JP2000069179A - Multispot conferencing device and its method, and terminal device for multispot conferencing - Google Patents

Abstract

PROBLEM TO BE SOLVED: To prevent a voice of transmit voice data from degrading in quality by putting together input voice datagenerated without a filtering process by decoding 1st transmitted voice encoded data and transmit voice data to be transmitted, and encoding the composite data into 2nd voice encoded data, and sending the data to a transmission destination different from the transmission source of the 1st voice encoded data. SOLUTION: Terminal devices 11A to 11N respectively fetch multiplexed data D5 through a 1st I/F circuit 18 and obtain 1st voice encoded data D21 and control data D4 at a demultiplexer 19. A control part 15 processes the data D21 on a 3rd decoding basis by placing a 1st decoding part 24 in operation according to the control data D4, and sends the data out to the side of a speaker 33 after performing a filtering process and to a 1st adder 26 without performing the filtering process. The 1st adder 26 puts together the output of the decoding part 24 and the transmit voice data D12 generated by its terminal device, and sends the composite data to a next stage through a 1st encoding part 39... a 2nd I/F circuit 42.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a multipoint conference apparatus and a method therefor, and is suitably applied to, for example, a television conference system and its terminal apparatus.

[0002]

2. Description of the Related Art Conventionally, in a video conference system, audio data and video data are transmitted and received between terminal devices installed at respective multipoints so that a video based on audio and video data based on the audio data is used. It is designed to be able to hold meetings.

FIG. 13 shows an example of the configuration of such a television conference system 1.
A to 2N are connected to a multipoint control unit (MCU: Multipoint Control Unit) 4 via the public line network 3, and each of the terminal devices 2A to 2N transmits audio data and video data to be transmitted to the public line network 3. The data can be transmitted to the multipoint control device 4 via the multipoint control device 4.

[0004] The multipoint control device 4 includes terminal devices 2A to 2A.
N, the voice data corresponding to each of the other terminal devices 2A to 2N except for one terminal device 2A to 2N are sequentially added and synthesized, and the voice data supplied from each of the terminal devices 2A to 2N. Of the received video data,
Each of the other terminal devices 2A to 2 except for the two terminal devices 2A to 2N
The video data corresponding to N is added and combined.

[0005] The multipoint control device 4 connects the public line network 3 to the terminal devices 2A to 2N corresponding to the audio data and the video data which are excluded when synthesizing the synthesized audio data and the corresponding synthesized video data. To be transmitted through.

Accordingly, each of the terminal devices 2A to 2N emits a synthesized voice based on the synthesized voice data supplied from the multipoint control device 4 from a speaker, and simultaneously outputs a plurality of images based on the synthesized video data. It is possible to collectively display the images on the monitor, and thus view the audio based on the audio data and the video based on the video data obtained in each of the other terminal devices 2A to 2N except the own terminal devices 2A to 2N.

In the video conference system 1 having such a configuration, since the multipoint control device 4 is used to connect the terminal devices 2A to 2N, the configuration of the entire system becomes complicated. Some systems are configured as shown in FIG.

In the television conference system 5, a plurality of terminal devices 6A to 6D are directly connected via a public line network 7, and the terminal devices 6A to 6D transmit audio data and video data to be transmitted, respectively. The data is transmitted to the other terminal devices 6A to 6D via the public line network 7.

[0009] Each of the terminal devices 6A to 6D synthesizes the respective voice data given from the other terminal devices 6A to 6D, and then emits a synthesized voice based on the synthesized voice data from the speaker, and Terminal devices 6A to 6D
After synthesizing each of the video data provided from the above, a plurality of videos based on the synthesized video data are collectively displayed on a monitor, and thus obtained in each of the other terminal devices 6A to 6D except the own terminal devices 6A to 6D. Audio based on audio data and video based on video data can be viewed.

[0010]

By the way, in the video conference system 5 having such a configuration, the terminal systems 6A to 6D are connected via the public line network 7 except for the multipoint control device, so that the whole system can be controlled. Although there is an advantage that the configuration can be simplified, there is a problem that the connection between the terminal devices 6A to 6D becomes complicated when the system is constructed.

In the television conference system 5, a public line network 7 is connected to each of the terminal devices 6A to 6D.
Is provided with an interface circuit for connecting each of the other terminal devices 6A to 6D via the interface. Therefore, when it is requested to connect a plurality of terminal devices larger than the number of the interface circuits, this can be easily performed. Had difficulties to deal with.

For this reason, as a television conference system which solves this problem, conventionally, a plurality of terminal devices are sequentially connected in series via a public line network, and each terminal device transmits audio data to be transmitted to the public line network. It is considered that the video data to be transmitted is transmitted while being sequentially combined in the next and previous stages through a public line network while being combined in the next and previous stages in both directions. According to this configuration, there is an advantage that the terminal devices can be easily connected to each other at the time of constructing the system, and the addition of the terminal devices can be easily performed.

In a television conference system having such a configuration, each terminal device encodes audio data (hereinafter referred to as input audio data) from the next and previous terminal devices in a predetermined manner during data transmission. Encoded data (hereinafter, referred to as first audio encoded data) obtained by performing encoding processing according to the system is provided via a public line network, and video data (hereinafter, referred to as input video data). ) (Hereinafter, referred to as first coded video data) obtained by performing coding processing on the video data according to a predetermined coding method.

Each of the terminal devices decodes the first coded audio data by a predetermined decoding method, and obtains the obtained input audio data in the transmission target audio data (hereinafter, referred to as “hereafter”) generated in its own terminal device. This is referred to as transmission voice data), and is subjected to coding processing according to a predetermined coding method. The obtained coded data (hereinafter referred to as second coded voice data) is transmitted to a public line network. Video data to be transmitted to the next and previous terminal devices via the first terminal, and the first video encoded data is decoded by a predetermined decoding method, and the obtained input video data is generated by the own terminal device as a transmission target video. After synthesizing with data (hereinafter referred to as transmission video data), the data is encoded by a predetermined encoding method,
The obtained encoded data (hereinafter, referred to as second encoded video data) is transmitted to the next and previous terminal devices via the public network.

Here, in the video conference system having such a configuration, when transmission of input audio data is considered,
Pulse code modulation with different compression ratios (Pulse Code M
coding method called A-law coding method (A-Law) (hereinafter referred to as a first coding method) and coding method called μ-law coding method (μ-Law) , This is referred to as a second encoding method), and the International Telecommunication Union Telecommunication Standardization Sector (ITU-T: International Teleco
mmunication Union-Telecommunication Standardizatio
Low Delay-Code Excited Linear Predic (LD-CELP)
) (hereinafter referred to as a third coding method) and Sub-band Adaptive Differential Pulse Coding (SB-ADPCM).
The input audio data is encoded using a desired encoding method among encoding methods called “de Modulation” (hereinafter, this is referred to as a fourth encoding method).

By the way, among the first to fourth encoding methods,
The first obtained by performing the encoding process according to the third encoding method
When decoding the audio encoded data of the first, the original input audio data is generated from the first audio encoded data by using a synthesis filter by a corresponding predetermined decoding method. Since the input voice data thus obtained contains audible noises and distortions (hereinafter collectively referred to as noise), the input voice data is usually subjected to a filtering process through a post filter. This eliminates the noise.

However, when the third encoding method is selected in a plurality of terminal devices, when the input audio data is repeatedly passed through the post-filter in the next-stage and preceding-stage terminal devices, the operation of the post-filter is performed. As a result, the input voice data is sequentially attenuated, and as a result, the quality of voice based on the input voice data is significantly deteriorated.

The present invention has been made in view of the above points, and a multipoint conference apparatus and method capable of transmitting input voice data to be transmitted while preventing deterioration in voice quality based on the input voice data. In addition, a terminal device for a multipoint conference is proposed.

[0019]

According to the present invention, there is provided a multi-point conference terminal installed at a multi-point, and a line network for sequentially subordinately connecting the multi-point conference terminal. The multipoint conference terminal device captures and outputs the first voice encoded data transmitted from the preceding and / or next stage multipoint conference terminal device via the network. Means for decoding the first encoded audio data output from the capturing means;
Decoding means for performing filtering processing for removing noise in order to emit the voice based on the input voice data, and outputting the input voice data without performing the filtering processing; and Audio data generation means for generating and outputting transmission audio data; input audio data output without filtering processing from the decoding means; and transmission audio data output from the audio data generation means, and synthesized. Synthesizing means for outputting the obtained output audio data, encoding means for encoding the output audio data output from the synthesizing means, and outputting the obtained second audio encoded data, and output from the encoding means. The obtained second voice-encoded data is connected to the preceding and / or subsequent multipoint conference terminal devices different from the transmission source of the corresponding first voice-encoded data. And to have a transmitting means for transmitting through.

As a result, it is possible to prevent the input voice data to be transmitted from attenuating due to the effect of the filtering process repeated in each multipoint conference terminal device.

Also, in the present invention, a first step of sequentially cascade-connecting the multipoint conference terminal devices respectively installed at the multipoint via a predetermined network, and the first step in each multipoint conference terminal device And / or decoding the first coded audio data transmitted from each multipoint conference terminal device at the next stage via the circuit network, and releasing the obtained input audio data based on the input audio data. A second step of performing filtering processing for removing noise in order to make the sound sound and outputting the filtered sound without performing the filtering processing, and generating transmission audio data to be transmitted, and generating the transmission audio data. Second audio encoded data is generated by synthesizing the input audio data output without being subjected to the filtering processing and encoding the obtained output audio data. And transmitting the second voice-encoded data to each of the preceding and / or subsequent multipoint conference terminal apparatuses different from the transmission source of the corresponding first voice-encoded data via the network. And a fourth step of performing the operation.

As a result, it is possible to prevent the input voice data to be transmitted from being attenuated by the effect of the filtering process repeated in each multipoint conference terminal device.

Further, in the present invention, decoding processing is performed on externally supplied first voice encoded data, and noise is removed from the obtained input voice data so as to emit a voice based on the input voice data. Decoding means that performs filtering processing and outputs the data without performing the filtering processing, audio data generating means that generates and outputs transmission audio data to be transmitted, and filtering processing that is not performed by the decoding means. A synthesizing unit for synthesizing the input audio data output to the unit and the transmission audio data output from the audio data generating unit, and outputting the obtained output audio data; and encoding the output audio data output from the synthesizing unit. An encoding means for processing is provided.

As a result, it is possible to prevent input voice data transmitted by a plurality of multipoint conference terminals sequentially connected in cascade via a predetermined line network from being attenuated by the action of the filtering process which is sequentially repeated. it can.

[0025]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described below in detail with reference to the drawings.

(1) Configuration of Video Conference System According to the Present Embodiment In FIG. 1, reference numeral 10 denotes a video conference system to which the present invention is applied as a whole, and a plurality of terminal devices 11A to 11A installed at multiple points. 11N is an Integrated Services Digital Network (ISDN)
k) and the like are sequentially connected in series via a public line network 12.

In each of the terminal devices 11A to 11N, as shown in FIG. 2, a control unit 15, a user interface 16 such as a keyboard for inputting various control information, and a corresponding terminal device 11A to 11N are operated. And a non-volatile memory 17 in which a predetermined program is stored in advance.
Memory 17 based on various control information input through
Of the terminal devices 11A to 11N based on the activated programs.

When each of the terminal devices 11A to 11N is connected to a line between the previous and next terminal devices 11A to 11N at the time of data transmission, the first terminal device 11A to 11N transmits a first signal to the first terminal device 11A to 11N. the speech encoded data D2 _1,
The _first encoded video data D3 ₁ and the terminal device 11 in the preceding stage
A multiplexed data D5 obtained by time-division multiplexing control data D4 for setting a common operation mode among A to 11N is transmitted via the public line network 12, and transmitted to the first interface. It is taken into the circuit 18.

The first interface circuit 18 adds the first identification information representing the terminal equipment 11A to 11N at the preceding stage of the transmission source corresponding to the multiplexed data D5 taken in, and sends it to the demultiplexer 19.

Upon receiving the multiplexed data D5 from the first interface circuit 18, the demultiplexer 19 detects the first identification information added thereto, and then converts the multiplexed data D5 to the first audio code. Data D2
_1, the first video coded data D3 _1, control data D
And 4.

Then, the demultiplexer 19
Based on the identification information of the first audio encoded data D2 ₁
Is added to a first identification number (for example, “0”) previously assigned to the terminal devices 11A to 11N at the preceding stage, and the first and second recognizers 2 of the first and second voice processing units 20 and 21 are added.
Sends out the 2 and 23, the first video coded data D3 ₁ sends to the image processing unit (not shown) by adding a first identification number and the control data D4 by adding a first identification number It is sent to the control unit 15.

In this case, the first recognizer 22 outputs the first coded speech data D2 given from the demultiplexer 19.
Recognize the identification number added to ₁ and, based on the result of the recognition, determine whether the first identification number specified in advance is added to the first identification number.
Selecting only the speech encoded data D2 _1, and sends the first decoding unit 24 by removing the first identification number.

[0033] The first decoding unit 24, a first encoded audio data D2 ₁ supplied from the first recognizer 22 processes decoded by a predetermined decoding method, resulting input audio data the D1 ₁ sends out to the first adder 26 via a first rate converter 25, and sends to the third adder 29 sequentially through the second and third rate converters 27 and 28.

At this time, in each of the terminal devices 11A to 11N, a sound signal S1 to be transmitted is obtained by collecting sound through the microphone 30, and this is transmitted through the analog / digital converter 31. D1 ₂
After that, it is sent to the echo canceller 32.

The echo canceller 32, the transmission voice data D1 ₂ provided from the analog / digital converter 31
Is suppressed by the acoustic coupling (the sound emitted from the speaker 33 is mixed into the microphone 30) and the fourth rate converter 3
First and second adders 26 and 36 via 4 and 35
To send to.

Thus, the first adder 26 receives the signal from the own terminal 1 via the fourth rate converter 34.
Transmission audio data D1 generated in 1A to 11N
_2, supplied through a first rate converter 25, by adding processing the input audio data D1 ₁ obtained from the preceding terminal device 11A~11N synthesized speech obtained data (hereinafter, the output is referred to as audio data) D1 ₃ successively through the first selector 37 and the fifth rate converter 38 is sent to a first encoding unit 39.

The first encoding unit 39, the output audio data D1 ₃ supplied through the fifth rate converter 38 to process encoded by a predetermined encoding method, the resulting second
The speech encoded data D2 ₂ delivered to a first adder 40.

The first adder 40 includes a first encoding unit 39
, A second identification number (for example, "1") assigned in advance to the next terminal device 11A to 11N at the transmission destination corresponding to the _second encoded voice data D22 given by Send to 41.

In this case, the multiplexer 41 combines the input video data obtained from the terminal devices 11A to 11N in the preceding stage from the video processing unit with the transmission video data to be transmitted generated in the own terminal devices 11A to 11N. , predetermined processing encoded by the coding scheme, and the second with the second video coded data D3 ₂ is obtained by adding the identification number given, the control from the control unit 15 predetermined data D
4 is provided so that the multiplexer 41
Is multiplexed data and the second video coded data D3 _2, the control data D4, by time multiplexing the second and the speech encoded data D2 ₂ given from the first adder 40 Generate D5.

At this time, the multiplexer 41
The second identification number added to the audio encoded data D2 ₂ and the second video encoded data D3 ₂ is detected, and based on the detection result, the corresponding multiplexed data D5 is included in the second encoded data D5. After the identification number is removed, the data is transmitted to the next terminal device 11A to 11N at the corresponding transmission destination via the second interface circuit 42 and the public line network 12 sequentially.

On the other hand, each terminal device 11A to 11N, a first speech encoded data D2 ₃ from the next stage of the terminal unit 11A to 11N, the first video coded data D3 _3, a control data D4 The multiplexed data D5 obtained by time-division multiplexing is transmitted via the public line network 12, and is taken into the second interface circuit 42.

The second interface circuit 42 adds second identification information indicating the next terminal device 11A to 11N of the transmission source corresponding to the multiplexed data D5 taken in, and sends it to the demultiplexer 19. .

When the multiplexed data D5 is supplied from the second interface circuit 42, the demultiplexer 19 detects the second identification information added thereto, and converts the multiplexed data D5 into the first voice encoded data. and data D2 _3, the first video coded data D3 _3, separates the control data D4.

The demultiplexer 19 outputs the second
Based on the identification information of the first audio encoded data D2 ₃
To the first and second recognizers 22 by adding a second identification number.
And 23, and the first video encoded data D
3 ₃ was sent to the second video processing unit (not shown) by adding the identification number, also sends control data D4 to the second adds the identification number controller 15.

In this case, the second recognizer 23 outputs the first encoded speech data D2 given from the demultiplexer 19.
_3. Recognize the identification number added to ₃ and, based on the recognition result, specify the first identification number
Selecting only the speech encoded data D2 _3, is sent to the second decoding unit 43 by removing the second identification number.

[0046] The second decoding unit 43, a first speech encoded data D2 ₃ supplied from the second recognizer 23 processes decoded by a predetermined decoding method, resulting input audio data the D1 ₄ sends out the second adder 36 through a first rate converter 44, and sends to the third adder 29 sequentially through the second and third rate converters 45 and 46.

At this time, the second adder 36 transmits the transmission audio data D given through the fourth rate converter 35.
And 1 _2, was given via a first rate converter 44, by adding processing the input audio data D1 ₄ obtained from the next stage of the terminal device 11A~11N synthesized, obtained output audio data D1 ₅ is transmitted to the second encoding unit 49 via the second selector 47 and the fifth rate converter 48 in order.

The second encoding unit 49, the output audio data D1 ₅ supplied through the fifth rate converter 48 to process encoded by a predetermined encoding method, the resulting second
The speech encoded data D2 ₄ sent to the second adder 50.

The second adder 50 includes a second encoding unit 49
Adding first identification number corresponding to the transmission destination to the second speech encoded data D2 ₄ given from, and sends it to the multiplexer 41.

In this case, the multiplexer 41 combines the input video data obtained from the terminal devices 11A to 11N at the next stage from the video processing unit with the transmission video data to be transmitted generated in the own terminal devices 11A to 11N. after a predetermined processing encoded by the coding scheme, and the first together with the second video coded data D3 ₄ obtained by adding the identification number is given, the control from the control unit 15 predetermined data D
4 is provided so that the multiplexer 41
Is multiplexed data and the second video coded data D3 _4, the control data D4, by time division multiplexing and second speech encoded data D2 ₄ supplied from the second adder 50 Generate D5.

At this time, the multiplexer 41
Of detecting the first identification number added to the voice coded data D2 ₄ and the second video coded data D3 _4, first it contained the multiplexed data D5 corresponding based on the detection result to After the identification number is removed, the data is transmitted via the first interface circuit 18 and the public line network 12 sequentially to the corresponding terminal device 11A to 11N at the preceding stage of the transmission destination.

[0052] The third adder 29, addition processing input audio data D1 ₁ provided from the first speech processing section 20, the input audio data D1 ₄ supplied from the second speech processing section 21 and to synthesize the voice data obtained after suppress echo and howling due to acoustic coupling (hereinafter, referred to as sound emitting data) D1 ₆ via the echo canceller 32, a digital / analog converter 51 The signal is converted to an audio signal S2 through the speaker 33.

Thus, in each of the terminal devices 11A to 11N, based on the audio signal S2 from the speaker 33,
Other terminal devices 11A except the own terminal devices 11A to 11N
To 11N can be emitted.

In the terminal device 11A at one end,
Since the terminal device is not connected to the previous stage, the input audio data D1 ₄ obtained from the second speech processing section 21 third
Via the adder 29 is sent as a sound emitting data D1 ₆ to the echo canceller 32.

[0055] In the other end of the terminal device 11N, since the next stage the terminal device is not connected, the input audio data D1 ₁ obtained from the first speech processing section 20 third adder 29 through sending a sound emitting data D1 ₆ to the echo canceller 32.

In this way, each of the terminal devices 11A to 11N
In, and conveyed while sequentially synthesized to the next stage and the previous stage of the transmission voice data D1 ₂ to be transmitted, this time other terminals except for the own terminal apparatus 11A to 11N 11A～11
N so that the user can hear the sound obtained.

Incidentally, in each of the terminal devices 11A to 11N, the first and second voice processing units 20 and 2
By switching the contacts of the selector 37 and 47 at 1, instead of the output audio data D1 ₃ and D1 ₅ supplied from the first and second adders 26 and 36, the second
It sends the rate converters 27 and 45 input audio data D1 ₁ and D1 ₄ obtained through the the first and second encoding section 39 and 49 which correspond.

In each of the terminal devices 11A to 11N, the first identification number is added by the first adder 40 to the _second encoded audio data D22 given from the first encoding unit 39 at this time. while, adding the second identification number in the second adder 50 to the second speech encoded data D2 ₄ supplied from the second encoding unit 49.

In this way, in each of the terminal devices 11A to 11N, at the time of this failure inspection, the next and previous terminal devices 11A to 11N are connected.
A~11N to return the input audio data D1 ₁ and D1 ₄ which transmitted as a multiplexed data D5 from the next stage and the previous stage of the terminal device 11A~11N in, thus sent back to the input audio data D1 ₁ and D1 ₄ that transmitted compared to the input audio data D1 ₁ and D1 _4, it is made to be able to check whether a failure has occurred in its own terminal apparatus 11A~11N based on the comparison result.

(2) Detailed Configuration of Each of Terminal Devices 11A to 11N
In the case of, when encoding the output audio data D1 ₃ and D1 ₅ , a desired encoding method among the first to fourth encoding methods can be selected and designated via the user interface 16. The control unit 15 controls the first and second encoding units 3 based on the selected and designated encoding method.
At 9 and 49, the corresponding encoding process is executed.

Further, the control unit 15 controls the control data D included in the multiplexed data D5 with information indicating the designated encoding system.
4 and transmitted to the next and previous terminal devices 11A to 11N, whereby the next and previous terminal devices 11A to 11N are transmitted.
In, the control unit 15 causes the first and second decoding units 24 and 43 to execute the corresponding decoding processing based on the control data D4 given from the demultiplexer 19.

Actually, the first and second encoding units 39 and 4
9, first to fourth encoders 55 to 5 that can perform encoding processing based on the first to fourth encoding schemes, as shown in FIG.
8 and corresponding fifth rate converters 38 and 4
Output audio data D1 given via the 8 ₃ and D1 ₅
Through the selector 59 to the first to fourth encoders 55 to 58 corresponding to the selected first to fourth encoding schemes.

Then, the first and second encoding units 39 and 4
9, the output audio data D1 ₃ and D1 ₅ in the corresponding first to fourth encoders 55-58 treated encoding second audio encoded data D2 ₂ and D2 ₄ obtained
To the corresponding first and second adders 40 and 50 via the selector 60.

The first and second decoding units 24 and 43
Has first to fourth decoders 61 to 64 that can perform decoding processing according to the first to fourth decoding methods corresponding to the first to fourth encoding methods, as shown in FIG. And the corresponding first
And a first speech encoded data D2 ₁ and D2 ₃ via the selector 65 supplied from the second recognizer 22 and 23, the first corresponding to the first to fourth coding scheme selected
To the fourth decoders 61 to 64.

Then, the first and second decoding units 24 and 4
3, the corresponding first to fourth this first speech encoded data D2 ₁ and D2 ₃ treated decoded by the decoder 61 to 64, a selector input audio data D1 ₁ and D1 ₄ obtained It sends out to the first and second rate converters 25, 44 and 27, 45 via 66.

Here, in practice, the third encoder 57
When the output audio data D1 ₃ and D1 ₅ via the selector 59 is given by the coding method is selected,
With dividing the output audio data D1 ₃ and D1 ₅ sequentially five consecutive sample blocks, the divided-obtained sample block (hereinafter, referred to as input signal blocks) respectively corresponding to the previously After adjusting the gain of the stored 1024 codebook vector candidates, for example, a filtering process is performed through a synthesis filter to generate 1024 quantized vectors.

Then, the third encoder 57 selects, from the 1024 quantized vectors, the one that minimizes the root mean square error by performing frequency weighting processing together with the corresponding input signal block, and selects the selected vector. and sends to the selector 60 a codebook index representing, for example, 10-bit codebook vector corresponding to the quantized vector as the second speech encoded data D2 _2, and D2 _4.

The third encoder 57 periodically updates the gain and the coefficients of the synthesis filter by applying the backward based on the corresponding past gain-adjusted codebook vector and quantization vector.

As shown in FIG. 5, the third decoder 63 performs a decoding process in units of five sample blocks.
Coded data D2 ₁ and D2 ₃ (that is, the second
Coded data D2 ₂ and D2 ₄ ) in the excitation vector codebook 68.

In this case, the excitation vector code book 68
Preliminarily stores 1024 codebook vector similarly to the third encoder 57, based on the first speech encoded data D2 ₁ and D2 ₃ (i.e. index information), 1024 Code A corresponding codebook vector is extracted from the book vector, and the extracted codebook vector is transmitted to the synthesis filter 70 via the gain control circuit 69 as codebook vector data D8.

The synthesis filter 70 performs a linear predictive coding (LPC: Linear Predictive Coding) of order 50 on the feedback path.
ding) is constituted by 50-order all-pole filter with a predictor, the input audio data D1 ₁ of the original five sample block basis based on the codebook vector data D8 supplied through a gain control circuit 69 and D1 ₄ generates, after passing the post filter 71, the second rate converters 27 corresponding via selector 66,
45, and without passing through the post filter 71, through the selector 66 to the first rate converter 2
5 and 44.

The backward gain adaptor 72 receives the codebook vector data D8 of which the gain has been adjusted from the gain control circuit 69, and performs logarithmic domain based on the gain of the codebook vector data D8 previously given. , The gain of the codebook vector data D8 currently given to the gain control circuit 69 is predicted, and the gain of the gain control circuit 69 is controlled based on the prediction result.

The backward synthesis filter adaptor 73
The input speech data from the synthesis filter 70 D1 ₁ and D
1 ₄ and is provided to control the filter coefficients of the synthesis filter 70 based on the input audio data D1 ₁ and D1 _4.

Further, the third decoder 63 is adapted to periodically update the filter coefficient of the post filter 71 based on predetermined information obtained at the time of decoding processing.

Thus, the first and second decoding units 2
4 and 43, the third encoding scheme may selectively designated, performs filtering processing through the third decoder 63 obtained from the synthesis filter 70 in the input audio data D1 ₁ and D1 ₄ post filtering 71 together to remove perceptual noise is sent to the speaker 33 by, and the input audio data D1 ₁ and D1 ₄ obtained from the synthesis filter 70 so as not subjected to filtering processing without passing through the post-filter 71, The first corresponding to the transmission to the terminal devices 11A to 11N of the next stage and the preceding stage
And to the second adders 26 and 36.

Incidentally, in the first to third encoding methods, it is specified that data generated at a sampling frequency of 8 [kHz] is to be encoded, and in the fourth encoding method, It is stipulated that data generated at a sampling frequency of 16 [kHz] is encoded.

In each of the terminal devices 11A to 11N,
The analog / digital converter 31 executes an analog / digital conversion process at a sampling frequency of, for example, 16 [kHz], and the digital / analog converter 51 outputs, for example, a 16 kHz sampling frequency.
The digital / analog conversion processing is executed based on the sampling frequency of [kHz].

Accordingly, in each of the terminal devices 11A to 11N, the first to fifth rate converters 25, 27, 28,
34,35,38,44,45,46,48 selectively using input if necessary audio data D1 _1, D1 ₄ and transmitted voice data D1 ₂ and output the audio data D1 _3, D
1 converts ₅ of sampling frequency from 16 [kHz] to 8 [kHz] treated with (hereinafter, this is referred to as a decimation process), the input audio data D1 _1, D1 ₄ and transmission audio data D
1 ₂ and conversion output audio data D1 _3, D1 ₅ sampling frequency from 8 [kHz] to 16 [kHz] (hereinafter referred to as interpolation) to.

In this case, the first to fifth rate converters 2
5, 27, 28, 34, 35, 38, 44, 45, 4
The low-pass filters 6 and 48 are provided with a low-pass filter for attenuating a frequency component of about 4.0 [kHz] or more of data analog-to-digital converted at a sampling frequency of 16 [kHz]. When 16 input audio data D1 ₁ of the sampling frequency [kHz], D1 ₄ and transmitted voice data D1 ₂ and output the audio data D1 _3, D1 ₅ is given, after which through a low-pass filter, the specimen thinning sequentially every other value, thus the input audio data D1 _1,
D1 ₄ and transmitted voice data D1 ₂ and output the audio data D1 _3, D1 8 a sampling frequency of 16 [kHz] of ₅ [kHz
].

The first to fifth rate converters 25,
27, 28, 34, 35, 38, 44, 45, 46, 4
Sample 8, during the interpolation process, given the sampled input speech data D1 ₁ frequency, D1 ₄ and transmitted voice data D1 ₂ and output the audio data D1 _3, D1 ₅ of 8 [kHz], the mutually sequentially adjacent this after inserting the 0 amplitude sample values between values, passed through a low-pass filter, the resulting input audio data D1 _1, D1 ₄ and transmitted voice data D1 ₂ and output the audio data D1 _3, D1 ₅ amplitude doubled according to the need, thus the input audio data D1 _1, D1 ₄ and transmitted voice data D1 ₂ and output the audio data D1 _3, D
1 ₅ interpolating the sampling frequency from 8 [kHz] to 16 [kHz] of.

Incidentally, the first to fifth rate converters 2
5, 27, 28, 34, 35, 38, 44, 45, 4
6,48, if it is not necessary to thinning processing and interpolation processing sampling frequency of the audio data D1 ₁ ~ D1 _5, the corresponding input speech data D1 _1, D1 ₄ and transmitted voice data D1 ₂ and the output audio data D1 _3, directly output to the subsequent the D1 _5.

[0082] As shown in practice Figure 6, the output audio data D1 ₃ and D1 ₅ in the first and second encoding section 39 and 49 treated encoded by the first to third encoding scheme, also the first and the second decoding unit 24 and 43 the first speech encoded data D2 ₁ and D2 ₃ first to third
When the decoding process is performed by the decoding method of
Fifth rate converters 25, 27, 28, 34, 3
Of 5,38,44,45,46,48, third and fourth rate converters 28, 46 and 34, 35 only is operated, the specimen of the transmission voice data D1 ₂ obtained through the echo canceller 32 Of the optimized frequency via the corresponding fourth rate converters 34 and 35,
And interpolation processing through a second decoding unit third rate converters 28 and 46 corresponding to the sampling frequency of the input audio data D1 ₁ and D1 ₄ obtained through the 24 and 43.

[0083] Further, as shown in FIG. 7, the output audio data D1 ₃ and D1 ₅ in the first and second encoding section 39 and 49 treated encoded by the first to third encoding scheme, also the 1 and the second decoding unit 24 and 43 when the first audio encoded data D2 ₁ and D2 ₃ to decoding processing by the fourth decoding method, first to fifth rate converter 25, 27, 28, 34, 35, 3
8, 44, 45, 46, and 48, only the fifth rate converters 38 and 48 are operated, and the output audio data D1 obtained through the first and second adders 26 and 36 are output.
₃ and D1 ₅ sampling frequency via a fifth rate converter 38 and 48 corresponding to the thinning process.

[0084] As further shown in FIG. 8, the output audio data D1 ₃ and D1 ₅ in the first and second encoding unit 39 and 49 processes the encoding by the fourth coding system, also the first and second in the decoding unit 24 and 43 of 2 when the first speech encoded data D2 ₁ and D2 ₃ to decoding processing by the first to third decoding method, first to fifth rate converter 25, 27, 28, 34, 35, 3
Of the 8, 44, 45, 46, 48, only the first and second rate converters 25, 44 and 27, 45 were operated and obtained via the first and second decoders 24 and 43. interpolating process via the first and second rate converters 25,44 and 27 and 45 to the corresponding sampling frequency of the input audio data D1 ₁ and D1 _4.

[0085] As further shown in FIG. 9, the output audio data D1 ₃ and D1 ₅ in the first and second encoding unit 39 and 49 processes the encoding by the fourth coding system, also the first and second in the decoding unit 24 and 43 of 2 when the first speech encoded data D2 ₁ and D2 ₃ to decoding processing by the fourth decoding method, all of the first to fifth rate converter 25, 27 , 28,34,35,3
The operation of 8,44,45,46,48 stopped, interpolation and decimation process on the input audio data D1 _1, D1 ₄ and transmitted voice data D1 ₂ and output the audio data D1 _3, D1 ₅ sampling frequency Do not execute.

Thus, each of the terminal devices 11A to 11N
, The first and second audio processing units 20 and 21 each have a maximum of two first to fifth rate converters 25, 27, 28, 34, 35, 38, 44, 4
The interpolation processing and the thinning processing are executed by using 5, 46, and 48.

In this embodiment, each terminal device 1
1A to 11N, the first and second encoding units 39
And the output stage of 49 and the input stage of the first and second decoding sections 24 and 43 have a 16-bit data transmission path,
Treatment encoded output audio data D1 ₃ and D1 ₅ by the first and second and fourth encoding method, the sampling frequency may be transmitted in sequential 8-bit units becomes 8 [kHz] (64 [kbps] (Which has a transmission rate of the order), the second encoded audio data D2 ₂ and D2 ₄ can be generated, and the second encoded audio data D2 ₂ and D2 ₂ and to D2 ₄ adds the first and second identification number (for example, 2 bits).

When the output audio data D1 ₃ and D1 ₅ are encoded by the third encoding method, the sampling frequency is usually 1.6 [kHz] and can be transmitted sequentially in 10-bit units (16). having [kbps] of about transmission rate) but to produce a second speech encoded data D2 ₂ and D2 _4, the second speech encoded data D2 ₂ and D2
_{4 is set} to a sampling frequency of 8 [kHz], and is sequentially transmitted in units of 2 bits (having a transmission rate of about 16 [kbps]), thereby utilizing the 14-bit vacancy of the data transmission path. adding first and second identification numbers to the speech encoded data D2 _2, and D2 _4.

[0089] In this way each terminal device 11A to 11N, (including first and second identification number) second speech encoded data D2 ₂ and D2 ₄ 10-bit units at maximum
Therefore, the processing load inside the apparatus can be reduced because the data is sequentially transmitted.

In each of the terminal devices 11A to 11N, the first and second audio processing units 2 are provided on one semiconductor chip.
Since the terminals 0 and 21 are manufactured in a lump, the mounting of various circuit elements on the circuit board can be simplified when the terminal devices 11A to 11N are manufactured.

(3) Operation and effect of the present embodiment In the above configuration, the television conference system 1
0, when the third encoding system is selected as the encoding system in the first audio processing unit 20 in each of the terminal devices 11A to 11N, the output audio data D1 obtained via the first adder 26 ₃ through the first encoding unit 39 processes encoded by the third encoding scheme, resulting second video coded data D3 ₂ and the corresponding second encoded audio data D2 ₂ Time-division multiplexing is performed together with control data D4 indicating that the third encoding method has been selected, and the obtained multiplexed data D5 is transmitted to the next-stage terminal devices 11A to 11N via the public network 12.

The terminal devices 11A to 11N at the next stage
Then, the multiplexed data D5 is taken into the demultiplexer 19 via the first interface circuit 18, and the multiplexed data D5 is converted into the first audio encoded data D2 ₁ , the first video encoded data separated into D3 ₁ and control data D4, sends out the first audio encoded data D2 ₁ to the first decoding unit 24, and sends the first image coded data D3 ₁ to the image processing unit, Also, it sends out the control data D4 to the control unit 15.

At this time, the control unit 15 operates the third decoder 63 of the first decoding unit 24 based on the control data D4, and the third decoder 63 performs the first audio coding. the data D2 ₁ via the synthesis filter 70 and decoding processing by the third decoding method, the input audio data D1 ₁ obtained by performing the filtering process by Tsusuru post filter 71 is sent to the speaker 33 side At the same time, the signal is transmitted to the first adder 26 without being subjected to the filtering process by not passing through the post filter 71.

Thus, each of the terminal devices 11A to 11N
In, the first adder 26, the input speech data D1 ₁ obtained without passing through the post-filter 71, synthesized to the transmission voice data D1 ₂ generated in its own terminal apparatus 11A~11N adding process , the output audio data D1 ₃ obtained through the first encoding unit 39, and processed encoded by any of the first to fourth coding scheme selected, the resulting second speech code second video coded data D corresponding to data D2 ₂ via the multiplexer 41
After time-division multiplexed together with 3 _2, and control data D4, successively through and be transmitted to the next stage of the terminal device 11A~11N the second interface circuit 42 and the public network 12.

In each of the terminal devices 11A to 11N, when the third encoding system is selected as the encoding system in the second audio processing unit 21, the second audio processing unit 21
Performs the same processing as in the first audio processing unit 20 described above.

Therefore, this video conference system 10
So in each terminal device 11A~11N selected third coding scheme, the output audio data D1 ₃ and or D1 ₅
It was treated encoded by the third encoding scheme, even when transmitting the second audio encoded data D2 ₂ and or D2 ₄ obtained sequentially next stage and or front of terminal devices 11A to 11N, the next And / or previous terminal devices 11A-11
The first speech encoded data D2 ₁ and or D2 ₃ given from N when processing decoded by the third decoding method, the input speech to be transmitted to the next stage and or preceding terminal device 11A~11N the data D1 ₁ and or D1 ₄ since it is combined with transmission voice data D1 ₂ without passing through the post-filter 71, that the input audio data D ₁ and or the D1 ₄ is attenuated by the effect of the post filter 71 in the transmission Can be prevented.

Conventionally, when such a television conference system 10 is constructed, for example, the first and second teleconference systems are used.
Output audio data D1 ₃ in the encoders 39 and 49 of FIG.
And D1 ₅ treated encoded by the first to third encoding scheme, also the first speech encoded data D2 ₁ and D2 ₃ in the first and second decoding unit 24 and the 43 first to When the decoding process is performed by the third decoding method, the first rate converters 25 and 44 and the fifth rate converter 3
8 and 48, the second or third rate converter 27, 4
It is conceivable to use three rate converters, 5 or 28,46.

[0098] That is the sampling frequency of the first and second decoding unit 24 and the input audio data D1 ₁ and D1 ₄ obtained from 43 and the interpolation processing through a first rate converter 25 and 44, this input audio data D1 ₁ and D1 ₄
After you by adding processing and transmission audio data D1 ₂ through the first and second adders 26 and 36 corresponding synthesis, sampling frequency of the output audio data D1 ₃ and D1 ₅ obtained in the fifth while thinning processing through the rate converter 38 and 48, the sampling frequency of the first and second decoding unit 24 and the input audio data D1 ₁ and D1 ₄ obtained from the 43 second or third rate converter Interpolation processing is performed via 27, 45 or 28, 46.

However, in the video conference system 10 according to the present embodiment, the first to fifth rate converters 25, 27, 28, 34, 35, 38, 44, 4
5, 46, 48, at most two of the first to fifth rate converters 25, 27, 28, 34, 35, 38,
Since the interpolation process and the decimation process are performed using 44, 45, 46, and 48, the load of the data processing inside the apparatus can be reduced as compared with the above-described case, and the interpolation process and the decimation process are repeated. input audio data D1 ₁ generated by, D1 ₄ and transmitted voice data D1
₂ and output the audio data D1 _3, D1 ₅ of degradation can be minimized.

Therefore, this television conference system 10
In, with minimal degradation of the output audio data D1 ₃ and D1 _5, among the first to fourth encoding method, encoding processing using the first to fourth encoding method desired Thus, the usability of the entire system can be improved.

Further, the television conference system 10
In, the first speech encoded data D2 ₁ and D2 ₃ transmitted from the next stage and the previous stage of the terminal unit 11A to 11N, as well as recognition by adding the corresponding first and second identification numbers, the next stage and second speech encoded data D2 ₂ and D2 ₄ to be transmitted to the front stage of the terminal unit 11A to 11N, from processing by adding the corresponding first and second identification numbers, each terminal connected in cascade In the devices 11A to 11N,
Can be transmitted sequentially next stage and the previous stage of the second to be transmitted to the terminal device 11A~11N encoded audio data D2 ₂ and D2 ₄ to accurately.

[0102] According to the above configuration, the next stage and or first audio encoded data D2 ₁ obtained by performing encoding processing by the third coding method from the preceding terminal device 11A~11N
And or if the D2 ₃ is transmitted, the terminal device 11A~11N which receives this, performs decoding processing by the third decoding method corresponding to the first audio encoded data D2 ₁ and D2 _3, this The input audio data D to be transmitted to the next and / or previous terminal devices 11A to 11N obtained at this time.
1 transmission ₁ and or D1 ₄ a and coding process after combining the transmission voice data D1 ₂ generated in its own terminal apparatus 11A~11N without passing through the post-filter 71 at the next stage and or preceding terminal device 11A~11N by which is adapted to, each terminal input audio data D1 ₁ and D1 ₃ device 1
In 1A-11N, it is possible to prevent the attenuation of the input audio data to be transmitted by preventing the input audio data to be transmitted from deteriorating by passing through the post-filter 71 sequentially, thereby preventing the input audio data to be transmitted from deteriorating. A video conference system that can transmit the video.

(4) Other Embodiments In the above embodiment, each of the terminal devices 11A to 11A
Although the case where 11N is configured as shown in FIG. 2 described above has been described, the present invention is not limited to this, and each terminal according to another embodiment shown in FIG. 10 in which parts corresponding to those in FIG. As in the devices 80A to 80N, a voice detector 81 is provided between the echo canceller 32 and the fourth rate converter 35 of the second voice processing unit 21.
Sound based on the transmission voice data D1 ₂ provided from the echo canceller 32 detects whether the silence by,
When the sound is silent, the connection between the echo canceller 32 and the fourth rate converter 35 may be cut off.

Thus, when there is no sound, the microphone 3
The next stage or previous stage of the terminal device as a transmission audio data D1 ₂ noise environment (atmosphere) for collecting by 0 11A～11
N can be prevented from being transmitted, and thus the quality of transmitted voice can be improved.

Each terminal 82A-82 according to another embodiment shown in FIG. 11 in which parts corresponding to those in FIG.
Like 82N, a headset 83 (or a handset (not shown)) provided with a microphone and a speaker may be provided instead of the microphone 30 (FIG. 2) and the speaker 33 (FIG. 2). In this case, in each of the terminal devices 82A to 82N, since the headset 83 (or a handset not shown) has the function of the echo canceller 32 (FIG. 2), the echo canceller 32
, The circuit configuration can be simplified.

[0106] Further in the above embodiment, the output audio data D1 ₃ and D1 ₅ in the first and second encoding section 39 and 49 treated encoded by the first to third encoding scheme, also First and second decoding units 24 and 43
In decoding the first coded audio data D2 ₁ and D2 ₃ by the first to third decoding schemes,
And the case where only the fourth rate converters 28, 34, 35 and 46 are operated. However, the present invention is not limited to this, and as shown in FIG.
Operating the rate converter 27,34,35,45 only, the input audio data D1 ₁ and D1 ₄ for delivering the first and second decoding unit 24 and 43 to the third adder 29
Is converted to the second rate converters 27 and 45.
, The first and second audio processing units 20 and 21 can be configured without the third rate converters 28 and 46, and thus the terminal devices 11A to 11N can be configured. Can be simplified.

Further, in the above-described embodiment, each of terminal apparatuses 11A to 11N selects and uses desired first to fourth encoding schemes among the first to fourth encoding schemes. However, the present invention is not limited to this, and the conjugate structure algebraic code-excited linear prediction (CS-ACE
LP: Conjugate Structure Algebraic Code Excited Line
ar Prediction), the maximum likelihood quantization type multi-pulse (MP-MLQ excitation: Multi-Pulse Maxi
An encoding method called mum Likelihood Quantization excitation, and algebraic code excitation type linear prediction (ACELP: Al
It is also possible to select other various coding methods, such as a coding method devised based on code-excited linear prediction, such as a coding method called gebraic Code Excited Linear Prediction).

[0108] Incidentally the code excited type linear predicted and coding scheme devised as a base, the filtering as the input audio data D1 ₁ and D1 ₄ is attenuated by repeating the filtering process with a predetermined filter such as postfilter 71 when processing such as using coding scheme corresponding to a predetermined decoding scheme that requires, upon decoding of the corresponding first audio encoded data D2 ₁ and D2 _3, obtained input audio data D1 outputs by performing a filtering process in order to sound the ₁ and D1 _4, may be output without performing the filtering process in order to transmit, thereby similarly to the above-described embodiment to prevent deterioration of quality be able to.

In a coding method called maximum likelihood quantization type multi-pulse or a coding method called algebraic code excitation linear prediction, two types of post-filters called a pitch post-filter and a formant post-filter are usually provided to a decoder. from that have been made to use, it is possible to prevent the voice quality deterioration of by the input audio data D1 ₁ and D1 ₄ so as not to pass through the pitch postfilter and or formant postfilter output.

Further, in the above-described embodiment, the second encoded voice data D2
With ₁ ~ D2 ₄ has dealt with the case of transmitting the first and second identification numbers, the present invention is not limited to this, the data transmission path, the second speech encoded data D2 ₁
With ~ D2 ₄ since the first and second identification number have additional free be transmitted, using this idle may transmit the other various information.

Further, in the above-described embodiment, a case has been described in which the terminal devices 11A to 11N are sequentially connected in series via the public line network 12 as shown in FIG. The present invention is not limited to this, and each of the terminal devices 11A to 11N may be cascade-connected in a variety of other states such that the terminal devices 11A to 11N are sequentially cascaded via a predetermined network. Incidentally, when each of the terminal devices 11A to 11N is connected in various states, each of the terminal devices 11A to 11N
The audio processing unit may be provided in each of the terminal devices 11A to 11N in accordance with the number of terminal devices 11A to 11N connected thereto.

Further, in the above embodiment, the multipoint conference terminal device installed at each of the multipoints is
And output the audio data D1 ₃ and D1 _5, the output video data D
3 ₂ and D3 ₄ has dealt with the case of applying the terminal device 11A~11N for transmitting, the present invention is not limited to this, and as a terminal device such as a conference call system for transmitting only the audio data, In addition, the present invention can be widely applied to various multipoint conference terminal devices.

Further, in the above-described embodiment, the first audio coded data transmitted from the preceding and / or the next stage multipoint conference terminal device via the network is fetched and output by the fetching means. The case where the first and second interface circuits 18 and 42 are applied has been described. However, the present invention is not limited to this, and is transmitted from the preceding and / or next stage multipoint conference terminal device via the line network. As long as the first encoded audio data can be captured and output, various other capturing means may be applied.

Further, in the above-described embodiment, a case has been described where the microphone 10 and the analog / digital converter 31 are applied as audio data generating means for generating and outputting transmission audio data to be transmitted. The present invention is not limited to this, and various other audio data generating means may be applied as long as transmission audio data to be transmitted can be generated and output.

Further, in the above-described embodiment, the input audio data output from the decoding means without being subjected to the filtering process and the transmission audio data output from the audio data generating means are synthesized and obtained. Although a case has been described where the first and second adders 26 and 36 are applied as synthesis means for outputting output audio data, the present invention is not limited to this, and filtering processing is performed by decoding means. The input audio data output without being transmitted and the transmission audio data output from the audio data generating means,
As long as the obtained output audio data can be output, various other synthesizing means may be applied.

Further, in the above-described embodiment, the second audio encoded data output from the encoding means is converted into a multi-point signal of a preceding stage and / or a next stage different from the transmission source of the corresponding first audio encoded data. The case where the first and second interface circuits 18 and 42 are applied as transmission means for transmitting to the conference terminal device via the network has been described. However, the present invention is not limited to this. Can be transmitted via the network to the multistage conference terminal device at the previous stage and / or the next stage which is different from the source of the corresponding first voice encoded data. Alternatively, various transmission means may be applied.

Further, in the above-described embodiment, the first identification number assigned in advance to each of the multipoint conference terminal devices of the transmission source corresponding to the first encoded voice data output from the capturing means. Has been described as a case where the demultiplexer 19 is applied as the first identification number adding means for adding and outputting the first speech code. However, the present invention is not limited to this. If it is possible to add a predetermined first identification number assigned in advance to each multipoint conference terminal device of the transmission source corresponding to the data and output the same, various other first identification number adding means are applied. You may do it.

Further, in the above-described embodiment, of the first speech coded data output from the first identification number adding means and provided in correspondence with each of the decoding means, The first and second recognizers 22 and 23 are applied as a plurality of selecting means for selecting only the first coded audio data to which the first identification number is added and sending the selected data to the corresponding decoding means. However, the present invention is not limited to this, and the first audio encoded data output from the first identification number adding means is provided corresponding to each decoding means. If it is possible to select only the first coded audio data to which a predetermined first identification number is added and send it to the corresponding decoding means, various other selection means may be applied. May be.

Further, in the above-described embodiment, the second voice coded data output from the corresponding coding means is provided in correspondence with each of the coding means, and each of the multi-points of the corresponding transmission destination is transmitted. The first and second adders 40 and 50 are applied as a plurality of second identification number adding means for adding a predetermined second identification number assigned to the conference terminal device in advance and sending the same to the transmission means. However, the present invention is not limited to this, and the present invention is not limited to this, and is provided corresponding to each encoding means, and the second audio encoded data output from the corresponding encoding means is transmitted to the corresponding transmission means. If it is possible to add a predetermined second identification number assigned in advance to each of the above-mentioned multipoint conference terminal devices and send the same to the transmission means, various other second identification number adding means are applied. Even There.

[0120]

As described above, according to the present invention, a multipoint conference terminal device installed at each multipoint and a line network for sequentially subordinately connecting the multipoint conference terminal devices are provided. , Each of the multipoint conference terminal devices captures and outputs the first voice encoded data transmitted from the preceding and / or next stage multipoint conference terminal device via the network, and capture means. Decoding processing of the first coded audio data output from, and performing filtering processing for removing noise in order to emit sound based on the input audio data, and outputting the obtained input audio data;
Decoding means for outputting without performing the filtering processing, audio data generating means for generating and outputting transmission audio data to be transmitted, and input audio data output without performing the filtering processing from the decoding means, Synthesizing means for synthesizing the transmitted audio data output from the audio data generating means and outputting the obtained output audio data; and encoding the output audio data output from the synthesizing means to obtain the second Encoding means for outputting audio encoded data; and a second audio encoded data output from the encoding means for each of a previous stage and / or a next stage different from the source of the corresponding first audio encoded data. Since the point-to-point conference terminal device has transmission means for transmitting the data via the line network, the input voice data to be transmitted is sequentially repeated in each multi-point conference terminal device. Thus, it is possible to realize a multipoint conference apparatus that can prevent attenuation due to the operation of the filtering process and can transmit input audio data to be transmitted while preventing deterioration of audio quality based on the input audio data. it can.

A first step in which the multipoint conference terminal devices installed at the respective multipoints are sequentially cascaded via a predetermined network, and in each of the multipoint conference terminal devices,
The first audio coded data transmitted from the preceding and / or subsequent multipoint conference terminal device via the network is decoded, and the obtained input audio data is converted into a voice based on the input audio data. A second step of performing filtering processing for removing noise in order to emit sound and outputting the processed sound without performing the filtering processing, and generating transmission audio data to be transmitted, and generating the transmission audio data. A third step of synthesizing the input audio data output without being subjected to the filtering process and generating the second encoded audio data by encoding the obtained output audio data; and A fourth step of transmitting, via a network, the voice coded data to the multistage conference terminal device at the preceding and / or next stage different from the source of the corresponding first voice coded data. The input audio data to be transmitted can be prevented from being attenuated by the action of the filtering process repeated in each of the multipoint conference terminal devices in sequence. Can be realized while preventing deterioration in voice quality based on the input voice data.

Further, the first speech coded data supplied from the outside is decoded, and the obtained input speech data is subjected to a filtering process for removing noise in order to emit a speech based on the input speech data. A decoding means for outputting and outputting without performing the filtering processing, an audio data generating means for generating and outputting transmission audio data to be transmitted, and an input output without performing filtering processing from the decoding means. Synthesizing means for synthesizing audio data and transmission audio data output from the audio data generating means, and outputting the obtained output audio data, and encoding means for encoding the output audio data output from the synthesizing means Is provided, a plurality of multipoint conference terminal devices sequentially cascade-connected via a predetermined network are sequentially repeated. Input audio data transmitted by the filtering process can be prevented from being attenuated, and thus the input audio data to be transmitted can be transmitted while preventing the deterioration of the voice quality based on the input audio data. A point conference terminal device can be realized.

[Brief description of the drawings]

FIG. 1 is a block diagram showing one embodiment of a configuration of a television conference system according to the present invention.

FIG. 2 is a block diagram illustrating a circuit configuration of each terminal device.

FIG. 3 is a block diagram illustrating a circuit configuration of first and second encoding units.

FIG. 4 is a block diagram illustrating a circuit configuration of first and second decoding units.

FIG. 5 is a block diagram showing a circuit configuration of a third decoder.

FIG. 6 is a block diagram for explaining the operation of first to fifth rate converters;

FIG. 7 is a block diagram for explaining operations of first to fifth rate converters;

FIG. 8 is a block diagram for explaining operations of first to fifth rate converters;

FIG. 9 is a block diagram for explaining operations of first to fifth rate converters;

FIG. 10 is a block diagram showing a circuit configuration of each terminal device according to another embodiment.

FIG. 11 is a block diagram showing a circuit configuration of each terminal device according to another embodiment.

FIG. 12 is a block diagram for explaining an operation of a rate converter according to another embodiment.

FIG. 13 is a block diagram showing a configuration of a conventional video conference system.

FIG. 14 is a block diagram showing a configuration of a conventional video conference system.

[Explanation of symbols]

10 TV conference system, 11A to 11N,
80A to 80N, 82A to 82N ... terminal device, 15 ...
... Control unit, 18 ... First interface circuit, 19
... Demultiplexer, 22 first recognizer, 23
... second recognizer, 24 ... first decoding unit, 25, 44
... a first rate converter, 26 ... a first adder,
27, 45 ... second rate converter, 28, 46 ...
... third rate converter, 30 ... microphone,
31 ... Analog to digital converter, 34, 35 ...
Fourth rate converter, 36... Second adder, 3
8, 48... Fifth rate converter, 39... First encoding unit, 40... First adder, 42... Second interface circuit, 43. ......
Second encoding unit, 50... Second adder, 57.
Encoder, 63 ...... third decoder, 71 ...... postfilter, D1 _1, D1 ₄ ...... input voice data, D1 ₂
... Transmission audio data, D1 ₃ , D1 ₅ ... output audio data, D2 ₁ , D2 ₃ ... first audio encoded data, D2
₂ , D2 ₄ ... Second audio encoded data.

Claims (11)

[Claims] 1. A multi-point conference terminal device installed at each multi-point, and a line network for sequentially cascading the multi-point conference terminal devices, wherein each of the multi-point conference terminal devices comprises: Capturing means for capturing and outputting first voice coded data transmitted from the multistage conference terminal device of the preceding stage and / or the next stage via the network; and the first voice output from the capturing means. The audio coded data is decoded, and the obtained input audio data is subjected to a filtering process for removing noise in order to emit a sound based on the input audio data and output, and the filtering process is not performed. A sound data generating means for generating and outputting transmission sound data to be transmitted; and a decoding means for outputting the sound data without being subjected to the filtering processing. Synthesizing means for synthesizing the input audio data and the transmission audio data output from the audio data generating means, and outputting the obtained output audio data; and encoding the output audio data output from the synthesizing means. Encoding means for performing the encoding process and outputting the obtained second audio encoded data; and converting the second audio encoded data output from the encoding means into the corresponding first audio encoded data. Transmission means for transmitting the multistage conference terminal device of the preceding stage and / or the next stage different from the transmission source via the network. 2. The decoding means according to claim 1, wherein said decoding means comprises a first decoding method corresponding to an encoding method based on code excitation type linear prediction or said first decoding method requiring said filtering processing. 2. The multipoint conference apparatus according to claim 1, wherein the audio encoded data is decoded. 3. The multi-point conference terminal device, comprising: an output stage for applying the filtering process to the input audio data of the decoding unit and outputting the input audio data; and performing a filtering process on the input audio data of the decoding unit. And an output stage for performing output without performing the processing, an output stage of the audio data generating unit and an input stage of the synthesizing unit, and an output stage of the synthesizing unit and an input stage of the encoding unit. At least four rate converters for converting a sampling frequency of the input audio data, the transmission audio data, and the output audio data; and a sampling frequency defined by a decoding method used by the decoding means; Based on the sampling frequency specified in the encoding method used by the encoding means, each of the rate converters , Multipoint conferencing system according to claim 1, characterized in that it comprises a control means for controlling each of said rate converter so as to convert processing the sampling frequency in two respective said rate converter at the maximum. 4. The multipoint conference terminal device, wherein the decoding means provided corresponding to each of the multipoint conference terminal devices to which the second encoded voice data is transmitted, A plurality of data processing means comprising a synthesizing means and the encoding means; and a plurality of data processing means which are assigned in advance to the multi-point conference terminal devices of the transmission source corresponding to the first coded audio data output from the capturing means. First identification number adding means for adding and outputting the predetermined first identification number, and the first identification number adding means provided in correspondence with each of the decoding means, and the first identification number adding means output from the first identification number adding means. A plurality of selections for selecting only the first audio encoded data to which the predetermined first identification number is added from among one audio encoded data and transmitting the selected data to the corresponding decoding means. Means and on each The second audio encoded data output from the corresponding encoding means is provided in correspondence with each of the encoding means, and the second audio encoded data output from the corresponding encoding means is assigned to each of the multipoint conference terminal devices at the corresponding transmission destination in advance. A plurality of second identification number adding means for adding the second identification number to the transmission means and transmitting the second identification number to the transmission means, wherein the transmission means is configured to output the second identification number from each of the second identification number addition means. 2. The multipoint conference apparatus according to claim 1, wherein the voice coded data is transmitted to each of the corresponding multipoint conference terminal apparatuses based on the second identification number via the line network. 3. 5. A first method for sequentially cascade-connecting multipoint conference terminal devices respectively installed at multipoints via a predetermined network.
In each of the multipoint conference terminal devices, decoding the first encoded voice data transmitted from the preceding and / or next stage multipoint conference terminal devices via the network. ,
A second step of performing a filtering process for removing noise in order to emit the voice based on the input voice data and outputting the obtained voice data without performing the filtering process; By generating the transmitted audio data, synthesizing the generated transmitted audio data with the input audio data output without being subjected to the filtering process, and encoding the obtained output audio data. A third step of generating the second encoded voice data; and transmitting the second encoded audio data to the first and / or second stages different from the transmission source of the corresponding first encoded audio data. A fourth step of transmitting to the point conference terminal via the network. 6. In the second step, the first decoding method corresponding to a coding method based on code excitation type linear prediction or the first decoding method requiring the filtering processing is performed by the first decoding method. The multipoint conference method according to claim 5, wherein the audio encoded data is decoded. 7. In the third step, a sampling frequency specified in a decoding method used in the decoding process of the first coded audio data and a sampling frequency defined in the decoding process of the output audio data are performed. On the basis of the sampling frequency defined in the used encoding method, the sampling frequency of the input audio data and the transmission audio data and the output audio data corresponding to the two predetermined positions at the maximum are calculated. The multipoint conference method according to claim 5, wherein conversion processing is performed as needed. 8. In the second step, a predetermined first identification number previously assigned to each of the multipoint conference terminal devices of the transmission source corresponding to the first voice encoded data is added. In a plurality of processing systems provided corresponding to each of the multistage conference terminal devices of the preceding and / or subsequent stages to which the second audio encoded data is to be transmitted, And selecting and decoding only the first encoded audio data to which the identification number of 1 is added. In the third step, the second encoded audio data generated in each of the processing systems is selected. A predetermined second identification number assigned in advance to each of the multipoint conference terminal devices at the preceding and / or subsequent stages corresponding to the transmission destination, and in the fourth step, each of the second voices Encoded data Multipoint conference method according to claim 5, respectively, characterized in that transmission over the network to each said multipoint conference terminal device corresponding on the basis of the second identification number. 9. A multi-point conference terminal device installed at each of said multi-point conference devices, wherein said first voice code is provided from outside. Decoding the decoded data, performing filtering processing for removing noise in order to emit the sound based on the input sound data, and outputting the obtained input sound data, and outputting without performing the filtering processing Encoding means, audio data generating means for generating and outputting transmission audio data to be transmitted, input audio data output from the decoding means without being subjected to the filtering process, and audio data generating means Synthesizes the output transmission voice data,
A multipoint conference terminal device, comprising: synthesizing means for outputting the obtained output audio data; and encoding means for encoding the output audio data output from the synthesizing means. 10. The decoding means according to claim 1, wherein said first decoding means is a predetermined decoding method corresponding to an encoding method based on code excitation type linear prediction or said predetermined decoding method requiring said filtering processing. 10. The multipoint conference terminal device according to claim 9, wherein the encoded audio data is decoded. 11. An output stage of the decoding means for performing the filtering process on the input audio data and outputting the input audio data, and an output stage of the decoding means for outputting the input audio data without performing the filtering process. The input audio data, the transmission audio data, and the output corresponding to the output stage of the audio data generation unit and the input stage of the synthesis unit, and the output stage of the synthesis unit and the input stage of the encoding unit, respectively. At least four rate converters for converting the sampling frequency of the audio data, the sampling frequency specified in the decoding method used by the decoding means, and the coding method used in the coding method used by the coding means. At most two of the rate converters based on the sampling frequency. 10. The multipoint conference terminal device according to claim 9, further comprising control means for controlling each of said rate converters so that said sampling frequency is converted by said data converter.