JPH0583288A - Cell transmission control method - Google Patents

Abstract

(57) [Abstract] [Object] The present invention relates to a cell transmission control system for converting a voice signal into cells and transmitting the cells, without reducing transmission efficiency.
The purpose is to prevent discomfort during a call. [Structure] The transmission side 1 includes a voice detection unit 4 and an encoding unit 5.
And a cell assembling unit 6, and when the voice detecting unit 4 detects a voiced section with a voice signal, the encoding unit 5 performs high-rate encoding and detects a silent section without a voice signal,
The encoding unit 5 performs low rate encoding, and the cell assembling unit 6
The high-rate coded output signal is used as a voiced cell, the low-rate coded output signal is used as a silent cell, and the cell discard priority of this silent cell is set to be high, and is sent to the receiving side 3 via the exchange network 2. To do.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a cell transmission control system for converting a voice signal into cells and transmitting the cells. ATM (Asyncro
The nous Transfer Mode) method does not include an error detection / retransmission procedure and has a constant packet (cell) length, and in order to enable speedup, even when the switching network is congested There is no procedure for stopping the transmission of cells, and therefore, when the buffer memory of the node of the switching network becomes full, the arrival cells cannot be temporarily stored, resulting in the loss of cells. A voice signal can also be converted into cells and transmitted, and in that case, it is desired to enable efficient transmission without degrading voice quality.

[0002]

2. Description of the Related Art FIG.
1 is a telephone, 52 and 72 are terminal devices for transmitting a modem signal such as a facsimile, 53 and 73 are hybrid circuits,
Reference numerals 54 and 74 are AD converters (A / D), 55 and 75 are echo cancellers, 56 and 76 are modem detection units that detect a modem signal or a voice signal, 57 and 77 are high efficiency encoding units,
Reference numerals 58 and 78 are voice detection units for detecting whether there is sound or no sound.
Reference numerals 9 and 79 are cell assembly parts. Further, 62 and 82 are cell disassembling units, 63 and 83 are decoding units, 64 and 84 are DA converters (D / A), 65 and 85 are noise generating units, and 66 and 86 are adders. In general, 54 to 59 and 74 to 79 form a voice coder, 82 to 86 and 62 to 66 form a voice decoder, and a set of the voice coder and the voice decoder is called a voice codec.

Voice signals from the telephones 51 and 71 are passed through hybrid circuits 53 and 73 and AD converters 54 and 74.
Is converted into a digital signal by the echo canceller 5
5 and 75 cancel the echo component due to the sneak from the partner hybrid circuit, and the modem detection units 56 and 7
6, a high-efficiency coding unit 57, 77, a voice detection unit 58,
In addition, in the case of voice signals from the telephones 51 and 71, voice detection signals from the modem detection units 56 and 76 are added to the cell assembly units 59 and 79. In addition, the terminal devices 52, 7
In the case of the modem signal from 2, the modem detection signal is added to the cell assembling parts 59 and 79.

Further, the AD converters 54 and 74 allow, for example, 8 k.
If a high-efficiency encoder 57, 77 performs, for example, 4-bit ADPCM encoding on a PCM signal having a 64-bit bit rate of 64 Hz / sampling of Hz samples, and encodes the bit rate of 32 kb / s. It becomes an output signal. This encoded output signal is applied to the cell assembling units 59 and 79, and in the case of a voice signal, the voice detecting units 58 and 78 detect a voiced section with a voice signal and a voiceless section without a voice signal. Then, the cell assembling units 59 and 79 convert the encoded output signal of the voice signal in the voiced section into cells, and send the cells to the receiving side via the switching network.

On the receiving side, the cells received by the cell disassembling units 62 and 82 are disassembled and the coded signals are decoded into the decoding units 63 and
In addition to 3, the decoded output signal is added to the DA converters 64 and 84 via the adders 66 and 86 to be converted into an analog signal, and the telephone 5 is connected via the hybrid circuits 73 and 53.
A call can be made between the telephones 51 and 71 by adding the telephone numbers to the telephones 71 and 71. In that case, since no cell is transmitted in the silent section, when the silent section is identified by the cell disassembling sections 62 and 82, noise signals of a constant level from the noise generating sections 65 and 85 are added to the adders 66 and 86. , It is avoided that the silent section becomes completely silent.

In the ATM system, as described above, H
Unlike the DLC procedure, it does not use the error detection / retransmission procedure and transmits as a cell of a fixed length. When the buffer memory in the ATM node becomes full, the transmission of cells to the transmission side is stopped. Since there is no procedure for this, arriving cells that cannot be stored in the buffer memory are discarded. In that case, there is known a method of avoiding the discard of the data cell by lowering the cell discard priority of the cell of the data from the terminal devices 52, 72 and the like as compared with the cell of the voice signal. That is, in the cell assembling sections 59 and 79, when the modem detecting signals are added from the modem detecting sections 56 and 76, the cell discard priority of the cell to be transmitted is set low to avoid the cell discard at the time of congestion. When the voice detection signal is added, the cell discard priority of the cell to be transmitted is set high to allow the cell discard at the time of congestion.

[0007]

DISCLOSURE OF THE INVENTION Problems to be Solved by the Invention As described above, since no cell is transmitted in the silent section, the receiving side outputs a white noise signal of a constant level from the noise generators 65 and 85 in the silent section. , Is to be inserted in place of the background noise signal on the transmission side, but since the frequency characteristics and the level are significantly different from the background noise on the transmission side in the sound section,
There is a drawback that a sense of discomfort occurs each time the voiced section and the silent section are switched. Also, if the background noise in the silent section is transmitted in cells without distinguishing between the silent section and the voiced section, the above-mentioned uncomfortable feeling on the receiving side does not occur,
There is a drawback that the transmission efficiency cannot be improved. An object of the present invention is to prevent discomfort during a call without reducing transmission efficiency.

[0008]

The cell transmission control system of the present invention will be described with reference to FIG. 1, which is a cell transmission control system in which cells are transmitted from a transmission side 1 to a reception side 3 via a switching network 2. In the transmitting side 1, a voice detecting section 4 for detecting a voiced section with a voice signal and a silent section without a voice signal, and high rate coding and low rate coding according to a detection signal of the voice detecting section 4. And an encoding unit 5 for switching to and
And a cell assembly unit 6 for converting the encoded output signal of
The encoding unit 5 performs high-rate encoding of the voice signal in the voiced section and low-rate encoding of the background noise signal in the silent section, and the cell assembly unit 6 outputs the high-rate encoded output signal of the voiced section in the voice cell. , The low-rate encoded output signal in the silent section is cellized as a silent cell, and the cell discard priority of this silent cell is
Set higher than the cell discard priority of voiced cells, and switch network 2
It is sent to the receiving side 3 via.

Further, the cell assembling unit 6 has a low rate code of the background noise signal in a predetermined section of the silent section at an interval longer than the interval at which the high-rate encoded output signal of the voice signal in the sound section is transmitted as a voiced cell. It is provided with a structure for transmitting the converted output signal as a silent cell.

[0010]

The voice detecting section 4 detects whether the voiced section or the silent section is based on the level comparison or the like. Encoding unit 5
Is to add a coded output signal by either high-rate coding or low-rate coding to the cell assembling section 6, and perform high-rate coding by the detection signal of the voiced section from the voice detecting section 4. , Low-rate coding is performed by the detection signal of the silent section. That is, the voice signal in the voiced section is encoded by the high rate encoding similar to the conventional example, and the background noise signal in the silent section is encoded by the low rate encoding. Therefore, it becomes possible to improve the transmission efficiency as compared with the case where each of the voiced section and the silent section is simply made into a cell, and at the receiving side, it is possible to improve the transmission efficiency in a state close to the background noise of the transmitting side. Since the noise can be reproduced in the silent section, there will be no discomfort. Moreover, since the cell discard priority of a silent cell is set to be high, when the switching network 2 is congested, this silent cell is preferentially discarded. However, since it is a silent section, it does not affect the content of the call.

Further, the cell assembling section 6 does not send out all the silent sections as silent cells, but by converting the background noise signal into cells at a certain interval, the transmission efficiency can be further improved. .. In that case, the receiving side can reproduce the background noise in the silent section on the basis of the silent cells that are randomly received.

[0012]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS FIG. 2 is an explanatory view of an embodiment of the present invention.
Reference numerals 1 and 31 are telephones, 12 and 32 are terminal devices for transmitting modem signals such as facsimiles, 13 and 33 are hybrid circuits, 14 and 34 are AD converters (A / D), and 15 and 35.
Is an echo canceller, 16 and 36 are modem detection units for detecting whether it is a modem signal or a voice signal, 17 and 37 are high rate coding units, 18 and 38 are low rate coding units, and 19 and 39 are voiced sections or silences. Voice detection unit for detecting whether it is a section, 20, 4
0 is a selector, 21 and 41 are cell assembling sections, 22 and 42 are cell disassembling sections, 23 and 43 are decoding sections, and 24 and 44 are DA sections.
It is a converter (D / A).

High rate coding unit 1 in this embodiment
7, 37, the low rate coding sections 18, 38, and the selectors 20, 40 constitute the coding section 5 in FIG. 1, and the high rate coding sections 17, 37 are similar to those in FIG. Similar to the high efficiency encoding units 57 and 77 of the conventional example shown in FIG.
It can be a 4-bit ADPCM encoder. Therefore, the transmitting side corresponds to the configuration in which the low rate coding units 18 and 38 and the selectors 20 and 40 are added to the conventional example shown in FIG. The low rate coding units 18, 38
Can be, for example, a 3-bit ADPCM encoder. On the receiving side, it is not necessary to insert noise into the silent section, so that the noise generators 65 and 85 in the conventional example shown in FIG. 4 are omitted.

The voice detectors 19 and 39 detect a voiced section with a voice signal and a voiceless section without a voice signal based on the level and the like, and control the selectors 20 and 40 by the detection signal. The low rate coding output signals from the low rate coding units 18 and 38 are selected according to the section detection signal and added to the cell assembling units 21 and 41. In the case of the voiced section detection signal, the high rate coding unit 17 and The high-rate encoded output signal of 37 is selected and added to the cell assembling parts 21 and 41. Further, by the modem detection signals from the modem detection units 16 and 36, digitized facsimile signals and the like are added to the cell assembling units 21 and 41 with or without the high rate encoding units 18 and 38 to assemble a data cell. The cell discard priority of the data cell is set to the lowest. In addition, a silent cell is assembled by the voice detection units 19 and 39 to set a silent cell and the cell discard priority is set to the highest level.
A voiced cell is assembled and its cell discard priority is set lower than the cell discard priority of a silent cell.

The cells assembled in the cell assembling sections 21 and 41 are sent to the receiving side through an exchange network (not shown). On the receiving side, the cell disassembling units 22 and 42 disassemble the received cells, the encoded signals of the information fields are added to the decoding units 23 and 43 and are decoded, and the DA converters 24 and 44 convert the signals into analog signals. Hybrid circuit 13, 3
3 to the telephones 11, 31. In this case, since the background noise signal is transmitted by the silent cell even in the silent section, the receiving side can reproduce the background noise of the transmitting side. Therefore, the background noise does not change even when the voiced section and the silent section are switched, so that the user does not feel uncomfortable.

FIG. 3 is an explanatory view of a cell. The cell has a fixed length of 53 bytes, which is composed of a cell header of 5 bytes and an information field of 48 bytes.
An adaptive header for inserting byte user information and the like can be provided. The cell header includes a flow control unit GFC (Generic Flow Control) used when a plurality of terminal devices are connected to the user interface, and a logical path identifier VPI (Virtual Path Identifier).
) And a logical channel identifier VCI (Virtual Channel)
l Identifier), and a payload part P indicating network management
T (PayloadType) and reserve part RES (Reserv
e) and cell discard priority CLP (Cell Loss Priority)
) And a header error control unit HEC (Header Error Co)
ntrol) and. The cell header in this case is for the user network interface, and for the inter-network interface, the flow control unit GFC is the logical path identifier VPI.

In the cell assembling units 21 and 41, the logical path identifier VPI and the logical channel identifier VCI are set, and when the detection signal of the silent section from the voice detecting units 19 and 39 is added, the cell discard priority is given. Set a high CLP. Further, when the detection signal of the voiced section is added, the cell discard priority CLP is set low. In addition, the cell assembly parts 21, 4
Even when the modem detection signals are added to 1 from the modem detection units 16 and 36, the cell discard priority CLP is set low. It is also possible to set the cell discard priority of four stages by using a total of 2 bits of the cell discard priority CLP and the reserve part RES. In this case, for example, the cell discard priority of silent cells is set to the highest, the cell discard priority of data cells when a modem detection signal is obtained is set to the lowest, and the cell discard priority of voiced cells is set to the middle. Can be set.

In the ATM type switching network, as described above, when congestion occurs, cells having a high cell discard priority CLP are discarded, so that the terminal devices 12, 3 are discarded.
Data cells in the transmission of facsimile signals and data between the two are set without priority because the cell discard priority is set low, but the data cells are transmitted without being discarded, but in the silent section in the telephone call between the telephones 11 and 31. Since the silent cell of No. 1 is set to have the highest cell discard priority, it is preferentially discarded at the time of congestion. However, since this silent cell contains only the background noise signal, it does not affect the content of the call. In addition, the receiving side can avoid the unnaturalness of the background noise in the silent section by repeatedly reproducing the background noise signal from the previously received silent cell instead of the discarded silent cell.

As described above, in the cell disassembling units 21 and 41 and the decoding units 22 and 42, when the silent cells are discarded, the background noise signal is repeatedly reproduced. The background noise signal for each predetermined section within the silent section can be transmitted as a random silent cell. Therefore, the transmission efficiency can be further improved by increasing the transmission interval of the low rate coded silent cells.

The present invention is not limited to the above-mentioned embodiment, but various additions and modifications can be made. For example, the high rate coding units 17 and 37 and the low rate coding units 18 and 38 can be modified. As a configuration for switching and operating in accordance with the detection signals of the voice detection units 19 and 39.
0 can be omitted.

[0021]

As described above, according to the present invention, the transmitting side 1 includes the voice detecting section 4, the coding section 5 for switching between the high rate coding and the low rate coding, and the cell assembling section 6. Including,
The voice detection unit 4 performs high-rate coding by detecting a voiced section and low-rate coding by detecting a silent section, and the cell assembling unit 6 outputs a high-rate coded output signal of the voiced section to a cell. In this case, the low-rate encoded output signal in the silent section is converted into cells, and the cell discard priority of the cell is set to the highest, and the cells are transmitted. By performing rate coding, the background noise in the silent section can be reproduced on the receiving side with a slight decrease in transmission efficiency, and therefore, there is an advantage that no discomfort occurs in a call.

The cell discard priority of the silent cell is set higher than the cell discard priority of the voiced cell.
When the M-system switching network 2 is congested, silent cells having a high cell discard priority are preferentially discarded, but since the silent cells contain only background noise signals, they do not affect the content of the call. There are advantages.

Further, by transmitting the low-rate coded output signal of the background noise signal within the predetermined interval of the silent section as a silent cell, the interval of sending the silent cell is increased and the transmission efficiency can be improved. In that case, on the receiving side, by repeatedly playing back the background noise signal from a silent cell,
There is an advantage that it is possible to realize a call state that does not cause discomfort.

[Brief description of drawings]

FIG. 1 is a diagram illustrating the principle of the present invention.

FIG. 2 is an explanatory diagram of an example of the present invention.

FIG. 3 is an explanatory diagram of a cell.

FIG. 4 is an explanatory diagram of a conventional example.

[Explanation of symbols]

 1 transmitting side 2 switching network 3 receiving side 4 voice detecting section 5 encoding section 6 cell assembling section

Claims (2)

[Claims] 1. A cell transmission control method for converting a voice signal from a transmission side (1) into cells and transmitting the cells to a reception side (3) through a switching network (2), wherein the transmission side (1) is a voice signal. A voice detection section (4) for detecting a voiced section with a signal and a silent section without a voice signal, and coding for switching between high rate coding and low rate coding according to a detection signal of the voice detection section (4). A coding unit (5) and a cell assembling unit (6) for converting the coded output signal of the coding unit (5) into cells, and the coding unit (5) enhances the voice signal in the voiced section. Rate coding, low-rate coding the background noise signal in the silent section, and the cell assembling unit (6) sets the high-rate coded output signal in the sounding section as a sounding cell, and sets the low-rate coding in the silent section. The output signal is converted into a cell as a silent cell, and the cell discard priority of the silent cell is Set higher than the cell loss priority sounds cells, cell transmission control method, characterized by sending to the receiving side through the switching network (2) (3). 2. The cell assembling unit (6) has a background within a predetermined section of the silent section at an interval longer than an interval at which a high-rate encoded output signal of a voice signal in the sound section is transmitted as a voice cell. The cell transmission control system according to claim 1, wherein the low-rate encoded output signal of the noise signal is transmitted as a silent cell.