JPH0974446A - Voice communication controller - Google Patents

Abstract

PROBLEM TO BE SOLVED: To enhance the articulation of a transmission content by using a sound image control parameter in a multi-point conference system so as to conduct processing, thereby giving a sound image different from other voice to as least are person's voice. SOLUTION: An exchange section 11 selects communication line from a terminal equipment making a participant request to a conference and connects to a voice addition control section 10 via an input channel. A sound image control section 14 in the voice addition control section 10 applies processing to a distributed voice signal derived from each terminal equipment through the use of one or desired combinations of sound image control parameters. The processing provides the sound image of an object space position different from each terminal equipment to the distributed voice signal. An adder section 15 adds the distributed voice signals and a terminal dependent distribute section 16 distributes the sum signal to all terminal equipments.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a voice communication control apparatus for controlling voice signal addition in a multipoint communication conference involving voice communication such as a voice conference, a video conference and a multimedia conference conducted via a communication line.

[0002]

2. Description of the Related Art In a voice conference device, a multi-point video conference device, etc., a voice signal from each conference participant is multiplied by a weighting coefficient according to the number of simultaneous speakers, and the result is added to each conference participant. A voice communication control device for returning the generated voice signal is used. Conventionally used voice communication control devices include those having means for adding voice signals of all conference participants, those having means for permitting transmission of voice signals in response to a speech request, and the like.

However, in any type of voice communication control device, the channel of the voice signal for transmitting the selected and added voice signal to each terminal is limited to only one channel. In the method of adding the voice signals of a plurality of people, when a plurality of conference participants speak at the same time, the voices of a plurality of people are mixed and reproduced from one sound source (speaker). As a result, the listener's intelligibility of the content of the statement is reduced. At the same time, it becomes difficult to fix the speaker. Furthermore,
It is necessary for the participants to perform an operation for requesting to speak. Therefore, there is a problem that free conversation is difficult.

On the other hand, it is known that when the listener locates the sound image for the voice of each speaker at different spatial positions, the identification of each speaker becomes easy and the intelligibility of the uttered content is improved ( DRBegault, “Multichannel Spatial Audito
ry display for Speech Communication, ”Journal of t
he Audio Engineering Society, 42, pp. 819-826, 1994).
However, sound image localization is to determine the position of the sound that is received. Usually, the position of the sound source and the position where the sound image is localized coincide. However, a technique has been devised that allows a listener to localize a sound image at a desired spatial position. Here, the following method is known as a representative example of a sound image localization processing method for localizing a plurality of sounds in different virtual space positions. Convoluted with the audio signals S ₁ as shown in FIG. 2 the acoustic transfer functions H _1L from the sound source 1 in FIG. 1 to the right and left ears, the H _1R respectively. at the same time,
The acoustic transfer functions H _2L and H _2R from the sound source 2 to each of the left and right ears are convoluted with the audio signal S ₂ different from the audio signal S ₁ , respectively. The audio signals obtained by the convolution calculation are added, and the sound by the added audio signals is presented to both ears using stereo headphones or the like. As a result, sound stimuli S ₁ ^* H _1L ＋ S ₂ ^* H _2L and S ₁ ^* H _1R ＋ S ₂ ^* H _2R, which are equivalent to those when the sound sources 1 and 2 reach both ears as shown in FIG. Obtained in the ear. In that case, the listener can localize the sound for each of the sound source signals S ₁ and S _{2 at the} same spatial position as the sound sources 1 and 2 in FIG. Other methods are also described in "Spatial Acoustics" by Brauert, Goto, Morimoto.
He is familiar with books such as (Kashima Press).

As an example of the prior art in which the above knowledge is applied to multipoint voice communication, there is a conference call terminal device described in JP-A-4-10744. As shown in the schematic diagram of FIG. 3, a communication terminal device provided with means 3L, 3R for performing audio signal processing for localizing reproduced sound from audio signals received from each point to different target space positions is proposed. Was done. In order to operate this device,
Number (or terminal number) ID for each terminal (caller) at each point
In advance, the point number I
It is essential to transmit D as sender control information. The purpose of assigning the point number ID is to identify the source point and separate the audio signal for each point. That is, a signal received from another communication terminal device is separated into a voice signal and an ID code by a signal separation unit, and a pair of pairs corresponding to the spatial position assigned to the terminal of that ID by the switching control unit 2 is separated according to the separated ID code. The audio signal processing means 3R and 3L for convoluting the acoustic transfer function are selected, and the selected processing means and the audio signal separated from the signal separation unit 1 are supplied. As a result, the reproduced sound image is localized at the assigned spatial position. In order to use this device, there is a drawback that the device must be installed between each point and the communication system must be predetermined. The above drawbacks impede economical realization of multipoint voice communication in which the voice of the talker is localized in different virtual space positions between arbitrary points.

In a communication conference limited to two points, a device has been proposed in which the sound field of the conference room is recorded by a stereo microphone and the acoustic environment of the own point is transmitted to the other point by a stereo codec or the like (for example, USPatent 5,020,098). . However, in applying to multipoint communication connecting three or more points,
Lines must be connected between each point.
As shown in FIG. 4, it is necessary to provide a terminal 4 equipped with the sound image localization signal processing means 4A at each point and connect each terminal 4 to each point by a communication line. At this time, the required number of lines C _M is at least the number of combinations M of connected terminals M M × (M−1) / 2. This connection method is unrealistic because not only the number of lines required as the number of connected terminals M increases, but also communication between terminals connected in advance becomes impossible.

Other methods for realizing multipoint voice communication in which the listener localizes the voices of the respective speakers to different spatial positions by using the same sound image localization technology as described above, for example, in the literature (Cohen). M., Koizumi N. and Aoki S. "Design and Contro
l of Shared Conferencing Environments for Audio Te
lecommunication, "Proc. Int. Symp. on Measurement
and Control in Robotics, pp.405-412, Nov. 1992)
In the communication performed between multiple points as described in (1), there has been proposed a method of performing communication between arbitrary two or more groups of points among the points.

Also in this method, the voice signal processing means for processing the voice signal transmitted from each point through the communication line so as to localize the voice of each talker at each point at each point, and this voice signal processing means It is premised that an adding means for adding the generated audio signals is installed in each terminal. Furthermore, it is necessary that the point of origin is specified and that a voice signal be transmitted at each point. Therefore, there is a drawback that the communication method must be predetermined.

[0009]

DISCLOSURE OF THE INVENTION An object of the present invention is to understand the utterance content that occurs when a plurality of speakers speak at the same time without giving each terminal a large voice signal processing capability in a multipoint communication conference. Another object of the present invention is to provide a voice communication control device that realizes an excellent multipoint communication conference.

Another object of the present invention is to provide a voice communication control device which can be used by any terminal accommodated in a communication network. Still another object of the present invention is to provide a voice communication control device for simultaneously realizing communication between a combination of two or more connected terminals in which each listener localizes voices by a plurality of transmitters at different positions. Especially.

[0011]

SUMMARY OF THE INVENTION The present invention is a voice communication control device for connecting to at least three terminals through communication lines to hold a communication conference, and for connecting to the above terminals through communication lines. An exchange unit, a plurality of input channels connected to the exchange unit, to which voice signals from the terminal are input, and a branch voice signal of a plurality of channels, each of which has predetermined input voice signals from the plurality of input channels. A channel branching unit for branching to, a branching sound signal of the plurality of channels corresponding to each of the terminals, a sound image control unit that performs a sound image control process by a predetermined sound image control parameter, and generates a sound image control sound signal, The sound image control audio signals of the plurality of channels corresponding to the respective terminals are added for each channel to generate an added audio signal of the plurality of channels. An adder, the sum audio signals of the plurality of channels distributed in correspondence to the terminal,
A branch unit corresponding to a terminal for giving to the exchange unit, and the sound image control unit, for branch audio signals of the plurality of channels corresponding to at least one terminal, the plurality of channels corresponding to other terminals. A process of giving a sound image different from the sound signal is performed.

[0012]

FIG. 5 shows an outline of a multipoint communication conference system using a voice communication control device according to the present invention. The voice communication control device 100 of the present invention has an exchange section 11. This line conversion unit is, for example, ISDN or LAN.
Etc., and can be used from terminals at any connected points. Voice communication control device 1
The maximum number of participants (the number of terminals) N that can participate in the conference at the same time is determined in advance due to the limitation of the size of 00 and the processing capacity.
N is an integer of 3 or more. The conference participation terminals TM-1 to TM-4 are selectively connected to the _N input channels C _{1 to} C _N by the exchange unit 11. The selectively connected terminal has an input channel C ₁
To C _N are connected to the voice addition control unit 10 to form a multipoint communication conference system in which terminals can talk with each other. As will be described in detail later, the voice addition control unit 10 adjusts the volume (attenuation rate), the delay time, the phase, and the transfer function of the branched voice signal originating from each terminal (these are referred to as sound image control parameters). Treatment using any one of them or a desired combination. By this process each terminal
In the TM, at least one of the voices of the plurality of participants played at the same time gives a different sound image to the rest of the voices.

FIG. 6 is a block diagram showing the basic configuration of the voice communication control device 100 of the present invention used in the system of FIG. The switching unit 11 selects a communication line from a terminal that is requested to participate in the common conference and selects N input channels C ₁ to
The voice addition control unit 10 is connected through C _N. The voice addition control unit 10 predetermines K (K is an integer of 2 or more, K = K in FIG. 6) of voice signals input to _N input channels C _{1 to} C _N from a maximum of N connected terminals. 2 and
Branch channels B _JL , B for each left and right)
A channel branching unit 13 that branches to a branching audio signal on _JR (J = 1, ..., N), and a sound image control unit 14 that controls the K sets of N branching audio signals by a predetermined sound image control parameter. , N sets of N-channel added audio signals, and an addition unit 15 that adds the audio signals of N branch channels corresponding to N sets of image-controlled audio signals to generate an added audio signal of K channels. It is composed of a terminal-corresponding branching unit 16 which is branched and is given to the switching unit 11 as N sets of K channel signals. Channel branching unit 13, sound image control unit 14
The adder 15 and the adder 15 constitute an audio signal processor 25.
The switching unit 11 converts N sets of K channel signals into N sets.
It is sent to each of the participating terminals TM-1 to TM-N. FIG. 6 shows that voice signals of 2 channels (K = 2) for N communication channels are transmitted from the switching unit 11 to the participating terminals through two downlinks.

As described above, K = in the principle configuration shown in FIG.
2 shows the case of each branch point of the channel branching unit 13.
Input audio signals are branched into 2-channel signals at 3-1, 3-2, ..., 3-N. According to convention, the two channels correspond to the left and right channels. Here, L is added to the symbol relating to the left channel, and R is added to the symbol relating to the right channel. The sound image control unit 14 includes N sets of signal processing units 4-1L, 4-1R, 4-2.
It consists of L, 4-2R, ..., 4-NL, 4-NR, and processes the branched audio signals using sound image control parameters of a predetermined type. As described above, volume, phase, delay time, transfer function and the like can be used as the types of sound image control parameters. For example, assuming that the number of branches of the input audio signal at each branch point of the channel branching unit 13 is 2, the effect of sound image control by various sound image control parameters will be briefly described below. (a) When using the volume (any of level, attenuation rate, amplification rate, etc.) as a sound image control parameter, by controlling the relative volume (level) for the left and right branch audio signals corresponding to each input audio signal, The direction of the sound image reproduced by the two speakers in the terminal based on the audio signal can be controlled to a desired direction between the left and right speakers. (b) When using the phase (in-phase and anti-phase) as a sound image control parameter, if the left and right branch audio signals corresponding to each input audio signal are controlled so that they are in phase or opposite phase, the reproduced sound image It is possible to give a sense of distance or eliminate the sense of distance. (c) When the delay time is used as a sound image control parameter, the direction of the reproduced sound image is adjusted to a desired direction in the horizontal plane of the front space by controlling the relative delay amount for the left and right branch audio signals corresponding to each input audio signal. Can be controlled. (d) When the acoustic transfer function is used as a sound image control parameter, the left and right branch audio signals corresponding to each input audio signal are convolved with a pair of acoustic transfer functions corresponding to the target position to reproduce a sound image reproduced by stereo headphones. It can be localized at the target position.

The sound image control parameters are the signal processing units 4-1L, 4
Parameter setting part 1 for -1R, 4-2L, 4-2R, ..., 4-NL, 4-NR
4C. For example, a method of determining the sound image control parameter according to the number of people participating in the communication conference is possible. In the case of FIG. 6, the audio signals from the signal processing units 4-1L, 4-2L, ..., 4-NL are added by the adder 5L of the adding unit 15, and the left channel added audio signal is generated. Signal processing section 4-1R, 4-2R,
The voice signal from 4-NR is added by the adder 5R of the adder 15 to generate the right channel added voice signal. Therefore, the K channel signal distributed from the terminal-corresponding branching unit 16 to each of the participating terminals TM-1 to TM-N includes a component of the voice signal transmitted from all the participating terminals.

Each of the terminals TM-1 to TM-N is constituted by a microphone MC, a transmitting section 51, a decoding section 52, and reproducing sections 53L and 53R as shown in FIG. Received K (=
2) The coded audio signal of the channel is decoded into the audio signal of each channel by the decoding unit 52, and converted into the audio by the reproducing units 53L and 53R, respectively. Therefore, the voice heard by the user of the terminal TM includes the voice transmitted from all the participating terminals.

In the present invention, in the sound image controller 14, N
By selecting different sound image control parameters for the pair of branched audio signals, it is possible to distinguish the sound image for the sound from at least one participating terminal and the sound image for the sound from another participating terminal in each participating terminal TM. Become. The characteristics of the controlled sound image are the spatial position and the sense of spaciousness of the sound that the listener perceptually perceives. For example, when the reproducing units 53L and 53R of the terminal are speakers, respectively, as the sound image control parameter for the left and right audio signals, any one of the level difference between channels, the time difference between channels, and the phase (same phase, opposite phase) or the level difference is set. The sound image can be controlled by controlling the combination of the time differences. Therefore, sound image control parameters given by the signal processing units 4-1L, 4-1R, 4-2L, 4-2R, ..., 4-NL, 4-NR for N sets of left and right audio signals in the sound image control unit 14 of FIG. By selecting each appropriately,
It is possible to give a desired sound image to the audio component from each terminal included in the audio reproduced by the participating terminals in FIG. 7. When the reproduction units 53L and 53R in FIG. 7 are stereo headphones, K is limited to 2. Further, the sound image control unit 14 of FIG. 6 performs convolution calculation of the transfer functions corresponding to the spatial positions of the desired sound sources with respect to the N sets of left and right audio signals, respectively, as the sound image control parameters, and reproduces from the reproduction units 53L and 53R of FIG. A sound image that localizes a desired spatial position is given to the audio component from each terminal.

In each of the following embodiments, the case of K = 2 is shown, but when a speaker is used as the reproducing unit of the terminal, K ≧ 3.
Is also possible. Hereinafter, specific embodiments of the present invention will be described with reference to the drawings. First Embodiment FIG. 8 shows a first embodiment of a voice communication control device based on the principle configuration of FIG. 6 according to the present invention. Multiple terminals TM-1 ~
The TM-M is accommodated in the voice communication control device 100 according to the present invention via the communication network 40. In this embodiment, by monitoring the voice signals on the input channels C _{1 to} C _N from the switching unit 11 connected to the plurality of participating terminals, the voice signal on which input channel is the voice signal by the main speaker. Is decided. Here, the sound image of the reproduced voice by the main speaker can be distinguished from the sound image of the voice by the other terminal.

The voice communication control apparatus 100 of this embodiment includes a switching unit 11, a separation / multiplexing unit 22 for voice signal data and control signal data or image data, and a voice signal decoding unit 23.
A, voice detection processing unit 23B, speaker selection unit 24, voice signal processing unit 25, main speaker voice subtraction unit 26, voice encoding unit 2
7, 28, downlink voice signal selection unit 29, signal processing control unit 2
0, the image display control unit 30 is included. The switching unit 11, the demultiplexing / multiplexing unit 22, the voice signal decoding unit 23A, and the voice detection processing unit 23B perform processing corresponding to the connected line, and have a processing capacity corresponding to the maximum number N of simultaneously connected terminals.

FIG. 8 shows, as an example of the terminals TM-1 to TM-M, an image conference terminal that simultaneously transmits and receives image information and audio information. The presence of image information is not essential to the configuration of the present invention, and the image display control unit 30 is not directly related to the gist of the present invention, and thus detailed description of the image display control is omitted. However, it is also possible to specify the conference in which each of the terminals TM-1 to TM-M participates and control the configuration of the conference based on the image information.
In this case, a signal related to the audio signal addition control is sent from the image display control unit 30 to the signal processing control unit 20.

The operation of the apparatus will be described below by taking as an example the case where M terminals (TM-1 to TM-M) are connected to the voice communication control apparatus 100 through the communication network 40. As the communication network 40, a line capable of immediate two-way voice communication is used. For example, N-ISDN line, leased line, analog telephone line, LA
Any of N lines, individual lines, and multiplexed logical lines can be applied. It does not matter whether it is wired or wireless. However, it is necessary to adapt the exchange unit 11 to the type of the communication network 40. In this embodiment, an N-ISDN line (for image, about 64 kbps,
An example of using a transmission band of about 64 kbps for voice will be described.

As an example of the terminals TM-1 to TM-M, N-ISD
It is possible to connect a video conference terminal using N lines. However, the terminals TM-1 to TM-M need to have a function of receiving audio signals for two channels. The terminals TM-1 to TM-M are connected to the switching unit 11 of the voice communication control device 100 through the communication network 40.
Connected to. For example, ITU-T H.264. The image information and audio signal information transmitted from each terminal TM-1 to TM-M, which are multiplexed into one channel according to the standard of 221 etc., and the audio or image control information for controlling the structure of the conference are separated. / Demultiplexed by the multiplexing unit 22. The separated image information and image display control information are sent to the image display control unit 30. The image display control is not directly related to the present invention, and therefore its explanation is omitted.

The voice control information is sent from the demultiplexing / multiplexing unit 22 to the signal processing control unit 20. Examples of the voice control information include control such as conference participation / exit request. The audio signal separated by the demultiplexing / multiplexing unit 22 is decoded by the audio signal decoding unit 23A. For example, it is converted into a PCM coded audio signal so that the subsequent processing can be performed. For simplicity, in the following processing, this will be simply called an audio signal.

In the voice detection processing section 23B, for example, a voice signal is sent.
The voice is detected by a method such as signal power detection. The voice is
When detected, the detection is indicated to the signal processing control unit 20.
Send a control signal. Sound in voice detection processing unit 23B
An example of the voice detection method is shown in FIG. Input voice signal waveform (Figure
9A) based on the power within a unit time (for example, 100 ms)
-The integrated value IT (Fig. 9B) is obtained. Where power integral value
IT is ON detection threshold E _ON, OFF detection threshold E_OFFLarge with
It is determined whether or not the state is the speaking state by comparing the small numbers.

In the first utterance identification method, the terminal is immediately determined to be in the utterance state when the power integrated value IT per unit time exceeds the ON detection threshold E _ON . Power integrated value IT is OFF
When it falls below the detection threshold E _OFF , it is immediately determined to be in the non-speech state. Therefore, the sections (ab, cd, fg) shown by hatching in FIG. 9C are determined to be the speaking state. In this first utterance identification method, the state is changed frequently between the utterance state and the non-utterance state.

Compared to the first utterance identification method, the second utterance identification method is in the utterance state for a fixed time (T shown in FIG. 9D) after the power integrated value IT of the unit time falls below the OFF detection threshold E _OFF . The difference is that the speech state identification is performed assuming that the speech state is continuing. According to this method, the section (a-e, f-h) indicated by hatching in FIG. 9D is identified as the speaking state.

In the voice detection processing section 23B
The voice signal from the identified terminal is the speaker selection unit 2 in FIG.
Selected in 4. The selected audio signal is the selected audio channel.
Le A ₁~ A_NAudio signal processing unit 25
Given to. The audio signal processing unit 25 is the basic structure of the present invention.
Channel branching unit 13, sound image control unit 14, adding unit
Includes 15.

The signal processing control unit 20 includes a demultiplexing / multiplexing unit 2
2. It operates based on a control signal received from the voice detection processing unit 23B, a conference control signal received from the image display control unit 30, or the like. Here, the selected voice channel corresponding to the terminal to be added and its priority are determined by discriminating from the current number of speakers, the number of speakers requesting, and the state of the terminal such as the chairperson who needs to always give the right to speak. Speaker selection unit 24
Connects the selected voice channels A _{1 to} A _N to be added to the input channel positions according to the priority order. In this embodiment, a voice signal for the main speaker is selected. Voice channel A
₁ , and the voice signals for the second to Nth speakers are passed to the selected voice channels A ₂ , A ₃ , ..., A _N.

The signal processing control section 20 is an audio signal processing section 25.
Control the behavior of. Audio signal processing is performed to add and distribute audio signals from a plurality of speakers without impairing the intelligibility of the main speaker's voice. In this example, the voice signal of the main speaker is sent to the left channel voice signal and the voice signals of the other speakers are all sent to the voice encoding unit 28 through the right channel voice signal. The voice encoding unit 28 uses a stereo encoder to output the voice signal processing unit 25.
The two-channel stereo addition audio signal from is encoded and multiplexed. The downlink voice signal selection unit 29 corresponds to the terminal corresponding branch unit 16 in FIG. Here, the encoded stereo addition voice signal from the voice encoding unit 28 is selected for the line corresponding to the terminal other than the main speaker. For the line corresponding to the terminal of the main speaker, the voice encoding unit 27
The stereo audio signal encoded by is selected. However, in order to cancel the echo, the main speaker voice subtraction unit 26 subtracts its own voice signal from the left and right channel added voice signals and then encodes it. The selected audio signals are sent to the demultiplexing / multiplexing unit 22, respectively.

The demultiplexing / multiplexing unit 22 multiplexes the stereo audio signal from the downlink audio signal selecting unit 29 and the image information from the image display control unit 30. The multiplexed signal is sent to each terminal TM-1 to TM-M via the switching unit 11 and the communication network 40. The audio signal processing unit 25 performs addition processing and sound image control processing based on human auditory characteristics and habits at the time of conference, which play an important role in the present invention. The audio signal processing unit 25 has the signal channel branching unit 13, the sound image control unit 14, and the adding unit 15 in FIG. 6 as described above. In this example, the branch audio signal of each selected audio channel is controlled using the attenuation amount that controls the audio signal level difference between channels as the sound image control parameter and the phase that controls the in-phase or anti-phase of the audio signal between the left and right channels. To do.

FIG. 10 shows a concrete example of the audio signal processing unit 25, and in this example, a branched audio signal is controlled by using a phase between channels as a sound image control parameter. The level control unit 14A and the phase control unit 14B form a sound image control unit 14. Selected audio channels A _{1 to} A _N
Sound signal is attenuators 4-1, 4-2, ..., 4 of the level control unit 14A.
-N attenuates the level to 2 ^-1/2 , N ^-1/2 , ..., N ^-1/2 times, respectively. The audio signals output from the attenuators 4-1 to 4-N are divided into left and right branch channels B _1L , B _1R , ..., B at branch points 3-1, 3-2, ..., 3-N of the channel branch unit 13, respectively. The left and right channel signals on the _NL and B _NR are branched, and the phase controller 14B uses the phase controllers 4-1L, 4-1R, 4-2L, 4-2R, ... It is controlled to each phase or antiphase. The signal processing control unit 2 sets the sound image parameters such as the attenuation amount and the phase.
The parameter setting unit 14C performs the control from 0.

The voice signal of the main speaker, that is, the signal of the selected voice channel A ₁ is attenuated by the attenuator 4-1, for example to 2 ^-1/2 , and is branched into the left channel and the right channel at the branch point 3-1. The phase controllers 4-1L and 4-1R are provided to the adders 5L and 5R in the same phase between the left and right channels. Adder 5L, 5R
The left-channel audio signal and the right-channel audio signal on the output side of are corresponding to the left-channel audio signal and the right-channel audio signal at the receiving terminal. Therefore, when the listener of the receiving terminal is listening in stereo reproduction, the selected audio channel A ₁
The (main speaker) 's voice is heard with a sense of distance, with the sound image position being localized in the center.

[0033] As the sum of the reproduced sound power level in each terminal of the speaker selection voice channels A ₂ to A _N is smaller than the main speakers equal to or reproduced audio, select audio channels A ₂ to A _N Of the audio signal is attenuated by, for example, N ^-1/2 times (N is the number of selected channels A _{1 to} A _N ) by attenuators 4-2 to 4-N and then branched to branch points 3-2 to 3-N. , The right channel audio signal is given to the adder 5L in the same phase by the phase controllers 4-2L to 4-NL, and the left channel audio signal is phase controller 4-2R to 4-NR. It is phase-inverted (multiplied by -1) and then given to the adder 5R.

In the stereo reproduction, when sounds of opposite phases are presented from the left and right channels, the listener perceives a sound image without a sense of distance. By utilizing this auditory characteristics, minor speech signal entering said selected voice channel A ₂ to A _N, when the listener of the receiving terminal is listening in stereo reproduction, lost sense of distance around the listener Is perceived. On the other hand, the voice of the main speaker is localized and perceived in a certain direction. Level control unit 1 of audio signal processing unit 25 of FIG.
The attenuators 4-1 to 4-N in 4A are for making the voice level of the main speaker reproduced in the terminal higher than the sum of the voice levels of other speakers. The difference in the localization between the sound image of the reproduced voice and the sound image of the reproduced voice of the other speaker is given by the phase control unit 14B exclusively depending on whether the left and right channels have the same phase or opposite phases.

FIG. 11 shows another embodiment of the audio signal processing unit 25. In the terminal, for example, only the main speaker's voice is reproduced from the left speaker, and the mixed voice of all speakers is reproduced from the right speaker. This is a case where the power level is reproduced to be equal to or lower than the power level of the main speaker voice. An attenuator having an attenuation rate N ^-1/2 for each of the right branch channels B _{1R to} B _NR branched by the channel branch unit 13.
4-1R, 4-2R,…, 4-NR is inserted and left branch channel of the main speaker
The attenuation amount of the attenuator 4-1L of B _1L is set to 0, and the attenuation factors N ^{-1/2 of the} right branch channel are used for the attenuators 4-2L, ..., 4-NL of the other left branch channels B _{2L to} B _NL. A much larger, eg infinite, attenuation is set (ie the channel is blocked). Therefore, the left channel adder 5L receives only the voice signal of the main speaker's selected voice channel A ₁ without being attenuated, and the right channel adder 5R receives the signals of all selected voice channels A _{1 to} A _N. It is given after being attenuated to a proper volume, for example, N ^-1/2 at 4-1R to 4-NR.

The listener of the receiving terminal simultaneously listens to the voice of the main speaker and the voice of the subordinate who have been subjected to the sound image control processing, but the perceived sound image position is different. Therefore, the listener can always hear the speech of the main speaker clearly, and at the same time, it is possible to realize a multipoint communication conference in which the speech of each point by other speakers can be simultaneously grasped.

An example of the method of determining the main speaker in the signal processing control section 20 and the method of generating the added audio signals of the left and right channels by using the audio signal processing section 25 of FIG. 11 is shown in FIG.
It is shown in FIG. In FIG. 12, as a method of determining the main speaker, “a terminal first recognized as the speaking state among terminals recognized as the speaking state at a certain time is regarded as the main speaker while the speaking state continues. It shows the "decision" method. When there is only one terminal in the speaking state, that terminal is determined to be the main speaker. When a plurality of terminals are in the speaking state at the same time, when the main speaker is in the non-speaking state, the terminal that is in the speaking state earliest is determined as a new main speaker. Row A of FIG.
Symbols D to D respectively indicate the speaking periods NA, NB, NC, and ND of the speakers in the terminals TM-1 to TM-4 in shaded areas. Row E shows the audio content of the left channel, and row F shows the audio content of the right channel. The sound of the right channel shows how the sounds of the participants other than the main speaker are added.

FIG. 13 shows another example of the method of determining the main speaker in the signal processing control section 20 and the method of generating the added audio signals of the left and right channels by using the audio signal processing section 25 of FIG. In the example shown in FIG. 13, a speech priority is given to a specific terminal (terminal TM-1 in the example of FIG. 13). This control method corresponds to the giving of the preferential speaking right to the chairperson at the conference, the giving of the preferential speaking right to the speaker at the lecture, and the like. Rows A to D of FIG. 13 show the message contents NA to ND of the terminals TM-1 to TM-4. Row E shows the contents of the left-channel audio signal, and row F shows the addition of the right-channel audio signal.

Further, the two identification methods shown in FIG. 9 are also suitable / unsuitable depending on the contents of the conference held in the communication conference. For example, when a plurality of peer terminals hold free discussions, the identification method 1 that exchangs speakers quickly is superior. On the other hand, in the case of a discussion form in which each person makes an opinion in turn, the identification method 2 in which unintended main speaker change is unlikely to occur is superior.

Therefore, in the signal processing control unit 20, as shown in FIG.
It has a plurality of speaker detection / addition control algorithms as illustrated in FIG. 12 and FIG. 13, and the speaker detection / addition control is performed by operations from the terminals TM-1 to TM-M according to the progress of the conference. It is effective to provide a means for switching the algorithm. The speaker selection unit 24 is the same as the speaker selection unit 24 of FIG.
This is provided when the sound image control parameter is fixedly set in the audio signal processing unit 25 shown in FIG. 0, and for which set of left and right branch channels B _JL and B _JR the audio signal of the audio signal processing unit of FIG. However, if the phase can be set to the same phase or the opposite phase with each other and the attenuation rate can be selected and set to 1 or 1 / N ^1/2 for any selected audio channel A _J , which input Even if the voice signal of the channel is the voice signal of the main speaker, the parameter setting unit 14C sets the sound image control parameters for the channel and the sound image control parameters for the other channels by using the channel of the main speaker in FIG. (Selected audio channel A
_The speaker selecting unit 24 is not necessary if the setting is made in the same manner as the relationship between the parameter for ₁ ) and the parameters for the other selected voice channels A _{2 to} A _N. Similarly, in FIG. 11, if the parameter setting unit 14C can selectively set the attenuation rate for each branch channel to 0, 1 / N ^1/2 , or 1 / ∞, the speaker selection unit 24 It is unnecessary.

Further, in FIGS. 10 and 11, when the phase of the audio signal between the left and right branch channels is controlled so that the sound image of the main speaker's voice signal can be distinguished from the sound image of the other speaker's voice signal, the left and right channels are controlled. The case where the distribution of the audio signal to the controller is controlled is shown. In these cases, speakers arranged on the left and right are used as the reproduction units 53L and 53R in each terminal. On the other hand, the signal processing units 4-1L and 4- in FIGS.
Instead of the phase or the attenuation amount as a sound image control parameter in NL, 4-1R, and 4-NR, a left-right acoustic transfer function shown in FIG. 18 described later may be used. In that case, stereo headphones must be used as the reproduction units 53L and 53R in each terminal. Second Embodiment FIG. 14 shows an embodiment modified from the embodiment of FIG. 8 so that the user at each terminal can participate in a plurality of conferences. FIG.
In the embodiment shown in FIG. 4, Q voice signal processing devices 52-1 to 25-Q are provided corresponding to the Q conferences, and each terminal TM-1 to TM-1.
A meeting selection unit 21-1 to 21-N is provided for each TM-N, so that a plurality of meetings can be held. Furthermore, one terminal can simultaneously participate in two or more conferences.

The participants of the conference are the voice communication control devices 100.
One or a plurality of conferences in which the user participates are designated for the speaker selection unit 24 of. However, when a plurality of conferences are designated, one conference (main) to which the user's voice is added is designated. The conference that processes the audio signal from the terminal as an addition target is uniquely determined. For other conferences, the audio signal from the terminal is not an addition target, and the participant only listens to the added voice of the conference.

As one of the logical conferences, for example, there is a dialogue between two or more specific participants of the conference. In this case, one terminal in FIG. 14 can simultaneously talk with a specific party while listening to the conference. as a result,
It becomes possible to have the same natural conversation as when all participants physically participate in the same conference room. In the embodiment of FIG. 14, the inside of the speaker selection unit 24 is logically divided into a plurality of conferences (conference 1 to conference Q), and the same story as described in the description of FIG. 9 is given for each conference room. Person detection.

The configurational difference of the embodiment of FIG. 14 from that of FIG. 8 is that a sufficient number of audio signal processing units 25-1 to 25-Q are provided for the number of meetings Q that can be held, which is logical. One of the audio signal processing units 25-1 to 25-Q is assigned to one conference, and conference selecting units 21-1 to 21-N are provided at the subsequent stages, respectively. As an embodiment of each audio signal processing unit 25-1 to 25-Q in this embodiment, for example, the audio signal processing unit shown in FIG. 10 or FIG. 11 can be used.

FIG. 15 shows a configuration example of the conference selection unit 21-J corresponding to the terminal TM-J in the conference selection units 21-1 to 21-N in the second embodiment of FIG. . The conference selection unit 21-J is a conference selection switch 7S-1 to 7S-Q to which respective left and right channel audio signals of the Q signal processing control units 25-1 to 25-Q are given.
And all of those conference selection switches 7S-1 to 7S-Q are connected to the left channel adder 2-L and the conference selection switches 7S-1 to 7S-Q are connected to the right channel output. And the added right channel adder 2-R. The signal processing control unit 20 turns on one or more conference selection switches 7S-P (1 ≦ P ≦ Q) corresponding to the designated conference based on the control signal that designates the participating conference received from the terminal TM-J. Then, the designated conference is selected.

Left and right audio signal outputs from the audio signal processing devices 25-1 to 25-Q corresponding to the conferences 1 to Q are branched by the terminal corresponding branching unit 16 to be included in the conference selecting unit 21-J corresponding to each terminal. Sent to the conference selection switches 7S-1 to 7S-Q. Thereby, the left and right audio signals of one or more conferences designated by each terminal TM-J are selected, and the left and right channel adders 2-
Given to L, 2-R. For example, when participating in two conferences at the same time, the left channel audio signals of the two participating conferences are added by the adder 2-L and output as the left channel audio signal. The right channel audio signal of is added by the adder 2-R and output as a right channel audio signal. Since the left and right channel audio signals are encoded by the corresponding audio encoding unit 27-J of FIG. 14 and transmitted to the corresponding terminal TM-J, the mixed audio of the two conferences is mixed in the reproduction unit of the terminal TM-J. Is played.

As the configuration of the conference selection part, instead of performing the conference selection corresponding to the terminal with the conference selection switches 7S-1 to 7S-Q shown in FIG. 15, the terminal correspondence branching unit 16 shown in FIG.
Is configured by a switch matrix whose input / output is logically 2Q × (2Q · N), and ON / OFF control of its contacts is performed by the signal processing control unit 20 based on the conference selection instruction from the terminal. It is also possible to provide only the audio signals of the conference selected by the terminal to the units 21-1 to 21-N. Third Embodiment In the first embodiment shown in FIGS. 8 and 11, the main speaker in a conference is extracted, the voice signal of this main speaker is assigned to the left channel, and the voices from other participants are assigned to the right channel. The case where signals are added and assigned is shown. The audio signals from the left and right channels are transmitted to the participating terminals at each point, and the sound is reproduced at each point by using one sound source for each channel. When this system is applied to communication between terminals at three or more points, voice signals from two points are simultaneously added in the right channel. In this case, there is a drawback that the listener cannot hear the sound image positions of the sounds from the two points separately for each point. Further, even if the main speaker is switched, the voice signal from each terminal is not always distributed to a fixed channel. Therefore, there is a drawback that the listener does not always hear the sound image position of each speaker in a constant manner. As a result, the identification of each speaker and the improvement of the intelligibility of the utterance content are limited. An embodiment in which this point is improved is shown in FIG.
Shown in

The embodiment shown in FIG. 16 is also based on the principle structure according to the present invention shown in FIG. The voice communication control device 100 according to the embodiment of FIG. 16 uses different sets of acoustic transfer functions for the voice signals of the N participants so that the reproduced voices of the N participants are localized at different spatial positions. Perform processing by using it as a control parameter. As a result, voice communication is possible by simultaneously locating the voices of N speakers at different positions in a communication conference between maximum N points. However, each terminal must use stereo headphones as the audio playback units 53L and 53R (Fig. 7). The voice signal of each one line is transmitted from the terminal at each point, and the voice communication control device 1
00 is input. On the other hand, a voice signal for each one line is output from the voice communication control device 100 and is returned to the terminal at each point. However, the output voice signal of each one line is a two-channel stereo voice signal synthesized in the voice communication control device 100 and multiplexed in one channel. Of course, voice signals of three or more channels may be multiplexed on one line.

Voice communication control apparatus 100 according to this embodiment
6, the channel branch points 3-1, ..., 3-N of the channel branch section 13 and the left / right signal processing section of the sound image control section 14 in FIG.
A set of 4-1L, 4-1R, 4-2L, 4-2R, ..., 4-NL, 4-NR corresponding to each terminal is set as a sound image processing unit 8-1, 8-2, ..., 8-N. It is shown as. FIG. 17 shows the one sound image processing unit 8-1 as a representative. The sound image processing unit 8-1 uses the principle described in FIG. 2 to convolve the acoustic transfer functions H _1L and H _1R with respect to the left and right audio signals branched at the channel branching point 3-1 by the convolution units 4-1L and 4-1R, respectively. The convolution operation is performed by. The audio signals obtained by the convolution are given to the adders 5L and 5R of the adding unit 15 of FIG. 16 as a left channel audio signal and a right channel audio signal. The transfer functions H _1L and H _1R that are convoluted with the branched audio signal of each channel can be determined in association with the desired spatial position where the reproduced audio of the audio signal is to be localized.

The switching unit 11 selects a communication line J (1≤J≤M) from an unspecified number of communication networks 40 that constitute the circuit network. However, M indicates the number of terminals connected simultaneously, and generally M ≦ N. The selected lines are connected to two channels for each terminal that simultaneously performs voice communication. Among them, one channel mediates the input of the audio signal in the example of the present invention and is connected to the decoding unit 23-J (J is 1, 2, ..., N). The other one channel mediates the output of the audio signal and is connected to the multiplexing / coding unit 22-J through the input channel C _J. Each decoding unit 23-J decodes the audio signal input from the connected terminal. The audio signal decoded by the decoding unit 23-J is supplied to the amplification factor setting unit 35 and the amplifier 36-J, respectively.

The signal processing control unit 20 changes from each terminal to the switching unit 1.
1 receives a control signal such as a connection confirmation transmitted via the control unit 1. The number M of connected terminals is detected from these control signals. In addition, the detected number M of connected terminals is set to the amplification factor setting unit 35, respectively.
And to the parameter setting unit 14C. Each amplifier 36-J
Respectively amplifies the input audio signal with an amplification factor G _J. The amplification factor G _J is determined by the amplification factor setting unit 35. As an example, the amplification factor G _J is set so that the short-time power integral value of the audio signal output from the amplifier 36-J becomes equal for each channel.

The parameter setting section 14C includes a sound image processing section 8-
At J, the acoustic transfer function H necessary for the processing for synthesizing the voice signals from the terminals TM-J at the respective points J to synthesize the voice signals for locating the reproduced sound at different target space positions θ _J , respectively.
Set _JL (θ _J ) and H _JR (θ _J ). Since there is a one-to-one correspondence between the target space position θ _J and the acoustic transfer functions H _JL (θ _J ) and H _JR (θ _J ), the target space position θ _J for each input signal.
Acoustic transfer function H convoluted with each audio signal if
_JL (θ _J ) and H _JR (θ _J ) can be determined. Here, the target spatial position θ _J for the voice transmitted from each terminal is determined based on the number M of connected terminals. As illustrated in the case of M = 5 in FIG. 18, the virtual space position θ _J passes from the left side (90 °) to the front side (0 °) on the horizontal plane and to the right side (−90 °).
Up to (°) at an equal angular interval Δθ = 180 / (M-1) degrees. The virtual space position θ _{J with} respect to the terminal TM-J at each point J is set to 90-180 (J-1) / (M-1) degrees depending on the number M of connected terminals. Therefore, the virtual space position interval Δθ becomes the smallest when the maximum number N of connectable points is used (M = N).

The sound image processing unit 8-J transfers the transfer signals H _JL (θ _J ) and H _JR (θ _J )
Is subjected to a convolution operation and applied to the adders 5L and 5R as a left channel audio signal and a right channel audio signal, respectively. If these left and right channel audio signals are reproduced by stereo headphones and listened to by both ears, the listener can localize the sound image at the virtual space position θ _J. Sound image processing unit 8-J
Left and right channel audio signals from D-JL, DJ
Also supplied to R.

The adder 5L adds all the left channel audio signals input from the sound image processing units 8-1 to 8-N and gives the obtained left channel added audio signal to the branch point 6L of the branch unit 16. The adder 5R adds all the right channel audio signals input from the sound image processing units 8-1 to 8-N and gives the obtained right channel added audio signal to the branch point 6R. The branch point 6L receives the added audio signal of the left channel input from the adder 5L by N
Branch to the subtractors 26-1L to 26-NL. The branch point 6R is
The right channel addition audio signal input from the adder 5R is set to N
Branch to the subtractors 26-1R to 26-NR.

On the other hand, each delay unit D-JL delays the input left channel audio signal by delay time τ _JL and outputs it to the subtractor 26-JL. The delay time τ _JL is the sum of the delay time related to the processing of the audio signal in the adder 5L and the delay time related to the processing of the audio signal at the branch point 6L. Therefore, the left channel audio signal output from the delay unit D-JL and the component given to the adder 5L from the sound image processing unit 8-J in the left channel added audio signal output from the branch point 6L have the same phase, Subtractors 26-JL cancel each other out. As a result, the voice signal component received from the terminal TM-J at the point J is canceled from the left channel addition voice signal branched and transmitted to the terminal TM-J, and echo can be prevented. Therefore, the only voice signal output from the subtractor 26-JL to be returned to the terminal TM-J is the added voice signal from the terminals other than TM-J. For the same reason, the delay unit D-JR delays the right channel audio signal input from the sound image processing unit 8-J with a delay time τ _JR and outputs it to the subtractor 26-JR. Delay time τ _JR is adder 5
The sum of the delay time related to the processing of the audio signal at R and the delay time related to the processing of the audio signal at branch point 6R.

Echo output from the subtracters 26-JL and 26-JR
The erased left and right channel audio signals are multiplexed and encoded.
Section 22-J, multiplexed with each other, coded and
It is transmitted from the conversion unit 11 to the terminal TM-J at the point J. like this
In addition, each multiplex / encoding unit 22-J has two left and right channel audio signals.
The signal is multiplexed into one channel and encoded. as a result,
Each multiplexed 1-channel audio signal is exchanged after encoding
One line for each point J (1 ≤ J ≤ M) by section 11
Transmitted in. Therefore, the transmission of 2 channel stereo signals
Avoiding the delay time difference between lines due to using two lines for transmission
And the number of lines can be saved. Multiplexed at each terminal
If the decoded audio signal is reproduced and the audio is reproduced,
For the listener, the voice from the terminal at another point can be sent to the desired virtual sky.
Position θ _JCan be localized to. As a result, each listener sends another
The person can be easily identified and conversation intelligibility is secured. Also, each place
Using sound image position processing means related to sound image localization at points
Eliminates the need and realizes an economical system.

Incidentally, unlike the above usage conditions, in the embodiment shown in FIG. 16, each point J (1 ≦ 1
J ≦ M), it is assumed that two communication lines are used for transmitting stereo audio signals of two channels.
In that case, one line is used for each of the left and right channel audio signals, and the switching unit 11 needs to switch three lines for each point J. Also, the multiplexing / encoding unit 22-J in FIG.
Also, although multiplexing and demultiplexing at each point are not required, two encoding units 23-JL and 23-JR are required instead of one multiplexing / encoding unit 22-J for each terminal. Also, it is necessary to connect subtractors 26-JL, 26-JR to the input sides of the coding units 23-JL, 23-JR, respectively, and input the speech signal to the two coding units 23-JL, 23-JR. is there.

As described above, according to the embodiment shown in FIG. 16, even if each terminal does not have a voice signal processing unit for the purpose of sound image localization, the listener can localize voices from other points to different positions. Can be heard. Therefore, it becomes easy for the listener at each point to identify the speaker, and a good conversation intelligibility can be secured. At the same time, it is not necessary to predefine the communication method.

As described above, even in the case of binaural listening using stereo headphones or the like, it is possible to economically realize voice communication between arbitrary multipoints in which the listener localizes the voices of the respective speakers at different positions. Become. Further, when the number M of connected terminals is smaller than the maximum number N of terminals that can be simultaneously connected, the sound image localization position interval of each speaker can be expanded. Fourth Embodiment By the way, as shown in FIG. 19, terminals TM-1 to
Consider a case where communication is performed between the TM-6s via the voice communication control device. Here, terminal combinations TM-1 to TM-3
It is assumed that the communication conference X and the communication conference Y are realized in each of the terminal combinations TM-3 to TM-6. At this time, the voices of the users of the terminals TM-1 and TM-2 are transmitted to the terminals TM-4 to TM-6.
Users cannot listen. On the contrary, the users of the terminals TM-1 and TM-2 cannot hear the voices of the users of the terminals TM-4 to TM-6.
In addition, the users of the terminal TM-3 are the terminals TM-1, TM-2, TM-4 to TM-6.
The voice of any of the users can be listened to, and the voice of the user of the terminal TM-3 can be listened to by any of the users of the terminals TM-1, TM-2, TM-4 to TM-6. By this method, it is possible to hide the communication contents from users who do not belong to the communication conference between the terminal combinations, and conversely to let users who belong to multiple simultaneous communication conferences grasp any communication contents. Various applications become possible. Furthermore, in this multipoint-to-point communication method, if the voice of each speaker is localized and listened to in different virtual space positions, the identification of each speaker becomes easier and the intelligibility of the utterance content is improved. Not only that, the communication is expected to be as natural as if the talker were present in the same space.

However, according to the voice communication control apparatus shown in the embodiments of FIGS. 14 and 15, the terminal TM-3 in FIG.
Can select both of the communication conferences X and Y to simultaneously receive, but can transmit only to the selected one of the communication conferences X and Y. Further, when both the conference calls X and Y are simultaneously received, the received voices of the conferences X and Y are reproduced separately from the left and right speakers, but from the plurality of terminals in the received voice signal of the conference X or Y. It is not possible to localize the voice of to different positions.

FIG. 20 shows the basic configuration of the communication control apparatus according to the fourth embodiment which has improved the above problems. The main configuration of the communication control device 100 according to the fourth embodiment is that the transfer function between the sound source and both ears is convoluted with the voice signal sent from the exchange unit 11 and each terminal TM-1 to TM-6. A sound image processing unit 8-J (J = 1,2, ..., N where N = 6) that performs a voice process for locating the sound source position of the speaker is assigned, and terminal combination allocation is performed for multiple conferences. Terminal terminal combination assigning section 19 to perform and adding / branching section 17-P (P = 1, ..., Q, here Q = 2)
And the inter-combination addition unit 12. Each adder / branch unit 17-P has left and right adders 5L and 5
Branch point 6L for branching R and their addition output,
It consists of 6R. Inter-combination inter-combination adder 12 includes N adders 2-JL, 2-JR (J = 1, 2, ..., N, here N = 6) for each of the left and right sides.
Consists of In addition, in order to identify one element from the same kind of constituent elements, it was distinguished by subscripts J (1 ≦ J ≦ N) and P (1 ≦ P ≦ Q). In addition, subscripts L and R are used to identify which of the left and right channels is the component that processes the audio signal.
Distinguished by.

The operation of the voice communication control device 100 according to the present embodiment will be described with the configuration shown in FIG. Exchange unit 1
Reference numeral 1 selects a communication line J (1≤J≤M) from an unspecified number of communication networks 40 constituting the circuit network. M indicates the total number of terminals connected simultaneously. Generally, M ≦ N. N is the maximum value that can be connected. The exchange unit 11 selects the communication line J based on control signals such as communication start / end, connection terminal designation, connection confirmation signal, etc. received from one terminal, and in this example, the sound image processing unit 8-through the input channel C _J. Connect to J. The configuration of each sound image processing unit 8-J is the same as that shown in FIG. 17, and corresponds to the combination of one branch point 3-J and the signal processing units 4-JL, 4-JR of the left and right channels in FIG. are doing.

Each sound image processing unit 8-J performs a process of convoluting the transfer function into the voice signal sent from each terminal TM-J and localizing the voice of the speaker of each terminal TM-J to the virtual space position. .. Therefore, the audio signal output from the sound image processing unit 8-J is a stereo audio signal. The stereo audio signals generated by the respective sound image processing units 8-J are input to the terminal combination allocation unit 19 and sorted according to the terminal combination. In the illustrated example, as shown in FIG. 19, terminals TM-1 to TM-3 and TM-3 to TM-6 belong to communication conferences X and Y, respectively.

The stereo audio signals divided into the communication conferences X and Y by the terminal combination allocating unit 19 are given to the adding / branching units 17-1 and 17-2, respectively, and the terminals belonging to the same communication are added by the adders 5L and 5R. The left and right channels are added together. The addition results of the X group (and the Y group) are respectively calculated by the left and right branch points 6L and 6R in the addition / division base unit 17-1.
The left-side adders 2-1L to 2-3L and the right-side 2-1R to inside the inter-combination addition unit 12 corresponding to all terminals TM-1 to TM-3 of the same set X, respectively.
Allocated to 2-3R. Similarly, the addition result of the Y group is stored in the inter-combination adding section 12 corresponding to all the Y terminals TM-3 to TM-6 of the same group by the left and right branch points 6L and 6R in the adding / branching section 17-2. To the left channel adders 2-3L to 2-6L and the right channel adders 2-3R to 2-6R. The adders 2-JL and 2-JR of each pair add the audio signals of all the communication conferences to which the pair belongs to the left and right channels to generate a stereo audio signal, and the stereo audio signal is supported via the exchange unit 11. Is transmitted to each of the terminals TM-1 to TM-6.

In this example, the audio signal transmitted to each terminal TM-1 to TM-6 is a stereo audio signal, and each terminal TM-1 to TM-6
The receiver of -6 receives the voice of the speaker of the other terminal as the sound image processing unit 8-J.
21A at the virtual space position determined by the transfer function convolved with
And 21B, it can be localized and listened.
That is, in the example of FIG. 20, the convolution operation of the transfer function is performed on each of the voice signals sent from the terminals TM-1 to TM-3,
As shown in FIG. 21A, the terminals are added and transmitted.
Only TM-1 to TM-3 speakers can listen. The communication conference of this set of terminals is called communication conference X. Also, the convolution operation of the transfer function is performed on the voice signals sent from the terminals TM-3 to TM-6, and the voice signals of the terminals TM-3 to TM-6 are added to add the voice signals to the terminals TM-3 to TM. -6, the listeners of the terminals TM-3 to TM-6 transfer the transfer functions obtained by convolving the voices of the speakers of the terminals TM-3 to TM-6 in the sound image processing unit as shown in FIG. 21B. It is possible to localize and listen to the virtual space position associated with.
A communication conference of this set of terminals is referred to as a communication conference Y. However,
Here, since the listener of the terminal TM-3 participates in both the communication conferences X and Y, the listener of the terminal TM-3 receives the terminal TM of both the communication conferences X and Y as shown in FIG. 21C. -1 to TM-6 can be localized and listened to in different virtual space positions. Fifth Embodiment An embodiment in which the voice communication control device 100 having the basic configuration shown in FIG. 20 is specifically configured will be described with reference to FIGS. 22 and 23. As the usage conditions of the embodiment shown in FIG. 22, it is assumed that communication between a plurality of terminals and the voice communication control apparatus 100 proposed in this embodiment is performed by transmitting a voice signal using one round-trip line for each terminal. In addition, this embodiment simultaneously controls a maximum of Q communication conferences among a maximum of N terminals.
Here, the digital voice signal of each channel is transmitted from each terminal and input to the voice communication control device 100. The voice communication control device 100 outputs a digital voice signal for each channel and transmits it to each terminal. However, the audio signal of each one channel that is output is the audio communication control device 10
Each of the two channels of stereo audio signals generated at 0 is multiplexed into one channel.

In FIG. 22, the exchange unit 11 and the decryption unit 23-J
(J = 1, ..., N), the signal processing control unit 20, the amplification factor setting unit 35,
Amplifier 36-J (J = 1, ..., N), parameter setting unit 14C, sound image processing unit 8-J (J = 1, ..., N), multiplexing / encoding unit 22-J (J = 1, ..., N)
The configuration and operation of N) are the same as the corresponding ones in the embodiment of FIG. 16, so description thereof will be omitted. The difference is that the terminal terminal terminal combination allocation unit 19, terminal selection control unit 9C, addition / branching unit 17-P (P = 1, ..., Q), intermittent combination determination unit 7C, intermittent unit 7-P
(P = 1, ..., Q), and the point that the inter-combination addition unit 12 is provided. The terminal / terminal combination assigning unit 19 has Q × N terminal selecting units 9 _P -J (P = 1, ..., Q; J = 1, ..., N), and the inter-combination combination adding unit 12 N pairs of adders 2-JL, 2-JR (J =
1, ..., N).

As shown in FIG. 23, each adder / branch unit 17-P has adders 5L and 5R, branch points 6L and 6R, and a delay unit DJ.
L, D-JR (J = 1, ..., N), Subtractor 26-JL, 26-JR (J = 1, ..., N)
N) and. The functions of the characteristic parts in the embodiment will be described below. The audio signal from the connected terminal is decoded by the decoding unit 23 in the same manner as described in the embodiment of FIG.
-J and the amplifier 36-J and is given to the sound image processing unit 8-J.

The signal processing control unit 20 changes from each terminal to the exchange unit 1.
1 receives control signals transmitted via 1 such as communication start / end, connection confirmation, and belonging of each connected terminal to a communication conference.
From these control signals, the number of connected terminals M, the connected terminal TM-1
~ Detects TM-n, start / end of communication, and belonging to communication between combinations of connection terminals. Further, the detected connection terminal and the number M of the connection terminals are sent to the amplification factor setting unit 35 and the parameter setting unit 14C, and the communication start / end is set to the terminal selection control unit 9C.
Also, the combination selection control unit 7C is notified to the terminal selection control unit 9C of belonging to the communication between the combinations of the connection terminals TM-1 to TM-n.

The parameter setting section 14C is used for each sound image processing section.
Playback sound for audio signals from each terminal TM-J at 8-J
For each terminal TM-J_JSet to
Sound transmission required for processing to generate a voice signal
Reaching function H_L(θ_J), H_R(θ_J) Is set. Virtual space position θ_J
And acoustic transfer function H_L(θ_J), H_R(θ_J) Is a one-to-one correspondence
Therefore, the virtual space position θ_JIf you decide
Reaching function H_L(θ_J), H_R(θ _J) Can be set. In this embodiment
Is the voice transmitted from each terminal based on the number M of connected terminals
Virtual space position θ with respect to_JTo determine. Illustrated in Figure 21C
So that the virtual space position θ_JTo the left of the horizontal plane (+90
Angle from the front (0 °) to the right (-90 °)
The interval is 180 / (M-1) degrees. In other words, for each point J
Virtual space position θ_JIs defined as 90-180 (J-1) / (M-1) degrees
You.

In the example shown in FIG. 21C, since M = 6,
Each virtual space position θ _J of each terminal TM-1 to TM-6 is θ _{J = 1} = 90 ° −180 ° × (1-1) / (6-1) = + 90 ° θ _{J = 2} = 90 ° −180 ° × (2-1) / (6-1) ＝ + 54 ° θ _{J = 3} ＝ 90 ° −180 ° × (3-1) / (6-1) ＝ + 18 ° θ _{J = 4} = 90 ° -180 ° × (4-1) / (6-1) =-18 ° θ _{J = 5} = 90 ° −180 ° × (5-1) / (6-1) =-54 ° θ _{J = 6} = 90 ° -180 ° x (6-1) / (6-1) = -90 °.

The sound image processing unit 8-J applies the audio signal given from the amplifier 36-J to the terminal TM in the parameter setting unit 14C.
The convolution operation is performed on each of the acoustic transfer functions H _L (θ _J ) and H _R (θ _J ) set according to -J to generate a stereo audio signal for each of the left and right channels. When a sound is reproduced from the generated stereo audio signal to be listened to by both ears, the listener hears the virtual space position θ.
_The sound image is localized at _J. The left and right channel audio signals from the sound image processing units 8-J are Q terminal selection units 9 ₁ -J, 9 ₂ -J, ..., 9 _Q -J.
Distributed to Terminal selection control unit 9C, each communication start is transmitted from the signal processing control unit 20 / Exit and based on belonging to a communication conference for each connection terminal, to determine the intermittent information, transmitted to the terminal selection unit 9 _P -J To do. For example, when the communication conference P to which the terminal TM-J belongs is started or ended, the terminal selection unit
9 Transfer control signals to _P- J to pass or block audio signals respectively. Terminal selection unit 9 _P -J is intermittently audio signal according to the control signal transferred from the terminal selection control unit 9C. Thereby, for example, in the terminal combination shown in FIG. 21C, the voice signals originating from the terminals TM-1 to TM-3 are added to the adding / branching unit 17-1 and the voice signals originating from the terminals TM-3 to TM-6. It is assigned to the adding / branching unit 17-2.

FIG. 2 shows the internal structure of the adder / branch unit 17-P.
3 will be described. The left and right channel audio signals input from each terminal selection unit 9 _P -J are given to the adders 5L and 5R, respectively, and also given to the delay units D-JL and D-JR, respectively. The adder 5L adds the input N left-channel audio signals and outputs the result to the branch point 6L. The adder 5R adds the input N right channel audio signals, and outputs a branch point 6R.
Output to The branch point 6L is the N-channel subtractor 26-JL (J = 1, ..., N) for the right channel addition audio signal input from the adder 5L.
Branch to. The branch point 6R is the N-channel subtractor 26-JR (J = 1, ..., N) for the right-channel added audio signal input from the adder 5R.
Branch to.

The delay unit D-JL delays the left channel audio signal input from the terminal selection unit 9 _P -J with a delay time τ _JL and outputs it to the subtractor 26-JL. However, the delay time τ _JL is the sum of the delay time related to the processing of the audio signal in the adder 5L and the delay time related to the processing of the audio signal at the branch point 6L. Therefore, the left channel audio signal output from the delay unit D-JL is synchronized with the left channel audio component output from the terminal selection unit 9 _P -J in the left channel added audio signal output from the branch point 6L. Delay time D-JR delay time τ
_{JR is} also determined in the same manner, and is output from the terminal selection unit 9 _P -J in the right channel audio signal output from the delay unit D-JR and the right channel addition audio signal output from the branch point 6R. Synchronize with the right channel component.

The subtractors 26-JL and 26-JR are provided with delay units D-JL from the audio signals supplied from the branch points 6L and 6R, respectively.
And the audio signals given from D-JR are respectively subtracted.
As a result, the same components as above are canceled out, and the addition / branching unit 17-P
In, in the channel corresponding to each terminal TM-J, the addition result of the audio signal derived from the other channel K (J ≠ K) is obtained. The added voice signal is output to the interruption unit 7-P in FIG. In other words, the voice signal from each terminal TM-J is
-It is not added to the audio signal transmitted to J. Therefore,
Reverberation due to the presence of the voice communication control device 100 of this embodiment can be suppressed.

Returning to the explanation of FIG. 22, the combination selection control unit 7C sends the communication conference P transmitted from the signal processing control unit 20.
The intermittent information is determined based on the start / end information of. This gating information is transferred to the gating section 7-P. For example, when starting or ending the communication conference P, a control signal instructing to pass or block the voice signal is transferred to the terminal selection unit 9 _P -J. The interruption unit 7-P is a combination selection control unit 7
In accordance with the control signal from C, the audio signal output from the adder / brancher 17-P, that is, the subtractors 26-JL and 26-JR is interrupted all together.

The adders 2-JL and 2-JR are connected from the J-th channel of each of the Q adder / branchers 17-P (P = 1, ..., Q) corresponding to the Q terminal combinations to the intermittent part. Add the left and right channel audio signals of the terminal combination P _S selected by 7-P (P = 1, ..., Q). The symbol P _S is the number of terminal combinations to be added to each other, and up to Q can be selected from the range of 0 ≦ P _S ≦ Q. As a result, the left and right channel audio corresponding to the selected terminal combination is added and transmitted to the terminal TM-J, so the terminal TM-J participates in multiple selected terminal combinations (multiple communication conferences). It is possible to localize and listen to the sound from all other terminals. Further, the voice transmitted from the terminal TM-J will be transmitted to all other terminals that have selected the terminal combination including the terminal TM-J.
Each multiplexing / encoding unit 22-J multiplexes and encodes the left and right channel audio signals input from the adding units 2-JL and 2-JR. That is, the multiplexing / encoding unit 22-J multiplexes and encodes the stereo audio signals corresponding to the left and right two channels into each one channel. As a result, the coded multiplexed voice signal of each channel is independently output to the switching unit 11 for each terminal TM-J, and is multiplexed to one channel toward each terminal TM-J (1 ≦ J ≦ M). The converted voice signal is transmitted by one line.

According to the embodiment of FIG. 22, for example, FIG.
As shown in C, even if a communication conference is being held at the terminals TM-1 to TM-6, the listener localizes the voice from each terminal at different positions at 36-degree intervals, and the terminal TM- 1
The communication conference X closed between TM3 and TM-3 and the communication conference Y closed between terminals TM-3 to TM-6 can be realized at the same time. At this time, the terminal
The listener in TM-3 can hear both voices not only from the terminals TM-1 to TM-2 but also from the terminals TM-4 to TM-6. Other,
Even if the communication conference in all terminals including the terminals TM-1 to TM-6 is in progress, the communication conference X between the terminals TM-1 to TM-3 can be realized. In the above, the terminal TM-3 in FIG. 21C is used.
As described above, the operation when a certain terminal participates in a plurality of communication conferences X and Y has been described. However, when participating in only one communication conference X, such as the terminal TM-1 in FIG. It suffices to pass the audio signal from only one interrupting unit corresponding to the communication conference X to the adding units 2-1L and 2-1R corresponding to TM-1.

As shown in FIG. 21C, in the discussion between specific speakers during the general conference, and in the application of voice communication between other terminals to monitoring for each communication conference, the listener transmits the speech at each terminal. The person's voice can be localized and heard in different virtual space positions. As a result, it becomes easy to identify each speaker, and the intelligibility of the utterance content is improved. According to the present embodiment, there is an economic advantage that it is not necessary to introduce a sound image processing unit for realizing sound image localization to a virtual space position in each terminal TM-J. From the above effects,
Economically possible natural communication equivalent to when all participants are present in the same space.

Thus, in the embodiment of FIG. 22, Q
Q addition / branching units 17-1 to 17-Q corresponding to one communication conference
Is provided, and when the signal processing control unit 20 receives from each terminal TM-J a control signal designating one or more communication conferences in which the user of the terminal intends to participate (speak), the control signal To the terminal selection control unit 9C. The terminal selection control unit 9C, among the Q number of terminal selection units 9P-J (P = 1, ..., Q) for the voice signal from the terminal TM-J, performs addition / correspondence corresponding to the communication conference designated by the control signal. 1 for branch 17-P
One or a plurality of terminal selection units 9P-J are brought into conduction. As a result, the voice signal from the terminal TM-J can be connected to one or more communication conferences requested by the terminal.
Users can send audio to the teleconference. Further, in the embodiment shown in FIG. 22, there are Q addition / branching units 17-1 to 17-Q.
Is provided with Q connecting / disconnecting units 7-1 to 7-Q, and a control signal for designating one or a plurality of communication conferences to be heard by a user of the terminal from each terminal TM-J. Is received by the signal processing control unit 20, the control signal is given to the combination selection control unit 7C. The combination selection control unit 7C turns on and off the interrupting unit 7-P corresponding to the one or more adding / branching units designated by the control signal, so that the contents of the designated one or more communication conferences are changed. It is sent to the terminal TM-J.

Therefore, each terminal TM-J sends a control signal to the voice communication control device of the present invention at any time,
It is possible to change, add, delete, and leave the communication to participate (speak), and to change, add, delete, and leave the communication (conference) to be heard. The voice communication control device 10 of FIG.
At 0, instead of sending each two-channel audio signal (stereo audio signal) to each terminal TM-J by one line, it may be sent by using two communication lines. In that case, one line is used for each voice signal of each channel, and the switching unit 11 needs to switch three lines for each terminal TM-J for input and output. In this case, multiplexing in the multiplexing / encoding unit 22-J and demultiplexing in each terminal TM-J are unnecessary. However, in the case of such a configuration, two multiplexing / encoding units 22-J are required for the left and right channels, which causes a disadvantage of complicating the configuration. Sixth Embodiment FIG. 24 shows a principle configuration of a modification of the embodiment shown in the principle configuration of FIG. The usage conditions are the same as those in the embodiment shown in FIGS. In the voice communication control device of the embodiment shown in FIG. 24, the terminal / terminal combination assigning unit 19 is used as the sound image processing unit 8-J.
20 is different from the embodiment of FIG.
The other structure is the same as that of the embodiment of FIG. By providing the terminal / terminal combination assigning unit 19 on the upstream side of the sound image processing unit 8-J, audio signal processing for sound image localization is performed after the combination between connected terminals is determined. Therefore, in FIG.
26B, each terminal TM for each communication conference X or Y
It is possible to freely set the sound image position of -1 to TM-3 or TM-3 to TM-6. Seventh Embodiment FIG. 25 shows a specific example of the configuration of the example shown in the principle configuration of FIG. The parts corresponding to those in FIG. 22 are designated by the same reference numerals. Most of the configuration and functions of this embodiment are similar to the embodiment of FIG. Also in this embodiment, each terminal can participate in a plurality of communication conferences at the same time, and the listener at each terminal can perform arbitrary multipoint voice communication in which the voices of the talkers of other terminals are localized at different positions. become. Furthermore, it is the same as the embodiment of FIG. 22 in that it is not necessary to introduce a sound image processing unit for realizing sound image localization to the virtual space position in each terminal TM-J or each terminal combination P.
The difference from the embodiment of FIG. 22 will be mainly described. Each part will be described below.

The signal processing control unit 20 operates from each point to the exchange unit 1.
Control signals such as communication start / end, connection confirmation, belonging to the communication conference P between the combinations of the connection terminals TM-J, etc. transmitted via 1 are received. From these control signals, the connection terminal TM-J, the number M of connection terminals, the communication start / end, the affiliation to the communication conference P between the combinations of connection terminals, and the number of terminals belonging to each communication conference P are detected. Further, the detected connected terminals and the number of connected terminals M are sent to the amplification factor setting unit 35, the communication start / end is sent to the terminal selection control unit 9C and the combination determination unit 7C, and each connected terminal TM
-J belongs to the communication conference P to the terminal selection control unit 9C, and the number of terminals M _P belonging to each communication conference P to the parameter setting unit 14C
Communicate to.

The parameter setting unit 14C is used to combine each terminal.
All terminals in each combination P for each combination P_P-Sound from P
These reproduced sounds are different in virtual space position with respect to the voice signal.
Place θ _PJAcoustic transfer function H_L(θ_PJ), H_R(θ_PJ)
Each is set in the sound image processing unit 8P-J. Where virtual space
Position θ_PJ, Then the acoustic transfer function H_L(θ_PJ), H
_R(θ_PJ) Can be set. In this example, the signal processing
A terminal belonging to the communication conference P detected by the control unit 20
Number M_PBased on the
Idea space position θ_PJTo determine. As illustrated in Figure 21C
, The virtual space position θ_PJOn the horizontal plane from the left side (90 °)
180 / (M
_P-1) Determine in degrees. At this time, terminals belonging to the communication conference P
Number for TM-J in order of J_P(1≤J_P≤ M_P),
The virtual space position θ_PJ90-180 (J_P-1) / (M-1) degree
Is determined.

The one-channel audio signal originating from each terminal TM-J is distributed to the Q terminal selecting units 9 _P -J (P = 1, ..., Q). Each terminal selection unit 9 _P -J is intermittently audio signal of each 1-channel in accordance with a control signal from the terminal selection control unit 9C, the terminal selecting unit 9 _P
The audio signal passing through -J is given to the corresponding sound image processing unit 8 _P -J. The sound image processing units 8 _P -J respectively apply the acoustic transfer functions H _L (θ _PJ ), H _R (θ _PJ ) set in the parameter setting unit 14 C to the audio signals output from the terminal selection unit 9 _P -J. The convolution operation is performed to obtain a two-channel audio signal, which is given to the adding / branching unit 17-P. There are N sound image processing units 8 _P -J for each terminal group P (J = 1,
, N), whereas in FIG. 22, only N sound image processing units 8-J are provided. Also, each terminal selecting unit 9 _P -J of FIG. 22 is different from the embodiment of FIG. 25 in that two-channel audio signals are intermittently connected. The configuration and operation of each adding / branching unit 17-P in FIG. 25 are exactly the same as those shown in FIG.

The embodiment of FIG. 25 and the embodiment of FIG. 22 differ in the order of processing for audio signals. In the embodiment of FIG. 25, one-channel audio signal from each terminal TM-J is grouped for each terminal combination P (P = 1, ..., Q) and then independently for each communication conference P. A two-channel audio signal for the purpose of sound image localization is generated. Therefore, for each communication conference P, it is allowed to set a different virtual space position θ _PJ to each terminal TM-J belonging to the communication conference independently. That is, in the parameter setting unit 14C, the virtual space position θ _PJ that is different for each terminal TM-J belonging to each communication conference P is given to the _listener.
The acoustic transfer functions H _L (θ _PJ ) and H _R (θ _PJ ) can be set as sound image control parameters for localizing the voice.

Here, as a method of expanding the virtual space position interval for the voice from each terminal TM-J in each communication conference P, a method of determining from the number M _{P of} terminals belonging to the communication conference P will be shown. As an example, it is applied to the terminal combinations shown in FIGS. 26A and 26B. As shown in FIG. 26A,
Since the communication conference X is held between the three terminals, the virtual space position interval is 90 °. The virtual space positions for terminals TM-1, TM-2 and TM-3 are distributed to the left (90 °), front (0 °) and right (-90 °) in order. The communication conference Y is held between the four terminals, and the virtual space position interval is 60 °. Terminal TM-3, TM
The virtual space positions for -4, TM-5, and TM-6 are to the left (90
), Left front (30 °), left front (-30 °), right side (-90 °).

For comparison, consider the case where the virtual space position is determined using the total number M of connected terminals. The total number of connected terminals M is 6
Therefore, the virtual space position interval is 36 °, and FIG.
As shown in C, the virtual space positions for the terminals TM-1 to TM-6 are sequentially left side (90 °), left rear (54 °), left front (18).
°), right front (-18 °), right front (-54 °), right side (-90 °)
Distributed in. 21A and 21B, since the terminal combination allocation is performed after performing the processing for sound localization, it is not possible to set the virtual space position independently for each communication conference P. Therefore, the localization position for the voice from one terminal is constant regardless of the combination of terminals. At this time,
As shown in FIGS. 21A and 21B, for the communication conference X including the terminals TM-1 to TM-3, the localization position distribution is to the left (+90).
Only from (°) to the front left (+ 18 °). For the communication conference Y consisting of the terminals TM-3 to TM-6, the localization position distribution is left front (+18
From (°) to the right (-90 °).

As described above, in the embodiments of FIGS. 24 and 25, the localization position can be freely set for each communication conference between the terminal combinations. Further, by setting the localization position according to the number of terminals M _P belonging to each terminal combination (that is, communication conference) P, the distribution and interval of the localization positions with respect to the voice from each terminal can be expanded as compared with the embodiments of FIGS. 20 and 22. . When the listener listens to the sound from each terminal, FIG.
It becomes easier to identify each talker than the embodiment of
The intelligibility of the utterance content is further improved.

According to the embodiment shown in FIG. 25, when the number of terminals participating in an arbitrary communication conference changes, the sound image localization position for the terminals participating in the communication conference can be changed. In this case, the signal processing control unit 20 of the voice communication control device 100 predetermines the sound image positions for all the participating terminals of the communication conference in accordance with the number of each participating terminal, and the placement model of the sound image localization position for each terminal (acoustic transmission). Function H _L (J), H
_R (J) group). That is, when the number M of participants changes due to a request to leave the conference or a request to participate, the virtual space position to be assigned to each participant is newly determined by referring to the placement model corresponding to the updated number of participants. And correspondingly the corresponding set of acoustic transfer functions
H _L (J), H _R and (J) respectively selected, set for each sound image processing unit 8-J. It is also possible to predetermine possible positions corresponding to the number of participants as an initial procedure of the communication conference and allow the participants to select the allocation of participants to those positions.

In each of the embodiments shown in FIGS. 22 and 25, when the sound image control parameter is set by the parameter setting unit 14C as the initial setting of an arbitrary communication conference, the number of terminals participating in the communication conference based on the participation request signal from the terminal. When M is detected, the signal processing control unit 20 calculates the virtual space positions that can be assigned to these participating terminals. In the above description, the parameter setting unit 14C determines to which participant each of the obtained virtual space positions is to be assigned. However, while the communication conference is being performed, all the participants are assigned to themselves. It is also possible to change the assigned current virtual space position to another desired virtual space position. For example, the voice communication control device 100 transmits the virtual space position information initially determined for all the participants of a communication conference to all the participating terminals in advance. When a terminal user wants to change his or her current position to a desired position during a communication conference, a change request signal indicating the desired position is transmitted to the voice communication control device 100. Based on the received position change request signal, the signal processing control unit 20 of the voice communication control device 100 exchanges, for example, the current participant allocations for the requesting current position and the desired position, and notifies all participants of the new allocation information. To do.

In the embodiment of FIGS. 8 and 14 described above,
The switches SW-1 to SW-N are inserted in series in each channel on the output side of the voice detection processing unit 23B as shown by the broken line, and the voice detection processing unit 23B determines whether each channel is in the utterance state, Turn on the switch SW-J for that channel only during the utterance period.
Noise input to the channel may be blocked by setting N and turning OFF in other cases. The detection of the utterance state can be determined by whether or not the short-time power integrated value of the audio signal is equal to or more than the threshold value E _ON , as described with reference to FIG. Similarly, in each of the embodiments of FIGS. 16, 22, and 25, the switches SW-1 to SW-N are inserted on the output side of the decoding units 23-1 to 23-N as shown by broken lines to set the amplification factor. Part 35
Determines whether or not it is in the utterance state from the audio signal of each channel,
The switch may be turned on only during the utterance state.

[0091]

As described above in detail, according to the voice communication control device of the present invention, the voice signal from each terminal is branched into a plurality of channels, and the branch voice signal from each terminal is branched for each branch channel. Since the addition audio signals of multiple channels are added and the added audio signals of the multiple channels are branched and transmitted to each terminal, at least one person participates in the communication conference without special sound image localization processing at each terminal. The person's sound image can be reproduced by distinguishing it from the sound images of the voices of other participants.

[Brief description of drawings]

FIG. 1 is a diagram for explaining an acoustic transfer function for the purpose of sound image localization.

FIG. 2 is a diagram for explaining an example of audio signal processing for the purpose of sound image localization.

FIG. 3 is a block diagram showing a configuration example of a conventional multipoint voice communication terminal.

FIG. 4 is a block diagram showing a configuration example of a communication line network in conventional multipoint voice communication.

FIG. 5 is a block diagram showing an example of accommodating a communication line in which the voice communication control device 100 of the present invention is used.

FIG. 6 is a block diagram showing a basic configuration of a voice communication control device of the present invention.

7 is a block diagram showing a configuration example of each terminal used in the system of FIG.

FIG. 8 is a block diagram showing a voice communication control device according to a first embodiment of the present invention.

FIG. 9 is a diagram for explaining a message identification method.

10 is a block diagram showing a configuration example of an audio signal processing section 25 in FIG.

11 is a block diagram showing another configuration example of the audio signal processing unit 25 in FIG.

FIG. 12 is a time chart for explaining a first main speaker identifying method and an operation example of FIG. 11.

13 is a time chart for explaining a second main speaker identifying method and an operation example of FIG. 11.

FIG. 14 is a block diagram showing a voice communication control device according to a second embodiment of the present invention.

15 is a block diagram showing a configuration example of a conference selection unit in FIG.

FIG. 16 is a block diagram showing a voice communication control device according to a third embodiment of the present invention.

17 is a block diagram showing a configuration example of a sound image processing unit 8-1 in FIG.

FIG. 18 is a diagram for explaining a sound image localization position.

FIG. 19 is a diagram for explaining a combination of terminals belonging to each of a plurality of communication conferences.

FIG. 20 is a block diagram showing a voice communication control device according to a fourth embodiment of the present invention.

21A is a diagram showing a sound image localization position example in one communication conference according to the embodiment of FIG. 20, B is a diagram showing a sound image localization position example in another communication conference according to the embodiment of FIG. 20, and FIG. The figure which shows the sound image localization position at the time of participating in two communication conferences by the Example of 20.

22 is a block diagram of a fifth embodiment showing a specific configuration example of the embodiment of FIG.

23 is a block diagram showing a configuration example of each adding / branching unit 17-P in FIG. 22.

FIG. 24 is a block diagram of a sixth embodiment showing a modification of FIG. 20.

FIG. 25 is a block diagram of a seventh embodiment showing a specific configuration example of the embodiment of FIG. 24.

26A is a diagram showing an example of a sound image localization position possible in one communication conference according to the embodiments of FIGS. 24 and 25, and FIG.
The figure which shows the example of a sound image localization position possible in another communication conference by the Example of 4 and 25.

 ─────────────────────────────────────────────────── ─── Continued Front Page (72) Inventor Nobuo Hayashi 1-1-6 Uchisaiwaicho, Chiyoda-ku, Tokyo Nihon Telegraph and Telephone Corporation

Claims (28)

[Claims] 1. A voice communication control device for connecting a communication conference with at least three terminals through a communication network, and a switching section for exchanging a voice signal received from the terminals via the communication network. , N input channels which are connected to the exchange unit and into which audio signals from the respective terminals are input, and N is an integer of 3 or more, and the input audio signals from the N input channels are respectively input. A channel branching unit for branching into branch audio signals of K branch channels, and K is an integer of 2 or more, each of which has a predetermined type for the branch audio signals of the channels corresponding to the N input channels. A sound image control unit that performs processing using N parameter groups that respectively include K sound image control parameters, and that generates a sound image control audio signal for each K branch channel; At least one of the data groups differs from the other parameter groups in accordance with the target position of the sound image, and the sound image control audio signal of the K branch channel corresponding to each of the N input channels is output for each branch channel. And an addition unit for generating a K channel added voice signal, and a terminal corresponding branching unit that distributes the K channel added voice signal corresponding to each of the terminals and supplies the divided voice signal to the exchange unit. 2. The voice communication control device according to claim 1,
Input voice signals that are inserted between the N input channels and the channel branching unit and are added to each other from the input voice signals input to the N input channels through the exchanging unit are selected and output. It includes a speaker selecting unit and N selected audio channels for giving the input audio signals selected and output by the speaker selecting unit to the channel branching unit, and input to the N input channels through the exchanging unit. Output the input speech signal to be processed by the at least one parameter group from each said terminal to a predetermined one of said N selected speech channels, and the other said input signal to the remaining one. A signal processing control unit that controls the speaker selection unit to output to the selected voice channel. 3. The voice communication control device according to claim 1, wherein the voice signal of the main speaker and the voice signals of other speakers are assigned the highest priority and the other ones of the input voice signals of the respective input channels. The channel branching unit has a branching number K of 2, and the channel branching unit converts the input audio signal given from each input channel into a two-channel audio signal. The audio image control unit processes the branched audio signals of the two channels corresponding to the audio signal of the main speaker so that they are in phase with each other, and outputs the branched audio signals of the K-branched channel. It includes a phase control unit that processes the two-channel branched voice signals corresponding to the voice signal of the speaker so that they have opposite phases. 4. The voice communication control device according to claim 2,
The number of branches K of the channel branching unit is 2, and the signal processing control unit has the N units from the respective terminals.
From the individual input voice signals, the one with the highest priority and the other one are discriminated as the voice signal of the main speaker and the voice signal of other speakers, and the speaker selection unit determines the discrimination result by the signal processing control unit. According to the above, the main speaker's voice signal and the other speaker's voice signals are output to the predetermined one of the N selected voice channels and the other ones, respectively, and given to the channel branching unit. The branching unit branches the given N input audio signals respectively and outputs the first and second branch channel audio signals as the branched audio signals of the K branch channel, and the sound image control unit is the signal processing control unit. So that the first and second branch channel voice signals corresponding to the voice signal of the main speaker from the predetermined one selected voice channel are in phase with each other under the control of And a phase control unit that sets the first and second branch channel audio signals corresponding to the audio signals of the other speakers from the other selected audio channels so as to have opposite phases to each other. . 5. The voice communication control device according to claim 3 or 4, wherein the sound image control unit compares with the level of the voice signal of the main speaker according to the control of the signal processing control unit, and the voices of the other speakers. An attenuator for reducing the signal level is included. 6. The voice communication control device according to claim 1, wherein the number of branches K of the channel branching unit is 2, and the channel branching unit branches each of the given N input voice signals into a first branch. And the second branch channel audio signal to the above K
The signal processing control unit outputs the branched voice signal of the branch channel, and the signal processing control unit outputs the N input voice signals having the highest priority among the N input voice signals of the respective input channels and those other than the voice signal of the main speaker and the others. , And the signal processing control unit provides a sufficiently large amount of attenuation to the second channel branch audio signal compared to the first channel branch audio signal corresponding to the main speaker audio signal. The parameter section is provided to the sound image control section so as to give a sufficiently large amount of attenuation to the first channel branch audio signal as compared with the second channel branch audio signal corresponding to the respective voice signals of the other speakers. Set to. 7. The voice communication control device according to claim 2,
The number of branches K of the channel branching unit is 2, and the signal processing control unit determines from the N input voice signals from the respective input channels as the voice signal of the main speaker and the voice signal of other speakers. The speaker selecting unit outputs the main speaker's voice signal and the other speaker's voice signals to the predetermined one of the plurality of selected voice channels and the other, respectively, and respectively outputs the channel branching unit. The channel branching unit branches the given N input voice signals, respectively, and outputs first and second branch channel voice signals as branch voice signals of the K branch channel. Is assigned to the second channel branch audio signal as compared to the first channel branch audio signal corresponding to the main speaker's voice signal from the predetermined one selected voice channel. The sound image control of the parameter group so as to give a sufficiently large amount of attenuation and to give a sufficiently large amount of attenuation to the first channel branch audio signal as compared with the second channel branch audio signal from the other selected audio channels. Set to the department. 8. The voice communication control device according to claim 2,
Monitoring the level of the input voice signal through the input channel corresponding to each of the terminals from the exchange,
A voice detection processing unit that detects the presence or absence of speech of the terminal based on the monitoring result is provided, and the signal processing control unit determines the main speaker based on the presence or absence of the speech detected by the voice detection processing unit. The speaker selection unit is controlled based on the determination result. 9. The voice communication control device according to claim 3, wherein the level of the input voice signal passing through the input channel corresponding to each of the terminals from the switching unit is monitored, and based on the monitoring result. A voice detection processing unit that detects the presence or absence of speech of the terminal is provided, and the signal processing control unit determines the input channel of the main speaker based on the presence or absence of the speech detected by the voice detection processing unit, and makes the determination. The parameter group used for the signal processing of the sound image control unit is updated based on the result. 10. The voice communication control device according to claim 2, wherein Q number of voice signal processing units including the channel branching unit, the sound image control unit, and the adding unit are provided, and Q is 2 or more. And the terminal-corresponding branch unit branches the K channel added voice signals from the Q audio signal processing units respectively corresponding to N terminal units, and the terminal-corresponding branch unit unit From the Q groups of correspondingly divided K channel audio signals, one or a plurality of groups are selected under the control of the signal processing control unit, added for each channel, and output as one set of K channel audio signals. Of the input voice signal from the terminal to be joined according to the control based on the conference participation request signal of the signal processing control unit. Output to the selected audio channel corresponding to the audio signal processing unit. 11. The voice communication control device according to claim 10, wherein each of the conference selection units has one of the K channels from one or a plurality of the voice signal processing units designated by a participation request signal from a corresponding terminal. The added audio signal is added for each channel and output as the added audio signal of the K channel distributed to the corresponding terminal. 12. The voice communication control device according to claim 1, wherein a set of a corresponding one of the destination spatial positions of the sound sources different for each of the terminals is provided for the branch channel voice signal of the K channel corresponding to each of the terminals. The sound image control unit has a signal processing control unit that determines a sound image control parameter, and the sound image control unit uses the determined set of sound image control parameters to operate on the corresponding branch channel audio signal of the K channel, thereby each terminal. The K channel sound image control audio signal is generated for each time. 13. The voice communication control device according to claim 10, wherein the number of branches of the channel branching unit is 2, and each set of the sound image control parameters is a pair of acoustic transfer functions. 14. The voice communication control device according to claim 10, 11, 12 or 13, wherein the signal processing control unit detects a participation request signal from the terminal to participate in the communication conference to thereby participate in the communication conference. Number, determine the destination space position for each participating terminal based on the number of participating terminals, determine the sound image control parameter corresponding to each of the determined destination space position to the sound image control unit. give. 15. The voice communication control device according to claim 14, wherein the signal processing control unit detects the number N of the terminals connected by the switching unit, and symmetrically determines different destination space positions for each of the terminals. (N-1) degree interval is decided. 16. The voice communication control device according to claim 1, further comprising:
A subtraction unit that cancels the component of the K-channel sound image control audio signal corresponding to the terminal from the sound image control unit, from the K channel added voice signal distributed by the terminal corresponding to each terminal. including. 17. The voice communication control device according to claim 1, further comprising:
It includes a multiplexing unit that multiplexes the added voice signals of the K channels distributed by the terminal corresponding branch unit for each terminal into one channel and gives the multiplexed signal to the switching unit. 18. The voice communication control device according to claim 1, wherein the number of sets of the adder unit and the terminal corresponding branch unit is provided, and Q is an integer of 2 or more, and the voice communication control device further includes: A terminal combination assigning unit for giving the K-channel sound image control audio signal corresponding to each terminal from the sound image control unit to one or more designated addition units;
And an inter-combination adding unit that adds the K channel added voice signals respectively distributed from one or a plurality of the terminal corresponding branch units and outputs the added voice signals to the corresponding terminals. 19. The voice communication control device according to claim 1, wherein the channel branching unit, the sound image control unit, the adding unit, and the terminal-corresponding branching unit are provided as Q sets, and Q sets are provided.
Is an integer of 2 or more, and the voice communication control device further includes
A terminal combination allocating unit that supplies the input voice signal from each of the terminals connected by the switching unit to one or more designated channel branching units, and one or more designated terminal-corresponding branches. K distributed from each department
And an inter-combination adding unit for adding the added voice signals of the channels corresponding to the channels and outputting to the corresponding terminals. 20. The voice communication control device according to claim 18, wherein the K channel is one channel for each of the left and right channels, and the sound image control section is provided with different sound sources for branching voice signals of the left and right channels corresponding to each of the terminals. By performing a convolution operation of a pair of left and right acoustic transfer functions corresponding to the target space position as the sound image control parameter, a stereo audio signal for each left and right channel is generated as the sound image control audio signal for each terminal. 21. The voice communication control device according to claim 20, further detecting a participation request signal to the communication conference of each of the terminals to obtain the total number of terminals participating in the Q communication conferences.
Signal processing control for determining the number of the target space positions corresponding to the number of all the participating terminals, determining a set of acoustic transfer functions corresponding to the target space positions as the sound image control parameters, and giving the set to the sound image control unit. Including parts. 22. The voice communication control device according to claim 14, wherein the signal processing control unit determines the object based on a new number of participants every time the number of participants changes such as a participation request or cancellation of participation in the communication conference. The spatial position is updated, the transfer function is updated according to the updated target spatial position, and the updated transfer function is set in the sound image control unit. 23. In the voice communication control device according to claim 21, wherein the signal processing control unit represents the number of terminals participating in all communication conferences as M, the destination space position is symmetrical to 180 / (M-1) degree interval is decided. 24. The voice communication control device according to claim 19, wherein the plurality of channels are each one left and right channel, and the sound image control unit corresponding to each communication conference is a branch voice of the left and right channels corresponding to each of the terminals. By performing convolution calculation of a pair of left and right acoustic transfer functions corresponding to different target spatial positions of the sound source with respect to the signals as the sound image control parameters, the stereo image signals of the left and right channels are controlled for each terminal. Generate as an audio signal. 25. The voice communication control device according to claim 24, further calculates the number of participating terminals of each of the communication conferences based on a participation request signal from each of the terminals to each communication conference, It includes a signal processing control unit that determines each of the acoustic transfer functions of a pair corresponding to the number of target space positions corresponding to the number of participating terminals and sets the acoustic transfer functions in the sound image control unit corresponding to the communication conference. 26. The voice communication control device according to claim 25, wherein the signal processing control unit, when the number of participating terminals of each of the communication conferences is represented by M _P , makes the target space position 180 / symmetrically.
(M _P -1) degree interval. 27. The voice communication control device according to claim 25, wherein the signal processing control unit causes a change in the number of terminal participations in a communication conference such as a new participation request or cancellation of participation during any communication conference. Each time it occurs, the destination space position of the communication conference is updated based on the new number of terminal participation, and the acoustic transfer function is updated corresponding to the updated destination space position and set in the corresponding sound image control unit. To do. 28. The voice communication control device according to claim 1, further comprising:
A plurality of switches that are inserted in series in each of the input channels to pass or block the input voice signal, and whether or not the corresponding terminal is in the uttering state from the input voice signal on each of the input channels. The input voice signal is passed to the switch of the input channel determined to be in the utterance state, and the input voice signal is blocked to the switch of the input channel determined to be not in the utterance state. And a voice detector.