example, the mixed audio signal for point A is a mixture of the decoded digital audio signals at points B and C, and the mixed audio signal for point B is decoded at points A and C. The mixed audio signal for point C is a mixture of the decoded digital audio signals at points A and B. The reason why the local signal is not included in the mixed signal is to prevent the voice from returning as an echo as described above. Note that there is no limitation on the method of creating the mixed audio signal for each point. For example, when the PCM mixing unit 111 has the configuration shown in FIG. 2, the decoded digital audio signals at point B and point C are mixed ( The mixed audio signal for point A) is created by subtracting the decoded digital audio signal at point A from the mixture of the decoded digital audio signals at point A, point B, and point C. Similarly, the decoded digital audio signals at points A and C (mixed audio signals for point B) are mixed with the decoded digital audio signals at points A, B, and C. A signal obtained by subtracting the decoded digital audio signal of point B from the point and a mixture of the decoded digital audio signals of point A and point B (mixed audio signal for point C) It is created by subtracting the decoded digital audio signal at point C from the mixture of each decoded digital audio signal at point C. In addition, the present invention is not limited to the creation method for creating the mixed audio signal for the destination point by subtracting the decoded digital audio signal at the destination point from the total mixed signal obtained by mixing the decoded digital audio signals at all points. Alternatively, each decoded digital audio signal at each point other than the destination point may be mixed to create a mixed audio signal for the destination point.

  Next, the G.711 encoders 113a, 113b, and 113c perform G.711 encoding of the mixed audio signal in units of subframe lengths for each point, and output a mixed audio code. Specifically, the G.711 encoder 113a performs G.711 encoding on each mixed speech signal for the point A in subframe length units and outputs G.711 bits (mixed speech code). The G.711 encoder 113b uses G.711 encoder 113c for each mixed audio signal for point B in subframe length units, and G.711 encoder 113c uses G.711 encoder 113c for each mixed audio signal for point C in subframe length units. The same processing as the 711 encoder 113a is performed.

  Control information corresponding to each point is input to the point selection unit 116. The point selection unit 116 determines the main utterance point and the subordinate utterance point from time to time using the control information of each point, and outputs a control signal for controlling the extension bit switching unit 117 (step S6). The control signal is a signal generated corresponding to the utterance point. As described above, the extension bits correspond to speech codes for the low-frequency enhancement layer and the high-frequency enhancement layer excluding the G.711 core layer in the case of the G.711.1 speech coding scheme. In addition to the information that is multiplexed and transmitted independently in divided audio packets in subframe length units, the control information uses G.711 bits, PCM audio signals that are output from the G.711 decoder, or extension bits. Information created in the multipoint mixing unit may be used. For example, as control information, power of a voice signal (including PCM voice that is output from a G.711 decoder), voice / non-voice section information (VAD), voiced / unvoiced voice identification information, and the like are used.

  Based on the control signal from the point selection unit 116, the extension bit switching unit 117 selects the extension bit corresponding to the output point from the extension bits corresponding to each point input to the extension bit switching unit 117, and multiplexing The extension bits selected for the multiplexing unit corresponding to the output point among the units (MUX) 115a, 115b, 115c are output (step S7). This process is performed corresponding to the output point. That is, the extension bit switching unit 117 outputs the extension bit determined according to the control signal among the extension bits corresponding to each point, for the point determined according to the control signal among the points.

For example, when the output point is the main utterance point, the extension bit switching unit 117 selects the extension bit of the subordinate utterance point and outputs the selected extension bit to the multiplexing unit corresponding to the main utterance point. When the output point is an utterance point other than the main utterance point, the extension bit switching unit 117 selects the extension bit of the main utterance point and outputs the selected extension bit to the multiplexing unit corresponding to the output point.
As a specific example, assuming that the main utterance point at a certain point is A and the subordinate utterance point is B, the first switching control unit 117a of the extension bit switching unit 117 is based on the control signal from the point selection unit 116. The extension bit at point B is selected as the extension bit corresponding to A, and the second switching control unit 117b of the extension bit switching unit 117 outputs the selected extension bit to the multiplexing unit corresponding to the output point A. Further, based on the control signal from the point selection unit 116, the first switching control unit 117a of the extension bit switching unit 117 selects the extension bit of the point A as the extension bit corresponding to the output points B and C, and the extension bit The second switching control unit 117b of the switching unit 117 outputs the selected extension bit to each multiplexing unit corresponding to the output points B and C.

  The multiplexing units (MUX) 115a, 115b, and 115c combine the G.711 bit (mixed speech code) and the extension bit (second speech code) for the corresponding points, respectively, and unit the subframe length. A voice packet is output (step S8). That is, the multiplexing unit 115a outputs the G.711 bit in the subframe length unit for the point A output from the G.711 encoder 113a and the extension bit output from the extension bit switching unit 117 (in this example, the corresponding utterance point B). A unit voice packet in units of subframe length is output by combining the extension bits. Similarly, the multiplexing unit 115b combines the G.711 bit for the point B output from the G.711 encoder 113b and the extension bit output from the extension bit switching unit 117 (in this example, the extension bit of the main speech point A). Then, the unit voice packet in units of subframe length is output, and the multiplexing unit 115c outputs the G.711 bit for the point C output from the G.711 encoder 113c and the extension bit output from the extension bit switching unit 117 (this In the example, the extension bit of the main speech point A) is combined and a unit voice packet in units of subframe length is output. The unit voice packet in units of subframe length is a group (payload) of data related to voice information and does not need to be in a narrowly defined packet format for transmission / reception via a communication network.

  Next, the time direction combining units 121a, 121b, and 121c combine a plurality of unit voice packets in units of subframe lengths output from the multiplexing units 115a, 115b, and 115c in the time direction, and set the time unit of the packetization period. The transmitted voice packet is output toward each output point (step S9). That is, the time direction combining unit 121a combines a plurality of unit voice packets in subframe length units output from the multiplexing unit 115a in the time direction, and outputs a transmission voice packet having a time unit of a packetization period as an output point. Output to A. Similarly, the time direction combining unit 121b combines a plurality of unit voice packets in units of subframe length output from the multiplexing unit 115b in the time direction, and outputs a transmission voice packet having a time unit of a packetization period. The time direction combiner 121c outputs the signal toward the point B, combines the plurality of unit voice packets in units of subframe length output from the multiplexing unit 115c in the time direction, and transmits the unit having a packetization period time unit. The trusted voice packet is output toward the output point C. Note that the process of combining unit voice packets in the time direction is not limited to a combining method in which unit voice packets are simply arranged in time order, and codes (layers) may be exchanged within unit voice packets or between unit voice packets. For example, by combining the G.711 bits included in each unit voice packet in the time direction and the concatenation of extension bits included in each unit voice packet in the time direction, the time unit of the packetization period May be generated.

[Second Embodiment]
In FIG. 6, the structural example of the multipoint connection apparatus 300 of 2nd Embodiment by this invention is shown. FIG. 7 shows a processing flow of the multipoint connection apparatus 300 of the second embodiment. Unlike the multipoint connection device 200 shown in FIG. 4, the multipoint connection device 300 includes a control information calculation / point selection unit 116 p instead of the point selection unit 116. Since the processing and procedure of each function other than the demultiplexing units 114a, 114b, and 114c and the control information calculation / point selection unit 116p are the same as those in the first embodiment, this is used instead of the redundant description. Hereinafter, the process of the control information calculation / point selection unit 116p will be described as a subject.

  When the time direction division units 120a, 120b, and 120c divide the voice packet into subframe length units, the control information for the point selection unit 116 included in the divided voice packet does not correspond to control in subframe length units. There is. That is, in the first embodiment, the demultiplexing units (deMUX) 114a, 114b, and 114c separate the divided voice packets in units of subframe length corresponding to each point into G.711 bits, extension bits, and control information, respectively. However, there are cases where control information in units of subframe length cannot be obtained (step S2a). The second embodiment is a form corresponding to such a situation, and the control information calculation / point selection unit 116p uses the G.711 bit to determine an utterance point and generates a control signal (step S6a).

Each functional configuration example of the control information calculation / point selection unit 116p and its processing flow are shown in FIGS.
<Configuration example 1>
FIG. 8 shows a configuration example 1 of the control information calculation / point selection unit 116p. FIG. 9 shows a processing flow of Configuration Example 1 of the control information calculation / point selection unit 116p. The power calculators 1161a, 1161b, and 1161c calculate power in subframe length units from the G.711 bits of each point input in subframe length units (step S6a11). The power can be obtained by the sum of squares of the PCM signal obtained by decoding the G.711 bit by the G.711 decoder. Alternatively, as an alternative value of power, an average of the absolute values of the PCM signal or an average value of codes for each sample obtained by removing the positive / negative sign from G.711 bits may be used (hereinafter collectively referred to as power). ). The utterance point determination unit 1162 compares the power of each point, determines the point with the highest power as the main utterance point, and the point with the second highest power as the subordinate utterance point, and outputs information representing these ( Step S6a12). The control signal output unit 1163 outputs a control signal for controlling the extension bit switching unit 117 using the information of the main utterance point and the subordinate utterance point given from the utterance point determination unit 1162 (step S6a13).

<Configuration example 2>
FIG. 10 shows a configuration example 2 of the control information calculation / point selection unit 116p. FIG. 11 shows a processing flow of the configuration example 2 of the control information calculation / point selection unit 116p.
Configuration example 2 includes memories 1164a, 1164b, and 1164c corresponding to each point in addition to the functional configuration of configuration example 1 shown in FIG. Since the processing of each function other than the speech point determination unit 1162 and the memories 1164a, 1164b, and 1164c is the same as that in the configuration example 1, this is used instead of the duplicate description.

  The power at each point calculated for each subframe by the power calculators 1161a, 1161b, and 1161c is temporarily stored in the memories 1164a, 1164b, and 1164c (step S6a2). The utterance point determination unit 1162 uses the power of the current subframe corresponding to each point as well as the time series of the power of the subframes corresponding to each point stored in the memories 1164a, 1164b, and 1164c. The point and the subordinate utterance point are determined, and information representing them is output (step S6a22). For example, when the difference between the points for the power value of the current subframe corresponding to each point is smaller than the threshold value, the magnitude relationship between the points for the power value corresponding to each point one subframe before this Considering the size relationship, the main utterance point and the subordinate utterance point are determined. Or, when the magnitude relationship between the points for the power value corresponding to each point before one subframe is more significant than the magnitude relationship between the points for the power value of the current subframe corresponding to each point The main utterance point and the subordinate utterance point may be determined based on the magnitude relationship between the points for the power value corresponding to each point before one subframe. Alternatively, the main utterance point and the subordinate utterance point are determined by comparing the power value of the current subframe corresponding to each point and the average of the power value corresponding to each point before one subframe between the points. May be.

<Configuration example 3>
FIG. 12 shows a configuration example 3 of the control information calculation / point selection unit 116p. FIG. 13 shows a processing flow of configuration example 3 of the control information calculation / point selection unit 116p.
This configuration example 3 includes memories 1165a and 1165b corresponding to the utterance points for which the master-slave is determined, in addition to the functional configuration of the configuration example 1 shown in FIG. Since the processing of each function other than the utterance point determination unit 1162 and the memories 1165a and 1165b is the same as that of the configuration example 1, this is used instead of the duplicate description.

  In the configuration example 3, information regarding each master / slave utterance point determined by the utterance point determination unit 1162 is temporarily stored in the memories 1165a and 1165b (step S6a3). The utterance point determination unit 1162 changes the determination criterion for determining each utterance point of the master-slave according to the information about each utterance point of the master-slave one subframe before acquired from the memories 1165a, 1165b, and The main utterance point and the subordinate utterance point are determined using the power of the current subframe corresponding to each point, and information representing these is output (step S6a32).

  In the process of determining the main and subordinate utterance points, the main utterance point is determined first and then the subordinate utterance is determined. First, the utterance point determination unit 1162 first acquires the main utterance one subframe before acquired from the memory 1165a. The decision criteria for determining the main utterance point are changed according to the information about the point, and the main utterance point is determined using the power of the current subframe corresponding to each point, and this is expressed. Output information.

  For example, if the main utterance point one subframe before is the point A and the power of the point A is the maximum among the powers of the respective points in the current subframe, the main utterance point is continuously set as the point A. If the main utterance point one subframe before is point A, and the power at point B is the maximum among the powers at each point in the current subframe, the configuration example shown in FIG. 1, for example, when the power at the point B is greater than α times the power at the point A for the current subframe, the main utterance point is changed to the point B, but even if the power at the point B is larger, the point A If the power is less than α times the power, the processing is performed such that the main utterance point remains unchanged at point A (where α is a positive real number and is usually a positive number larger than 1).

  This process means that continuation of the main utterance point is easy and a high hurdle is provided for changing the main utterance point, and there is an effect of preventing the utterance point from being switched more frequently than necessary. It is also possible to employ a rule that raises the hurdle for switching the main utterance point to another point as the subframe determined as a main utterance point continues for a longer time. For example, when the same point is set as a main utterance point in four consecutive subframes, it is allowed to impose a restriction that a power difference of β times or more is required in the next subframe for switching the utterance point. This β is a positive real number set independently of α, and is usually a value larger than α.

  Further, following the process of determining the main utterance point described above, the utterance point determination unit 1162 determines the subordinate utterance point according to the information about the subordinate utterance point one subframe before acquired from the memory 1165b. In accordance with this criterion, the sub utterance point is determined using the power of the current subframe corresponding to each point, and information representing this is output.

  For example, << 1 >> In the above-described process for determining the main utterance point, when the subordinate utterance point B one subframe before is not promoted to the main utterance point, the following process is performed. If the secondary utterance point one subframe before is the point B and the power of the point B among the power of each point for the current subframe is the highest except for the main utterance point determined by the above-described processing, it will continue to follow. Let the utterance point be point B. If the sub utterance point one subframe before is the point B and the power of the point C is the maximum except for the main utterance point determined in the above process among the powers of each point for the current sub frame, simply Unlike the configuration example 1 shown in FIG. 8 in which the subordinate utterance point is the point C, for example, when the power of the point C is larger than the power of the point B by γ times or more in the current subframe, the subordinate utterance point is set to the point C. However, if the power at point C is larger than γ times the power at point B, the subordinate utterance point is not changed as it is at point B (where γ is a positive real number). And is usually a positive number greater than 1.)

  << 2 >> In the above-described process of determining the main utterance point, when the subordinate utterance point B one subframe before is promoted to the main utterance point, the following process is performed. The main utterance point A one subframe before is regarded as a sub utterance point one subframe before, and the same processing as << 1 >> is performed in the above case. That is, when the subordinate utterance point one subframe before is the point A and the power of the point A among the powers of the respective points in the current subframe is the maximum except for the main utterance point determined by the above-described processing. The subordinate utterance point is designated as point A. If the sub utterance point one subframe before is the point A, and the power of the point C is the maximum except for the main utterance point determined in the above process among the powers of the respective points in the current subframe, the simple Unlike the configuration example 1 shown in FIG. 8 in which the utterance point conforming to is the point C, for example, if the power at the point C is more than γ times greater than the power at the point A for the current subframe, the utterance point according to Change to C, but if the power at point C is greater than γ times the power at point A, a process is performed in which the subordinate utterance point is point A (where γ is a positive real number) Yes, usually a positive number greater than 1.)

  In the case of << 2 >>, the main utterance point A one subframe before is regarded as a sub utterance point one subframe before, and the following processing is performed to determine the sub utterance point subordinate to the current subframe. You may perform a process like this. That is, the content stored in the memory 1165b is cleared, and the utterance point followed by the current subframe is determined using the power of the current subframe in the same manner as the processing described in the configuration example 1.

  In this way, the process for determining the subordinate utterance point is also necessary for the process of determining the subordinate utterance point by making it easy to continue the subordinate utterance point and providing a high hurdle for changing the subordinate utterance point. As described above, the utterance point may be prevented from being frequently switched. It is also possible to employ a rule that increases the hurdle for switching the utterance point according to another point as the subframe determined as the utterance point followed by one point continues for a longer time. For example, in the case where the same spot is followed by the same spot in four consecutive subframes, it is allowed to impose a restriction that a power difference of θ times or more is required in the next subframe for switching the talk spot. This θ is a positive real number that is set independently of γ, and is usually larger than γ.

  In the configuration example 3, it is not essential to change the determination criteria for determining the subordinate utterance point. In this case, the memory 1165b is unnecessary (that is, the subordinate utterance point conforms to the configuration example 1). Determined.)

<Configuration example 4>
FIG. 14 shows a configuration example 4 of the control information calculation / point selection unit 116p. FIG. 15 shows a process flow of the configuration example 4 of the control information calculation / point selection unit 116p.
Configuration example 4 is a composite form of the functional configuration of configuration example 2 shown in FIG. 10 and the functional configuration of configuration example 3 shown in FIG. Since the processing of each function has already been clarified by the description of the configuration example 2 and the configuration example 3, this is used instead of the duplicate description.

  In the configuration example 4, the utterance point determination unit 1162 includes a time series of the power of the current subframe corresponding to each point and the power of each subframe corresponding to each point stored in the memories 1164a, 1164b, and 1164c. The master / slave utterance points in the current subframe are determined using the information about the master / slave utterance points stored in the memories 1165a and 1165b, and information representing these is output (step S6a4). As the determination process, first, the determination process according to the configuration example 2 is performed, then the determination process according to the configuration example 3 is performed, or the determination process according to the configuration example 3 is performed first, and then the determination process according to the configuration example 2 is performed. It is possible to implement a determination process in which the determination algorithm according to the configuration example 2 and the determination algorithm according to the configuration example 3 are merged. For example, when the main utterance point one subframe before is the point A, a determination process is performed in which the main utterance point is switched to the point B when the power of the point B reaches the maximum for the two immediately preceding subframes. .

  In the configuration example 4, it is not essential to change the determination criterion for determining the subordinate utterance point. In this case, the memory 1165b is unnecessary (that is, the subordinate utterance point is the configuration example 1 or the configuration example). 2).

  In each of the above-described embodiments, G.711.1 is adopted as the speech encoding method, but the present invention is not limited to this, and other speech encoding methods may be used. In the second embodiment, power as control information is calculated using G.711 bits, and an utterance point is determined using this power. However, the present invention is not limited to such processing. As the control information, acoustic attributes such as the power of voice signals (including PCM voice output from the G.711 decoder), voice / non-voice section information (VAD), voiced / unvoiced voice identification information, etc. It is good also as a process calculated | required from the output PCM audio | voice signal and an extension bit.

  For example, if the multipoint connection device is realized by a computer, the multipoint connection device according to the embodiment includes an input unit to which a keyboard or the like can be connected, an output unit to which a liquid crystal display or the like can be connected [these are simple relays. This is not always necessary when a multipoint connection apparatus is realized as a base station. ] A communication unit that can be connected to a communication device (for example, a modem) that can communicate with the outside of the multipoint connection device, DSP [CPU may be used. A cache memory or the like may be provided. ] RAM (Random Access Memory), ROM (Read Only Memory), and external storage devices such as hard disks, as well as the input unit, output unit, communication unit, DSP, RAM, ROM, data between the external storage devices It is equipped with a bus that is connected so that it can be exchanged. If necessary, the multipoint connection device may be provided with a device (drive) that can read and write a storage medium such as a CD-ROM.

  The external storage device of the multipoint connection device stores a program for multipoint connection and data necessary for processing of the program [for example, the program is stored in a ROM which is a read-only storage device. Form is also allowed. ]. Further, data obtained by the processing of these programs is appropriately stored in a RAM or the like.

  Specifically, in the external storage device, a program for dividing a voice packet transmitted from each of three or more points into subframe length units shorter than the packetization period and outputting a divided voice packet A program for extracting at least the first voice code and the second voice code from each of the divided voice packets corresponding to each point, and for decoding the first voice code corresponding to each point and outputting the first voice signal A program for mixing a first audio signal corresponding to each point and outputting a mixed audio signal for each point, and for encoding a mixed audio signal corresponding to each point and outputting a mixed audio code Program, a program for determining the utterance point from each point, and outputting a control signal corresponding to the utterance point, out of each point A program for outputting a second speech code determined according to a control signal among second speech codes corresponding to each point, a mixed speech code corresponding to each point, and a second code for a point determined according to the control signal A program for combining the second speech code corresponding to each point output in the speech code switching process to output a unit speech packet in units of subframe length, and combining a plurality of unit speech packets corresponding to each point A program for outputting a voice packet for transmission having a time unit of a packetization period is stored. In addition, a control program for controlling processing based on these programs is also stored as appropriate.

  In the multipoint connection device according to the embodiment, each program stored in an external storage device (or ROM, etc.) and data necessary for processing each program are read into the RAM as necessary, and interpreted and executed by the DSP. It is processed. As a result, the DSP has predetermined functions (time direction division unit, demultiplexing unit, decoder, mixing unit, encoder, point selection unit, second speech code switching unit (extension bit switching unit), multiplexing unit, time direction combination. Multi-point connection is realized.

  In addition, the multipoint connection apparatus / method according to the present invention is not limited to the above-described embodiment, and can be appropriately changed without departing from the gist of the present invention. In addition, the processing described in the above multipoint connection apparatus / method is not performed in chronological order according to the description order, but is performed in parallel or individually as required by the processing capability of the apparatus that performs the processing or as necessary. It may be.

  When the processing functions in the multipoint connection apparatus are realized by a computer, the processing contents of the functions that the multipoint connection apparatus should have are described by a program. And the processing function in the said multipoint connection apparatus is implement | achieved on a computer by running this program with a computer.

  The program describing the processing contents can be recorded on a computer-readable recording medium. As the computer-readable recording medium, for example, any recording medium such as a magnetic recording device, an optical disk, a magneto-optical recording medium, and a semiconductor memory may be used. Specifically, for example, as a magnetic recording device, a hard disk device, a flexible disk, a magnetic tape or the like, and as an optical disk, a DVD (Digital Versatile Disc), a DVD-RAM (Random Access Memory), a CD-ROM (Compact Disc Read Only). Memory), CD-R (Recordable) / RW (ReWritable), etc., magneto-optical recording medium, MO (Magneto-Optical disc), etc., semiconductor memory, EEP-ROM (Electronically Erasable and Programmable-Read Only Memory), etc. Can be used.

  The program is distributed by selling, transferring, or lending a portable recording medium such as a DVD or CD-ROM in which the program is recorded. Furthermore, the program may be distributed by storing the program in a storage device of the server computer and transferring the program from the server computer to another computer via a network.

  A computer that executes such a program first stores, for example, a program recorded on a portable recording medium or a program transferred from a server computer in its own storage device. When executing the process, the computer reads a program stored in its own recording medium and executes a process according to the read program. As another execution form of the program, the computer may directly read the program from a portable recording medium and execute processing according to the program, and the program is transferred from the server computer to the computer. Each time, the processing according to the received program may be executed sequentially. Also, the program is not transferred from the server computer to the computer, and the above-described processing is executed by a so-called ASP (Application Service Provider) type service that realizes the processing function only by the execution instruction and result acquisition. It is good. Note that the program in this embodiment includes information that is used for processing by an electronic computer and that conforms to the program (data that is not a direct command to the computer but has a property that defines the processing of the computer).

  In this embodiment, the multipoint connection apparatus is configured by executing a predetermined program on a computer. However, at least a part of these processing contents may be realized by hardware.

Claims (12)

A time direction division unit that divides each voice packet transmitted from each of three or more points into time units shorter than the packetization period (hereinafter referred to as subframe length units) and outputs divided voice packets;
A demultiplexing unit that extracts at least a first speech code and a second speech code from each of the divided speech packets corresponding to each of the points;
A decoder for decoding the first audio code corresponding to each of the points and outputting a first audio signal;
A mixing unit that mixes the first audio signal corresponding to each of the points and outputs a mixed audio signal for each of the points;
An encoder that encodes the mixed speech signal corresponding to each of the points and outputs a mixed speech code;
A point selection unit that determines a speech point from each of the above points and outputs a control signal corresponding to the speech point;
A second voice code switching unit for outputting a second voice code determined according to the control signal among the second voice codes corresponding to the points, for a point determined according to the control signal among the points; ,
Multiplex that outputs a unit voice packet in units of subframe length by combining the mixed voice code corresponding to each point and the second voice code corresponding to each point output from the second voice code switching unit. Cushing part,
Combining a plurality of the unit voice packets corresponding to each of the points, and outputting a voice packet for transmission having a time unit of a packetization period,
The point selection part
Either the sum of squares of the first audio signal corresponding to each point, the average of the absolute values of the first audio signal, or the average value of the code for each sample excluding the positive / negative code from the first audio code (Hereinafter referred to as power) acoustic attribute determination unit,
Applying a decision criterion to the power corresponding to each of the above points, an utterance point determining unit that determines the current utterance point from among the above points,
A memory for storing information representing an utterance point;
A control signal output unit that outputs a control signal corresponding to the determined utterance point,
The above decision criteria are:
If the power is a maximum point which is a main utterance location identical to the location of the previous one subframe obtained from the memory, mainly a main utterance location same point and the previous upper Symbol 1 sub frame of the current sub-frame Rather than the decision criteria for the above power, which is determined to be the utterance point,
If the power is a maximum point is the point different from the main utterance location before the 1 sub-frame, a point different from the main utterance location before the upper Symbol 1 subframe to be the primary utterance location of the current subframe determined The decision criterion for the power is higher ,
The utterance point determination unit determines an utterance point under the determination criterion. The multipoint connection device according to claim 1,
More certain one point to subframe lasts long it is determined as the main utterance location, to enhance the decision criteria when the power is the maximum point of a different point as the main utterance location of the previous subframe A multipoint connection device characterized by. The multipoint connection device according to claim 1,
The above decision criteria are:
The maximum power among the power corresponding to each of the above points for the current subframe is not less than α times the power corresponding to the main utterance point one subframe before (where α is a positive number greater than 1). In some cases, the main utterance point is changed to a point corresponding to the maximum power, and if it is less than α times, the main utterance point is not changed as the main utterance point one subframe before. Multipoint connection device. In the multipoint connection device according to claim 2,
The above decision criteria are:
The maximum power among the power corresponding to each of the above points for the current subframe is
More than α times the power corresponding to the main utterance point one subframe before (α is a positive number larger than 1) [However, the same utterance is the main point in a plurality of consecutive subframes up to the previous subframe. If it is a point, it is β times or more (β is a positive number larger than α). If it is, the main utterance point is changed to a point corresponding to the maximum power,
A multipoint connection device characterized by not changing the main utterance point as the main utterance point one subframe before when the number is less than α times. A time direction division unit that divides each voice packet transmitted from each of three or more points into time units shorter than the packetization period (hereinafter referred to as subframe length units) and outputs divided voice packets;
A demultiplexing unit that extracts at least a first speech code and a second speech code from each of the divided speech packets corresponding to each of the points;
A decoder for decoding the first audio code corresponding to each of the points and outputting a first audio signal;
A mixing unit that mixes the first audio signal corresponding to each of the points and outputs a mixed audio signal for each of the points;
An encoder that encodes the mixed speech signal corresponding to each of the points and outputs a mixed speech code;
A point selection unit that determines a speech point from each of the above points and outputs a control signal corresponding to the speech point;
A second voice code switching unit for outputting a second voice code determined according to the control signal among the second voice codes corresponding to the points, for a point determined according to the control signal among the points; ,
Multiplex that outputs a unit voice packet in units of subframe length by combining the mixed voice code corresponding to each point and the second voice code corresponding to each point output from the second voice code switching unit. Cushing part,
Combining a plurality of the unit voice packets corresponding to each of the points, and outputting a voice packet for transmission having a time unit of a packetization period,
The point selection part
Either the sum of squares of the first audio signal corresponding to each point, the average of the absolute values of the first audio signal, or the average value of the code for each sample excluding the positive / negative code from the first audio code (Hereinafter referred to as power) acoustic attribute determination unit,
A memory for storing the power corresponding to each of the points;
An utterance point determination unit that determines a current utterance point from among the points based on the current power corresponding to the points and the past power corresponding to the points stored in the memory; ,
A control signal output unit that outputs a control signal corresponding to the determined utterance point,
The utterance point determination unit
When the difference between the points regarding the power value of the current subframe corresponding to each point is smaller than the threshold, the magnitude relationship between the points regarding the power value corresponding to each point before one subframe is added to the above determination. A multipoint connection device characterized by: A time direction division unit that divides each voice packet transmitted from each of three or more points into time units shorter than the packetization period (hereinafter referred to as subframe length units) and outputs divided voice packets;
A demultiplexing unit that extracts at least a first speech code and a second speech code from each of the divided speech packets corresponding to each of the points;
A decoder for decoding the first audio code corresponding to each of the points and outputting a first audio signal;
A mixing unit that mixes the first audio signal corresponding to each of the points and outputs a mixed audio signal for each of the points;
An encoder that encodes the mixed speech signal corresponding to each of the points and outputs a mixed speech code;
A point selection unit that determines a speech point from each of the above points and outputs a control signal corresponding to the speech point;
A second voice code switching unit for outputting a second voice code determined according to the control signal among the second voice codes corresponding to the points, for a point determined according to the control signal among the points; ,
Multiplex that outputs a unit voice packet in units of subframe length by combining the mixed voice code corresponding to each point and the second voice code corresponding to each point output from the second voice code switching unit. Cushing part,
Combining a plurality of the unit voice packets corresponding to each of the points, and outputting a voice packet for transmission having a time unit of a packetization period,
The point selection part
Either the sum of squares of the first audio signal corresponding to each point, the average of the absolute values of the first audio signal, or the average value of the code for each sample excluding the positive / negative code from the first audio code (Hereinafter referred to as power) acoustic attribute determination unit,
A memory for storing the power corresponding to each of the points;
An utterance point determination unit that determines a current utterance point from among the points based on the current power corresponding to the points and the past power corresponding to the points stored in the memory; ,
A control signal output unit that outputs a control signal corresponding to the determined utterance point,
The utterance point determination unit
When the magnitude relationship between the points for the power value corresponding to each point before one subframe is more significant than the magnitude relationship between the points for the power value of the current subframe corresponding to each point, A multipoint connection apparatus, wherein a main utterance point is determined based on a magnitude relationship between points with respect to a power value corresponding to each point before one subframe. A time direction division step of dividing a voice packet transmitted from each of three or more points into time units shorter than the packetization period (hereinafter referred to as subframe length units) and outputting divided voice packets;
A demultiplexing step of extracting at least a first speech code and a second speech code from each of the divided speech packets corresponding to each of the points;
A decoding step of decoding the first speech code corresponding to each of the points and outputting a first speech signal;
Mixing the first audio signal corresponding to each point to output a mixed audio signal for each point; and
An encoding step of encoding the mixed audio signal corresponding to each of the points and outputting a mixed audio code;
A point selection step of determining a speech point from each of the above points and outputting a control signal corresponding to the speech point;
A second voice code switching step for outputting a second voice code determined according to the control signal among the second voice codes corresponding to the points, for a point determined according to the control signal among the points; ,
The mixed speech code corresponding to each of the points and the second speech code corresponding to each of the points output in the second speech code switching step are combined to output a unit speech packet in subframe length units. A multiplexing step;
Combining a plurality of the unit voice packets corresponding to each of the points, and outputting a voice packet for transmission having a time unit of a packetization period, and a time direction combining step,
The point selection step is
Either the sum of squares of the first audio signal corresponding to each point, the average of the absolute values of the first audio signal, or the average value of the code for each sample excluding the positive / negative code from the first audio code (Hereinafter referred to as power) acoustic attribute determination step,
Applying a decision criterion to the power corresponding to each of the above points, and determining an utterance point determining step for determining a current utterance point from among the above points;
A storage step of storing information representing the utterance point in a memory;
A control signal output step for outputting a control signal corresponding to the determined utterance point,
The above decision criteria are:
If the power is a maximum point which is a main utterance location identical to the location of the previous one subframe obtained from the memory, mainly a main utterance location same point and the previous upper Symbol 1 sub frame of the current sub-frame Rather than the decision criteria for the above power, which is determined to be the utterance point,
If the power is a maximum point is the point different from the main utterance location before the 1 sub-frame, a point different from the main utterance location before the upper Symbol 1 subframe to be the primary utterance location of the current subframe determined The decision criterion for the power is higher ,
In the utterance point determination step, the utterance point is determined under the determination criterion. The multipoint connection method according to claim 7,
More certain one point to subframe lasts long it is determined as the main utterance location, to enhance the decision criteria when the power is the maximum point of a different point as the main utterance location of the previous subframe A multipoint connection method characterized by The multipoint connection method according to claim 7,
The above decision criteria are:
The maximum power among the power corresponding to each of the above points for the current subframe is not less than α times the power corresponding to the main utterance point one subframe before (where α is a positive number greater than 1). In some cases, the main utterance point is changed to a point corresponding to the maximum power, and if it is less than α times, the main utterance point is not changed as the main utterance point one subframe before. Multipoint connection method to do. The multipoint connection method according to claim 8, wherein
The above decision criteria are:
The maximum power among the power corresponding to each of the above points for the current subframe is
More than α times the power corresponding to the main utterance point one subframe before (α is a positive number larger than 1) [However, the same utterance is the main point in a plurality of consecutive subframes up to the previous subframe. If it is a point, it is β times or more (β is a positive number larger than α). If it is, the main utterance point is changed to a point corresponding to the maximum power,
A multipoint connection method characterized by not changing the main utterance point as the main utterance point one subframe before if it is less than α times. A time direction division step of dividing a voice packet transmitted from each of three or more points into time units shorter than the packetization period (hereinafter referred to as subframe length units) and outputting divided voice packets;
A demultiplexing step of extracting at least a first speech code and a second speech code from each of the divided speech packets corresponding to each of the points;
A decoding step of decoding the first speech code corresponding to each of the points and outputting a first speech signal;
Mixing the first audio signal corresponding to each point to output a mixed audio signal for each point; and
An encoding step of encoding the mixed audio signal corresponding to each of the points and outputting a mixed audio code;
A point selection step of determining a speech point from each of the above points and outputting a control signal corresponding to the speech point;
A second voice code switching step for outputting a second voice code determined according to the control signal among the second voice codes corresponding to the points, for a point determined according to the control signal among the points; ,
The mixed speech code corresponding to each of the points and the second speech code corresponding to each of the points output in the second speech code switching step are combined to output a unit speech packet in subframe length units. A multiplexing step;
Combining a plurality of the unit voice packets corresponding to each of the points, and outputting a voice packet for transmission having a time unit of a packetization period, and a time direction combining step,
The point selection step is
Either the sum of squares of the first audio signal corresponding to each point, the average of the absolute values of the first audio signal, or the average value of the code for each sample excluding the positive / negative code from the first audio code (Hereinafter referred to as power) acoustic attribute determination step,
A storage step of storing the power corresponding to each of the points in a memory;
An utterance point determination step for determining a current utterance point from among the points based on the current power corresponding to the points and the past power corresponding to the points stored in the memory; ,
A control signal output step for outputting a control signal corresponding to the determined utterance point,
In the above utterance point determination step,
When the difference between the points regarding the power value of the current subframe corresponding to each point is smaller than the threshold, the magnitude relationship between the points regarding the power value corresponding to each point before one subframe is added to the above determination. A multipoint connection method characterized by: A time direction division step of dividing a voice packet transmitted from each of three or more points into time units shorter than the packetization period (hereinafter referred to as subframe length units) and outputting divided voice packets;
A demultiplexing step of extracting at least a first speech code and a second speech code from each of the divided speech packets corresponding to each of the points;
A decoding step of decoding the first speech code corresponding to each of the points and outputting a first speech signal;
Mixing the first audio signal corresponding to each point to output a mixed audio signal for each point; and
An encoding step of encoding the mixed audio signal corresponding to each of the points and outputting a mixed audio code;
A point selection step of determining a speech point from each of the above points and outputting a control signal corresponding to the speech point;
A second voice code switching step for outputting a second voice code determined according to the control signal among the second voice codes corresponding to the points, for a point determined according to the control signal among the points; ,
The mixed speech code corresponding to each of the points and the second speech code corresponding to each of the points output in the second speech code switching step are combined to output a unit speech packet in subframe length units. A multiplexing step;
Combining a plurality of the unit voice packets corresponding to each of the points, and outputting a voice packet for transmission having a time unit of a packetization period, and a time direction combining step,
The point selection step is
Either the sum of squares of the first audio signal corresponding to each point, the average of the absolute values of the first audio signal, or the average value of the code for each sample excluding the positive / negative code from the first audio code (Hereinafter referred to as power) acoustic attribute determination step,
A memory for storing the power corresponding to each of the points;
An utterance point determination step for determining a current utterance point from among the points based on the current power corresponding to the points and the past power corresponding to the points stored in the memory; ,
A control signal output step for outputting a control signal corresponding to the determined utterance point,
In the above utterance point determination step,
When the magnitude relationship between the points for the power value corresponding to each point before one subframe is more significant than the magnitude relationship between the points for the power value of the current subframe corresponding to each point, A multipoint connection method, wherein a main utterance point is determined based on a magnitude relationship between points with respect to a power value corresponding to each point before one subframe.

Images