JPH1188513A - Voice processing unit for inter-multi-point communication controller - Google Patents

Abstract

PROBLEM TO BE SOLVED: To realize a clear voice conference and extension of the number of channels in the voice processing unit for an inter-multi-point communication controller that connects video conference terminals installed at plural points to realize inter-multi-point video conference. SOLUTION: Since a voice detection section 4 silences or decreases a gain of voice data from a video conference terminal that is discriminated as a non first-come channel by a first-come channel discrimination section 5 and provides an output of the result to a voice adder section 6a, the voice adder section 6 is not required to select a channel for voice addition processing but has only to conduct adder processing for voice data of all the channels at all times. Thus, the processing is simplified and the processing capability of a processor has a margin, then the number of handled channels is extended.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an audio processing device of a multipoint communication control device for realizing a multipoint videoconference using three or more videoconference terminals.

[0002]

2. Description of the Related Art FIG. 5 is a block diagram of a conventional voice processing device of a multipoint communication control device disclosed in Japanese Patent Application Laid-Open No. Hei 4-84553, wherein reference numerals 1a to 3k denote a voice encoding / decoding unit, Reference numeral 21 denotes a voice detection unit including voice level detectors 21a to 21k, 22 denotes a priority selection unit, 6 denotes a voice addition unit, and 7 denotes a control unit.

[0003] Audio data from a video conference terminal (not shown) is transmitted via an audio encoding / decoding unit 1a through a network.
1k, and in the voice encoding / decoding units 1a to 1k,
The input audio data is decoded into digital PCM data and output to the audio detection unit 21 and the priority selection unit 22. The audio detector 21 includes audio level detectors 21a to 21a.
The sound part and the silent part of the sound are determined for each channel by k, and the result of the determination is output to the control unit 7 as the accompanying data of the sound data. The accompanying data includes a channel number, a time at which a sound part of the sound corresponding to the channel is recognized, sound information of the sound, and the like.

Next, the operation will be described. In FIG. 5, when a multipoint video conference is held, voice data from each video conference terminal is received by the voice encoding / decoding units 1a to 1k via the network, and after being decoded into digital PCM data, voice detection is performed. It is transmitted to the section 21 and the priority selecting section 22. The voice detection unit 21 monitors a voiced portion and a voiceless portion of the voice at regular time intervals, and determines whether the voiced portion and the voiceless portion of the voice remain in the same state even after the monitoring time has elapsed. Is determined.

The control unit 7, having received the presence / absence of voiced sound and the time of recognition from the voice detecting unit 21, first sets the number of channels in which the voiced portion of the voice was detected to N in advance.
It is determined whether the number is equal to or less than N (N is a positive integer). If the number is equal to or less than N, the priority selecting unit 22 is controlled, and only the channel in which the sound portion of the sound is detected is connected to the sound adding unit 6. Also,
If the number exceeds N, the time at which the sound portion of the sound of each channel is recognized is determined in chronological order, and the N channels are selected from the earliest one in accordance with the order in which the sounds are generated.
2 to connect the selected N channels to the voice adder 6. It should be noted that the channel on which the sound portion of the voice is detected is N
If the number is less than or equal to the number, the audio information is unconditionally added to the audio adding unit 6.
Output to

The speech adder 6 synthesizes the selected speech data of the maximum N channels and composes the speech encoders / decoders 1a to 1k.
The audio encoding / decoding units 1a to 1k encode the received data using the same encoding method as the received data, and transmit the encoded data to each video conference terminal via a network.

[0007]

Since the conventional voice processing device of the multipoint communication control device is configured as described above, the priority selection unit is controlled by the number of speaking terminals and the set value of N. However, the number of audio channels to be added, which is designated by the above, changes, the processing of the audio adding unit becomes complicated, and the number of lines to be added is limited.

SUMMARY OF THE INVENTION The present invention has been made to solve the above-described problems, and has a scalable multipoint communication control by simplifying the addition processing of a voice addition unit and reducing the restriction on the number of channels to be added. An object of the present invention is to obtain an audio processing device.

[0009]

SUMMARY OF THE INVENTION In view of the above-mentioned object, the present invention is an audio processing device of a multipoint communication control device for realizing a multipoint videoconference using videoconference terminals at three or more points, The audio data is decoded into digital data for each channel for receiving the audio data from each of the video conference terminals, and the final audio data subjected to audio processing is encoded by the same encoding method as the received data. The audio encoding / decoding means to be sent to the video conference terminal respectively, and the audio data of each channel decoded from the audio encoding / decoding means are discriminated at a given time interval to determine whether the audio is a voiced part or a silent part. And the data corresponding to the decoded audio data from the audio encoding / decoding means and the data when the decoded data is a silent part for each channel. A voice detection means for generating or Re, based on the identification result of the determination in the voice detecting means, the number of channels which sound section is detected more than N (where N is a positive integer)
In the case of (1), decoded speech data from the speech encoding / decoding means is generated for all the channels, and if the number of channels where the speech portion is detected exceeds N, a speech portion is recognized. For the N channels from the earliest in the order, generate the decoded audio data,
For the remaining channels, the first-arrival channel determination means for controlling the voice detection means so as to generate data corresponding to the silence part, and the voice data generated from the voice detection means and the data corresponding to the silence part are all added. Then, the audio data that has been subjected to the audio processing is
And a voice addition means for transmitting from the decoding means to each of the video conference terminals.

Further, in the above-mentioned voice detecting means, the data corresponding to the time when the decoded data is a silent part is data of a silent level. .

Further, in the audio detecting means, the data corresponding to the time when the decoded data is a silent part is 1 / M of the decoded audio data from the audio encoding / decoding means (where M is The audio processing device of the multipoint communication control device is characterized in that the data is reduced to the audio level of the number of video conference terminals.

Further, in the voice detecting means, the data corresponding to the time when the decoded data is a silent part is 1 / M (where M is a participant) from the decoded voice data from the voice encoding / decoding means. The audio processing device of the multipoint communication control device is characterized in that the data is data that gradually decreases to the audio level of the number of video conference terminals.

[0013]

DESCRIPTION OF THE PREFERRED EMBODIMENTS The present invention will be described below according to each embodiment. Embodiment 1 FIG. FIG. 1 is a block diagram showing the configuration of a voice processing device of a multipoint communication control device according to an embodiment of the present invention, wherein 1a to 1k are voice encoding / decoding sections, 4 is a voice detection section, and 5 is a first-come-first-served section. A channel determination unit, 6a is a voice addition unit, and 7a is a control unit.

FIG. 2 is a block diagram showing an example of the voice detection unit 4 of FIG. 1. Reference numerals 41a to 41k denote voiced / silence determination units, reference numerals 42a to 42k denote silence level generation units, and reference numerals 43a to 43k.
Reference numeral 3k denotes a selector for selecting either audio data to be output to the audio addition unit 6 or received audio data or silence level data.

Next, the operation will be described. In FIG. 1, when a multipoint video conference is held, voice data from each video conference terminal (not shown) is received by voice coding / decoding sections 1a to 1k via a network, and the digital PC
After being decoded into M data, it is transmitted to the voice detection unit 4.

The voice detection unit 4 monitors a voiced portion and a non-voiced portion of the voice at regular time intervals, and determines whether the voiced voiced portion of the voice remains in the same state even after the monitoring time has elapsed. / The discrimination judgment of the silent part is performed, and the sound part /
The silent part determination result is transmitted to the first-arrival channel determination unit 5.

The first-arrival channel determination unit 5 which has received the voice / non-voice determination result from the voice detection unit 4 first determines the number of channels on which the voice portion of the voice is detected by N which is set in advance by the control unit 7a. (N is a positive integer)
In the following cases, voices from all terminals are added to the voice adding unit 6
output to a. When the number of detected sound channels exceeds N, the time at which the sound portion of the sound is recognized in each channel is determined in a time series, and only the sound of the N channels from the earliest one in the order in which the sound is generated is determined. Is output to the sound adding unit 6a without level conversion, and the sound from the other channels is subjected to a silencing process and then output to the sound adding unit 6a.

The voice adding unit 6a constantly adds voices from all the channels participating in the multipoint conference, but the voice detecting unit 4 silences voices other than that of the first channel N. As a result, the first N channels are added. The audio data from each channel added by the audio adding unit 6a is transmitted to the audio detecting unit 4, and the audio encoding / decoding unit 1a is not processed.
To 1k, and the audio encoding / decoding units 1a to 1k output
The data is encoded by the same encoding method as the received data, and transmitted to each video conference terminal via the network.

In the voice detecting section 4 of FIG. 2, the received voice data 1-k from each video conference terminal input from the voice encoding / decoding sections 1a-1k are voice / non-voice determining sections 41a-41.
41k are respectively input. Sound / silence determination unit 41a
At ~ 41k, the sound / non-sound of the sound is monitored at regular intervals, and the sound / non-speech discrimination of the sound is determined based on whether or not the same state continues even after the monitoring time has elapsed. The sound / silence determination results 1 to k are transmitted to the first-arrival channel determination unit 5. The received voice data received by the sound / non-sound determining units 41a to 41k are directly used as selectors 43a to 43k.
Sent to

In the selectors 43a to 43k, the first-arrival channel judgment results 1 to k sent from the first-arrival channel judgment unit 5
, And selects either the audio data received from the sound / non-speech determination units 41a to 41k or the non-speech data from the silence level generation units 42a to 42k.
Output to

The voice data subjected to the voice addition processing in the voice addition unit 6a is input to the voice detection unit 4, but the voice detection unit 4 performs voice coding / non-processing as transmission voice data 1 to k without any processing. Output to the decoding units 1a to 1k.

By doing so, the voice detection unit 4
Since the audio data from the video conference terminal determined as the non-first-arrival channel by the first-arrival channel determination unit 5 is muted, it is not necessary to select a channel to be added by the audio addition unit 6a, and the audio of all the channels is not required. Should always be added. Therefore, the processing is simplified and the processing capacity of the processor has a margin, so that the number of channels can be expanded.

Embodiment 2 FIG. FIG. 3 is a block diagram showing another configuration of the voice detection unit in the voice processing device of the multipoint communication control device according to the present invention. Portions that are the same as or correspond to those in the above embodiment are denoted by the same reference numerals, and description thereof is omitted. Reference numerals 44a to 44k denote 1 / M gain controllers for lowering the audio level to a value obtained by dividing the received audio data by the number M of multipoint conference participation points, that is, the number M of participating video conference terminals.

Next, the operation will be described. Voice coding
The audio data 1 to k received from the video conference terminals input from the decoding units 1 a to 1 k are used as sound / non-speech determining units 41 a to 41.
1k and 1 / M gain control units 44a-4
4k is also input. 1 / M gain control units 44a to 44k
In advance, information on the number of participating points in a multipoint conference, that is, the number of participating video conference terminals, is obtained from the control unit 7a, and audio data obtained by reducing received audio data to an audio level of 1 / M is output to the selectors 43a to 43k. ing.

The sound / non-speech judging sections 41a to 41k monitor the sound / non-speech of the sound at regular intervals, and determine whether the same state continues even if the monitoring time has elapsed. A sound / silence discrimination determination is made, and the sound / silence judgment results 1 to k are transmitted to the first-arrival channel judgment unit 5. Sound /
The received voice data received by the silence determination units 41a to 41k are sent to the selectors 43a to 43k as they are.

In the selectors 43a to 43k, the first-arrival channel judgment results 1 to k sent from the first-arrival channel judgment unit 5
Based on the received data from the sound / non-sound determining units 41a to 41k or 1 / M from the 1 / M gain control units 44a to 44k.
One of the M-level audio data is selected and output to the audio adding unit 6a.

In the first embodiment, when no video conference terminal speaks, the silent level generation units 42a to 42a
Since the output of 42k is transmitted to the voice adding unit 6a, complete silent data is transmitted from the voice adding unit 6a to each terminal. The audio data transmitted from each video conference terminal includes, in addition to the audio data of the speaker, audio data having a low level of background noise which should be unnecessary.

However, when actually conducting a video conference, a complete silence state in which no sound is output from the terminal often causes unnaturalness to the user, and it is more uncomfortable to leave some background noise. Multi-point conferences can be held without any problems.

If the background noise from each terminal is output to the selectors 43a to 43k without any processing, the background noise is added by the number of conference participation points M, and the background noise level increases. Trouble. In order to solve this problem, the level of the background noise is reduced by the 1 / M gain control units 44a to 44k to a value obtained by dividing by the number M of conference participation points and output to the voice addition unit 6a, thereby adding the background noise. The noise level is the same as the level for one terminal, and a multipoint conference without discomfort can be held.

Embodiment 3 FIG. 4 is a block diagram showing still another configuration of the voice detection unit in the voice processing device of the multipoint communication control device according to the present invention. Portions that are the same as or correspond to those in the above embodiment are denoted by the same reference numerals, and description thereof is omitted. 45a to 45k include a fade-out function 1 / M with a fade-out function for lowering the voice level while fading out the received voice data to a value obtained by dividing the received voice data by the number M of multipoint conference participation points.
It is a gain control unit.

Next, the operation will be described. Voice coding
The audio data 1 to k received from the video conference terminals input from the decoding units 1 a to 1 k are used as sound / non-speech determining units 41 a to 41.
1M input with 1k and fade-out function
The signals are also input to the gain controllers 45a to 45k. In the 1 / M gain control units 45a to 45k with a fade-out function, information on the number of participating points of the multipoint conference, that is, the number of participating video conference terminals is obtained in advance from the control unit 7a, and the 1 / M gain control unit fades out the received voice data. Is output to the selectors 43a to 43k.

The sound / non-speech judging sections 41a to 41k monitor the sound / non-speech of the voice at regular intervals, and determine whether the same state continues even after the monitoring time has elapsed. The sound / silence determination is performed, and the sound / silence determination results 1 to 1 are transmitted to the first-arrival channel determination unit 5. The received voice data received by the sound / non-sound determining units 41a to 41k is transmitted to the selectors 43a to 43k as they are.

In the selectors 43a to 43k, the first-arrival channel judgment results 1 to k sent from the first-arrival channel judgment unit 5
Based on the received data from the sound / non-sound determining units 41a to 41k or the 1 / M gain control unit 45 with a fade-out function.
Any one of 1 / M level audio data from a to 45k is selected and output to the audio adding unit 6a.

In the above-described second embodiment, the channel is determined as the first-arrival channel, and when the terminal that has made the speech stops speaking, the first-arrival channel determination result becomes non-first-arrival and is notified to the selectors 43a to 43k of the voice detection unit 4. 1 / M gain control unit 44 from output data of sound / non-sound determination units 41a to 41k
The output data is suddenly switched to the output data of a to 44k. For this reason, there may be a sense of incongruity such as the end of the voice of the speaker being interrupted, or as soon as the utterance ends, the speaker suddenly becomes quiet.

Therefore, the 1 / M gain control units 45a to 45k with fade-out function use the first channel determination results 1 to k sent from the first channel determination unit 5, and the sound / non-sound determination units 41a to 41k during the first arrival. And outputs the received voice data to the selectors 43a to 43k without performing gain control. Thereafter, the gain of the received voice data is gradually reduced from the point in time when the remark is finished and it is determined that the voice is non-first-come-first-served, and finally 1 / M
Lower the gain. By performing such processing, it is possible to realize a sound addition processing without discomfort without ending the end of the voice of the speaker.

[0036]

As described above, according to the present invention, there is provided an audio processing apparatus of a multipoint communication control apparatus for realizing a multipoint videoconference using videoconference terminals at three or more points. The audio data is decoded into digital data for each channel for receiving the audio data from the terminal, and the final audio data subjected to the audio processing is encoded by the same encoding method as the received data and transmitted to each video conference terminal. Each of the voice encoding / decoding means to be transmitted, and the audio data of each channel decoded from the voice encoding / decoding means, each of which is discriminated between a voiced part and a non-voiced part at regular intervals, and For each channel, one of the decoded audio data from the audio encoding / decoding means and the data corresponding to the time when the decoded data is a silent part is generated. If the number of channels in which a sound part is detected is N or less (where N is a positive integer) based on the voice detection means and the identification determination result by the voice detection means, the voice coding is performed for all channels. Generating the decoded audio data from the decoding means, and when the number of channels in which the sound portion is detected exceeds N, for the N channels from the earliest in the order in which the sound portion is recognized, First-arrival channel determination means for controlling the voice detection means so as to generate the decoded voice data and to generate data corresponding to the silent portion for the remaining channels; and voice data generated from the voice detection means. And all the data corresponding to the silence portion, and adds the voice-processed voice data from the voice encoding / decoding means to each of the video conference terminals. With the addition of a stage, audio addition can be performed by always adding audio from all the video conference terminals, simplifying the processing, and creating a margin in the processing capability of the processor and the like used for audio addition. The number of channels that can be processed is expanded, and the effect of being able to provide a voice processing device of a multipoint communication control device that can process more voices of the video conference terminal is obtained.

Further, in the voice detecting means, the data corresponding to the time when the decoded data is a silent part is data of a silent level, so that it is possible to process more voices of the TV conference terminal with a simple configuration. Advantages such as providing a voice processing device for a multipoint communication control device can be obtained.

Further, in the voice detecting means, the data corresponding to the time when the decoded data is a silent part is replaced with the decoded voice data from the voice coding / decoding means by 1 / M (where M is the participating television). Since the data is reduced to the audio level of (the number of conference terminals), there is no completely silent state in which no audio is output from the terminal, and a multipoint communication control device that does not give a feeling of strangeness to the user who leaves some background noise And the like can be provided.

Further, in the voice detecting means, the data corresponding to the time when the decoded data is a silent part is converted by the decoded voice data from the voice encoding / decoding means into 1 / M (where M is (No. of conference terminals) The data is gently lowered to the audio level, so there is no complete silence where no audio is output from the terminal, leaving some background noise, and the ending of the voice of the speaker As a result, it is possible to provide an audio processing device of a multipoint communication control device that does not give the user a sense of incongruity such as a break or a sudden stop of speech as soon as the speech ends.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a configuration of a voice processing device of a multipoint communication control device according to an embodiment of the present invention.

FIG. 2 is a block diagram illustrating an example of a voice detection unit in FIG. 1;

FIG. 3 is a block diagram illustrating another example of the voice detection unit in FIG. 1;

FIG. 4 is a block diagram showing still another example of the voice detection unit in FIG. 1;

FIG. 5 is a block diagram of a conventional voice processing device of the multipoint communication control device.

[Explanation of symbols]

1a to 1k voice encoding / decoding section, 4 voice detection section, 5
First-arrival channel determination unit, 6a Voice addition unit, 7a control unit, 41a to 41k Voiced / silence determination unit, 42a to 42
k silent level generator, 43a to 43k selector, 4
4a to 44k1 / M gain control unit; 45a to 45k 1 / M gain control unit with fade-out function.

Claims (4)

[Claims] An audio processing device of a multipoint communication control device for realizing a multipoint videoconference using videoconference terminals at three or more points, wherein the channel receives the audio data from each of the videoconference terminals. Audio encoding / decoding means for decoding the audio data into digital data each time, and encoding the final audio data subjected to the audio processing by the same encoding method as the received data and sending the encoded data to each video conference terminal, The audio data of each channel decoded from the audio encoding / decoding means is discriminated at a predetermined time interval between a sound part and a silent part of the audio, and the audio encoding / decoding means is determined for each channel. Voice detection means for generating either of the decoded voice data from the device and data corresponding to the time when the decoded data is a silent portion; If the number of channels in which a sound part is detected is N or less (where N is a positive integer) based on the result of the identification determination by the means,
When the number of channels in which the sound part is detected exceeds N, the decoded sound data from the decoding means is generated. First-channel determining means for controlling the voice detecting means so as to generate decoded voice data, and for the remaining channels, generating data corresponding to the silent portion; voice data generated from the voice detecting means; Multi-point adding means for adding all data corresponding to a silent portion and transmitting the audio-processed audio data from the audio encoding / decoding means to each of the video conference terminals. Voice processing unit for communication control unit. 2. The multipoint communication control apparatus according to claim 1, wherein, in the voice detecting means, data corresponding to a time when the decoded data is a silent part is data of a silent level. Audio processing device. 3. In the audio detecting means, the data corresponding to the time when the decoded data is a silent part is 1 / M (where M is a participant) the decoded audio data from the audio encoding / decoding means. 2. The voice processing device of the multipoint communication control device according to claim 1, wherein the data is data reduced to a voice level of the number of video conference terminals. 4. In the voice detecting means, the data corresponding to the time when the decoded data is a silent part is 1 / M (where M is a participant) from the voice encoding / decoding means. 2. The voice processing device for a multipoint communication control device according to claim 1, wherein the data is data that gradually decreases to the voice level of the number of video conference terminals.