JP2828527B2 - Multipoint voice communication system - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for addressing problems such as an increase in the number of communication points accommodated in a multipoint voice communication, an unlimited number of connected terminals, a reduction in transmission costs, and a prevention of sound quality deterioration.

[0002]

2. Description of the Related Art FIG. 3 shows a configuration of a conventional conference trunk for realizing a multipoint voice communication system. This figure is described in Japanese Patent Publication No. Hei 2-24426, in which two conference trunks 11 and 12 are connected, 11 is B ₁ , and 12 is B ₂
Transmitting and receiving the conference ports # 1 to # 3 to the trunk B ₁ as 3
Line, sending and receiving to port # 6 to # 8 in the conference trunk B ₂ 3
Lines are connected, and conference trunks B ₁ and B ₂ are connected using ports # 4 and # 5, respectively.

Hereinafter, the drawings will be described in detail. The conference trunks B ₁ and B ₂ have adders 21 and 22, respectively, where the input signals from ports # 1 to # 3 are a, b and c, and the input signals from ports # 6 to # 8 are f, g and Assuming that h, the input signal connected to the trunk B ₁ is temporarily calculated by the adding circuit 21 of the trunk B ₁ as a + b + c + d (where d = f
+ G + h). Similarly, the adder circuit 22 Trunk _{B 2 e + f + g +} h ( where, e = a + b + c )
Are added.

The output signal of the port # 1 is b + c + d obtained by subtracting the input signal a of the port # 1 by the subtraction circuit 31 from the addition result obtained by the addition circuit 21. Similarly, the output signals of ports # 2 to # 8 are a + c + d and a +
b + d, a + b + c, f + g + h, e + g + h, e + f
+ H, e + f + f. Incidentally, since the trunk B ₁ and trunk B ₂ is connected using a # 5 and port # 4, as apparent from FIG. 3 e = a + b + c , d = f + g + h
Becomes Therefore, the output signal of port # 1 is b + c + d =
b + c + f + g + h, and the sum of all signals of the connection system other than the input signal of the port # 1 is obtained.

The same applies to lines # 2, # 3, # 6, # 7, # 8. In this way, by connecting a large number of trunks with respective ports, it is possible to easily increase the number of connection points.

[0006]

The conventional multipoint voice communication system has a configuration as shown in FIG.
To connect ports of conference trunks that perform addition,
Since audio signals from terminals at each point are connected in the same way as accommodated in each port, by accommodating a large number of terminals that can not be accommodated by a single trunk in multiple interconnected trunks, The number of communication points was easily increased.

Here, a case is considered in which voice communication is performed using a voice signal using high-efficiency coding in order to reduce transmission costs. This is shown in FIG. In the conference trunk 11, the compression codes input from the ports # 1 to # 3 and # 6 to # 8 are decoded into linear signals by the decoders 41 to 43 and 46 to 48, operated by the addition circuit 21, and then decoded by the port 21. # 1 to # 3, #
Encoders 51-53, 56 again for transmission to 6- # 8.
Compression encoding is performed at .about.58. Similarly, in the ports # 4 and # 5, the decoder 44,
Decoding is performed at 45, and encoding processing is performed at encoders 54 and 55 at the time of output.

That is, if a transmission line of the same quality as that of the terminal line is used for connecting the conference trunks as in the related art, encoding and decoding are repeated between the conference trunks. However, in general, the lower the bit rate of the encoding system, the more the encoding distortion is caused. Therefore, repeating the encoding and decoding processes causes deterioration in sound quality.

Table 1 shows the REAL-TIME IMPLE.
MENTATION AND PERFORMANCE
OF A 16KB / S LOW-DELAY CEL
PSPEECH CODER (Juin-Hway C
hen, Melvin J .; Melchner, Ric
hard V. Cox), 3
ADPCM coding method of 2 kb / s and L of 16 kb / s
The evaluation results are shown for the case where the D-CELP coding system is connected in one stage and the case where the codec is connected in multiple stages in a dependent manner.

[0010]

[Table 1]

In Table 1 above, MOS (Mean Op)
The inion score) is a five-level evaluation method of 5 to 1 called an opinion method (5: very good, 4: good, 3: normal, 2: bad, 1: extremely). This is the average of the results of subjective evaluation. From this table, it can be seen that the evaluation when the LD-CELP coding system is connected in multiple stages is remarkably low, and that the voice quality deteriorates when repeated high-efficiency coding and decoding are repeated. That is, in FIG. 4, in a system using a plurality of connections between conference trunks, there is a problem that comfortable voice communication cannot be performed.

[0012] The "multipoint image synthesizing method" disclosed in Japanese Patent Laid-Open No. Hei 2-203692 is disclosed in Japanese Patent Application Laid-Open No. Hei 2-203692. In a video conference system in which a screen is synthesized and transmitted to each point, the image synthesizing unit called a master device re-transmits two types of a normal screen and a reduced screen obtained by thinning out the normal screen. Even if a reduced screen is not created by performing code decoding, the transmitted reduced screen can be arranged at each position of the monitor and the screen can be synthesized.

On the other hand, the present invention relates to a system for holding a conference by conference participants at N or more locations using a plurality of conference trunks connected to conference participants at N locations and performing N-1 voice addition. In order to prevent code decoding from being repeated every time transfer between conference trunks and deteriorating sound quality, the speed is sufficiently high enough that high efficiency compression encoding is not required only for transfer between conference trunks ( Alternatively, in order to reduce the cost, a line (with an encoding speed using high-efficiency compression encoding that can tolerate sound quality degradation) is separately provided.

Therefore, the difference between the present invention and the preceding example is that the type of information to which the preceding example can be applied is a type in which data that does not require decoding to perform synthesis can be created in advance, and its applicable range Can be said to be an invention in the case where both the number of points to which the master device (the conference trunk in this patent) can be directly connected are satisfied. (Since one of the reduced screens of the normal screen is selected by the selection unit 17 for the output of the master device, only one of them is transmitted to the output destination master device, so it is considered impossible to connect the master devices.) To solve the problem of information quality degradation caused by repeated encoding and decoding in multipoint communication, there is a "multipoint image synthesizing method" disclosed in JP-A-2-203692.

In this prior patent, when the number of conference participation points is two, the image data sent from the other party is left as it is,
An image called a master device is transmitted by transmitting two types of a normal screen and a reduced screen obtained by performing a thinning process on the normal screen in a video conference system in which a screen is combined and transmitted to each point in a case where the number is three or more. It is possible to arrange the transmitted reduced screen at each position of the monitor and perform the screen synthesis without having to re-encode and decode to create a reduced screen.

However, in the prior art, like the synthesis of screens, the screens transferred from each point are reduced independently and each screen is arranged at a designated place on one monitor. It can be applied to the synthesis processing of the type of information that does not affect the information, but cannot be applied to the linear operation processing that adds information from each point like voice information. In addition, in order to realize the preceding invention, it is necessary to provide each terminal device with a reduced screen creation function by thinning-out processing, and the number of circuits is expected to increase. In addition, when two types of video data, that is, a normal screen and a reduced screen, are transmitted, there is a problem that the transmission cost rises and reduction cannot be expected at all.

SUMMARY OF THE INVENTION The present invention has been made to solve the above-described problems. In order to perform multipoint voice communication at low cost, a plurality of conference trunks are used when a high-efficiency voice coding technique is used. It is an object of the present invention to obtain a comfortable voice communication method without deteriorating voice quality even when communication is performed by connecting two or more peers.

[0018]

SUMMARY OF THE INVENTION In order to solve the above-mentioned problems, the first aspect of the present invention provides a multipoint voice communication system for interconnecting conference trunks as a line connecting conference trunks. In addition to a low-bit-rate line using a high-efficiency coding system that connects each terminal and the conference trunk, a high-speed transmission path that can transmit between the conference trunks without performing coding and decoding processing is provided.

In the second invention, each conference trunk
Lower than the bit rate when the connection line is not compression-coded
In addition to the fast low-speed transmission path,
That of the input and output stages of each connected conference trunk
In each case, compress and code to the bit rate of the low-speed transmission path.
Are also suitable for the input and output stages of the conference trunk for the terminal.
Performs encoding with less encoding distortion than the encoding method used
It is provided with a decoder and an encoder .

[0020]

[Action] multipoint voice communication method in the first invention, the connection times line with the conference trunk and each terminal reduces the transmission rate by using the high-efficiency coding scheme, aims to reduce transmission costs, cut The connection line between the conference trunks is equipped with a separate transmission line at a different speed from the one that connects each terminal and the conference trunk, and does not perform encoding and decoding, or at least the distortion of the connection between each terminal and the conference trunk By using less than the coding method applied to the above, it is possible to realize a high-efficiency voice code with remarkable deterioration of sound quality and comfortable voice communication without repeating decoding.

In the multipoint voice communication system according to the second aspect of the present invention , the connection lines of the respective conference trunks are compressed.
To a low-speed transmission line that is slower than the bit rate when
And at the same time,
The above low speed is applied to each of the input and output stages of the conference trunk.
Even if the code is compressed to the bit rate of the transmission path,
Coding scheme applied to the input and output stages of a conference trunk
Decoder and code for performing coding with less coding distortion
The cost of the system can be reduced by
And compression encoding of connections between each conference trunk
If you use a low-speed transmission path that is slower than the
Thus, even if coding / decoding is repeated, the coding distortion is small.
And can prevent deterioration of sound quality

[0022]

FIG. 1 shows an embodiment of a multipoint communication system according to the present invention, in which 11, 12, 21, 22, 31 to 38 are the same as the conventional one shown in FIG. The Port # 1 to # 3 and the conference trunk B ₁ and # 6 #
In 8 and connection path conference trunk B _2, conference trunk B _1, high efficiency decoder 41 to 43 in the input stage of B ₂ and 4
6 to 48, and the output stage is equipped with high-efficiency encoders 51 to 53 and 56 to 58 to be paired with the output stages. Further, a transmission line having a bit rate capable of transmitting the signal form subjected to the audio signal addition processing without encoding in the conference trunk is newly provided as a connection line between the conference trunks.

FIG. 2 shows an embodiment of a multipoint voice communication system using the present invention.
Reference numerals 1, 22, 31 to 38 are the same as those of the conventional example shown in FIG. And at port # 1 to # 3 and the conference trunk B ₁ and # 6 to # 8 and the connection path of the conference trunk B _2, the high efficiency decoder 41 to 43 and 46 to 48 in the input stage of the conference trunk, the output stage Are equipped with high-efficiency encoders 51-53 and 56-58 that are paired with the encoders. Further, the connection between the conference trunks includes a transmission line having a lower speed than the bit rate when no compression encoding is performed, and at least ports # 1 to # 3 and # 6 even if the bit rate of the newly equipped transmission line is compression-coded. Decoders 64 and 65 and encoders 74 and 75 for performing encoding with less encoding distortion than the encoding methods applied to # 8 are provided at the input and output stages of the conference trunk interconnection.

Next, the operation will be described in detail. In FIG. 1, a multipoint communication system according to the present invention employs a high-efficiency voice coding system, and a signal transferred from each terminal is, for example, a 16 kb / s APC (Adaptive P).
(reductive coding: linear predictive coding)
It is assumed that high-efficiency coding of the system has been performed. B ₁ ,
To perform an operation in each adder circuits 21 and 22 shown in B ₂ is a linear PCM audio signal that is APC coded
It needs to be converted to a signal. This processing is performed by the decoders 41 to 4
3 and 46-48. Therefore, ports # 1 to # 3
Let A, B, and C be the high-efficiency coded signals input from the decoders 41 to 43, and convert the respective signals to linear linear signals a, b, and 14-bit having a sampling period of 5 kHz and 14 bits, for example.
c.

Similarly, assuming that the signals input from ports # 6 to # 8 are F, G, and H, the signals are
At 48, the signals are converted into linear signals f, g, h having a sampling period of 8 kHz and 14 bits. The connection line between the trunk 11 and the trunk 12 is a high-speed transmission line of 112 bits that can transmit this 14-bit linear signal without compression encoding. Therefore, the linear signals d and e (where d = f + g
+ H, e = a + b + c), the conventional processing is performed without performing the above-described coding and decoding processing, and each port obtains the sum of all signals of the connection system except the signal transmitted from the port to the conference trunk. . Then, each signal is re-encoded by the encoder 5
APC encoding of 16 kb / sk is performed at 1 to 53 and 56 to 58, and output to each terminal from ports # 1 to # 3 and # 6 to # 8.

In a multipoint communication system using a high-efficiency voice coding technique to reduce the speed of a line connecting each terminal to a conference trunk, a high-speed transmission path as shown in FIG. If used, the problem of sound quality degradation is solved, but the price can be further reduced by the method shown in FIG. Instead of 112KB / s of the line shown in FIG. 2 in FIG. 1, equipped with a line of example 32 kb / s of at least a lower bit rate than the line, the coding scheme ADP
Use CM. Thus, the cost can be reduced as compared with FIG. Also, 112 kb / s linear coding line and 3
Conference trunk such as 2 kb / s ADPCM coded line
Used for terminal line acquisition without using a dedicated line for connection
16 kb / s LD-CELP input / output line
By connecting to each other, 16 kb / s LD-CELP
When a connection between conference trunks is made with a subordinate connection of the encoding method
Conference traffic over 32 kb / s ADPCM coded line
This method for connection between links is 32 kb / s ADPCM code.
32 kb / s even if coding / decoding is repeated by the coding method
LD-C towards the ADPCM encoding method is 1 6kb / s of
Since the encoding distortion is much smaller than the ELP encoding method (see Table 1), even if the encoding / decoding is repeated between ports # 4 and # 5 interconnecting each conference trunk, the encoding distortion is reduced.
And the deterioration of sound quality can be prevented.

The operation of FIG. 2 is as follows.
In FIG. 2, similarly to FIG. 1, each of the ports # 1 to # 3 and # 6 to # 8
, A signal compressed and encoded to 16 kb / s is input,
This signal is converted into a linear signal by the decoders 41 to 43 and 46 to 48, and N-1 addition processing is performed by the same processing as in FIG. However, since the conference tracks are connected by a line having a transmission rate of 32 kb / s, signals d and e transmitted between ports # 4 and # 5 (where d = f + g + h, e =
a + b + c) are converted from linear signals by encoders 74 and 75 to AD
PCM code D. E and transferred to the decoders 64 and 6
At 5, the ADPCM code is again converted into linear signals d and e. Thereafter, the same operation as that of FIG.
Multipoint voice communication can be realized between the terminals connected to # 3, # 6, and # 8.

In the above embodiment, the signal form between the terminal and the conference trunk is 16 kb / s APC code.
The purpose of the present invention is to reduce the transmission speed by companding coding or high-efficiency coding to reduce the cost. For example, if the coding speed is 16 kb / s, the LD-
In the case of the CELP system or 32 kb / s, the same effect can be obtained by a system using an effective high-efficiency code aiming at a low speed such as the ADPCM coding system or the ADM coding system. Similarly, in the above-described embodiment in FIG. 2, the 32 kb / s ADPCM encoding method is used for the connection lines between the conference trunks. If the transmission speed between the terminal and the conference trunk is 16 kb / s, 32 kb / sA
The same effect can be obtained by the DM coding method or the PCM coding method of 64 kb / s PCM. In the present embodiment, the number of conference trunks to be connected is two, but the same effect can be obtained when two or more conference trunks are connected. As an example, FIG. 5 shows an embodiment in which the number of conference trunks is three.

[0029]

As described above, according to the first aspect,
In a multipoint voice communication system that can easily realize a conference trunk that can accommodate more terminals than a single conference trunk can accommodate by connecting multiple conference trunks, the The transmission cost can be reduced by applying a high-efficiency audio coding technique to the transmission of the audio signal and setting a low bit rate.

According to the second invention, each conference trunk
Than the bit rate when the compression connection is not encoded
In addition to performing on a low-speed low-speed transmission line,
The input and output stages of each of the interconnected conference trunks.
The compression code is set to the bit rate of the low-speed transmission
The input and output stages of the conference trunk for the terminal
Encoding with less encoding distortion than the encoding method applied to
System with a `` decoder and encoder to perform
System cost, and each conference trunk
Than the bit rate when the compression connection is not encoded
The coding / decoding is repeated by performing on a low-speed low-speed transmission path.
Even if returned, there is little coding distortion, and deterioration of sound quality is prevented.
There is an effect that can be.

[Brief description of the drawings]

FIG. 1 is a configuration diagram of an embodiment of a multipoint voice communication using the first invention of the present invention.

FIG. 2 is a block diagram of an embodiment for further reducing the transmission cost in multipoint voice communication using the second invention.

FIG. 3 is a configuration diagram of a conventional example that realizes multipoint voice communication.

FIG. 4 is a configuration diagram of a conventional multipoint voice communication in which cost reduction is realized.

FIG. 5 is a configuration diagram of an embodiment when three conference trunks using the present invention are connected.

[Explanation of symbols]

11 Conference trunk B ₁ 12 Conference trunk B ₂ 21,22,23 Adder circuit 41-49,410 16 kb / s LD-CELP decoder 51-59,510 16 kb / s LD-CELP encoder 63-68 32 kb / s ADPCM decoding 73-78 32 kb / s DPCM encoder

──────────────────────────────────────────────────続 き Continuation of the front page (56) References JP-A-59-156058 (JP, A) JP-A-63-54859 (JP, A) JP-A-2-309458 (JP, A) JP-A 61-1986 49548 (JP, A) JP-A-61-170173 (JP, A) (58) Fields investigated (Int. Cl. ⁶ , DB name) H04M 3/42 H04M 3/56 H04Q 3/58

Claims (2)

(57) [Claims] 1. An audio signal transmitted from a terminal at N different points is received as an input signal from each input stage, and a total of N-1 audio signals excluding the signal at the point is obtained as N.
In a multipoint voice communication system in which a plurality of conference trunks configured to output the sum of N voice signals from an output stage as a reception signal at the point are output from an output stage, each terminal and the conference trunk are connected. A high efficiency encoder / decoder is provided between the input stage and the output stage of
A multipoint voice communication system characterized in that a high-speed transmission path capable of transmitting a signal without compression encoding is provided in a connection between the conference trunks. 2. An audio signal transmitted from a terminal at N different points is received as an input signal from each input stage, and a sum of N-1 audio signals excluding the signal at the point is obtained as N. In a multipoint voice communication system in which a plurality of conference trunks configured to output the sum of N voice signals from an output stage as a reception signal at the point are output from an output stage, each terminal and the conference trunk are connected. high efficiency coding / decoding apparatus provided with compression coding, each conference tiger between the input and output stages
Bit rate when compression encoding is not performed on connections between links
This is performed on the lower-speed transmission line,
Input stage and output of each conference trunk interconnected by
Each stage compresses to the bit rate of the low-speed transmission path
Even if it is coded, the input stage of the conference trunk
Code with less coding distortion than the coding method applied to the power stage
A multipoint voice communication system comprising a decoder and an encoder for performing conversion .