JPH01300739A - Equipment for making voice packet multiplexing transmission - Google Patents

Abstract

PURPOSE:To make tentative overload control suitable for the state of voice traffic so as to eliminate deterioration of the speech quality by providing a storing means utilizing rate detecting means and voice operation rate detecting means and using the information received from the detecting means as discriminating conditions for controlling channel assignment. CONSTITUTION:A voice detecting section 42 detects significant voice signals contained in input signals of channels CH1-CHn at every channel and outputs a detected input trunk channel number to an assignment processor 52 and n-channel voice operation rate detecting section 44. The section 44 calculates voice operating rates (alpha) from inputted detecting information at every channel and informs the processor 52 of the calculated results. A buffer utilizing rate detecting section 60 detects an accumulating capacity Q of the memory section 36 accumulating voice information and calculates a buffer utilizing rate (beta) which is informed to the processor 52. The processor 52 does not accumulate the voice information of the trunk channels whose voice operating rates (alpha) exceed a threshold (a) in the memory 36, but annuls. Moreover, when the utilizing rate (beta) exceeds a threshold (b), the processor 52 causes an input regulating section 62 to regulate the originating call of a relevant input truck channel.

Description

DETAILED DESCRIPTION OF THE INVENTION (Field of Industrial Application) The present invention relates to a packet device, and particularly to a voice packet multiplexing transmitting device connected to a high-speed digital line and a dedicated line.

(Prior art) Regarding packetized voice communication, for example, see "Packet voice communication" by Toru Tsuda, Journal of the Institute of Electronics and Communication Engineers.

Vol. 62, No. 2, pp. 182-184 (February 1979)
month). This describes that the audio channel is one channel, and the audio packet transmitting side device samples the input audio original signal at a frequency such as 8KH2 (sampling period: 125 pLs),
8) PC: After encoding into an M signal, this is temporarily stored until a predetermined amount of information is reached. When the accumulated voice information reaches a predetermined amount, for example, destination information and sender information are added as headers, and the voice information is packetized.The voice information is encoded at a speed of, for example, 64Kbit.
/s, the packetized audio information, that is, the audio packet, is transferred to a PC with a speed of 64 Kbit/s or more, for example.
The signal is sent out onto a high-speed transmission path such as a high-speed digital line of the Nth order group. The audio packet is then input to the receiving device, where it is reproduced into an original signal.

Since it is rare for both parties to speak at the same time in a normal circuit-switched telephone conversation, it is said that 60% of conversational voices are non-calls. Therefore, by performing voice packetization only on significant voice, other information can be sent during no calls, and the efficiency of using the transmission path can be improved. Silence compression methods that utilize the statistical properties of conversational speech include, for example, digital voice insertion (Digital
italSpeech Interpolation:
There is a method (hereinafter referred to as DSI). This DS1 system relies on the large grouping effect obtained by bundling the number of lines in a certain location, and if the number of lines on the receiving side is, for example, 60 or more, the logical number of transmission lines on the sending side will be the same. It can be reduced to less than half. Therefore, if the gain obtained by this method, that is, the DSI gain, is, for example, 2.5, 80 channels can be transmitted on a transmission line capable of transmitting 24 channels, and a transmission line with 60 side lines can be secured for reception.

The audio packet multiplexing transmitting device that multiplexes multiple channels to create audio packets time-division multiplexes the signals input from multiple input trunk channels and writes only significant audio signals to the buffer memory using the DS1 method. enter,
When the significant audio signal written in the buffer memory reaches, for example, a predetermined size, a header is added to it, audio packetization is performed, and the audio packet is sent to the audio packet multiplexing receiving side device via a transmission path. The audio packet receiving device decomposes the audio packet into a header and an audio signal, absorbs delay fluctuations, reproduces the audio signal, and sends it to a predetermined output channel. The receiving device selects an output channel based on the information in the header.

(Problem to be Solved by the Invention) However, in such conventional technology, the audio signal of the trunk channel where the audio has been detected is unconditionally written into a buffer memory with a finite capacity. In a temporary voice traffic overload state where the state increases, the delay time increases temporarily due to the increase in the packet queue in the buffer memory, and the number of packets discarded due to packet overflow in the finite length buffer memory increases temporarily. However, there was a problem in that the voice communication quality deteriorated significantly.

The present invention eliminates the drawbacks of the prior art, and provides voice packet transmission that prevents the quality of calls from deteriorating and prevents voice conversations from becoming difficult to communicate even in a temporally overloaded state of voice traffic. The purpose is to provide side equipment.

(Means for Solving the Problems) In order to solve the above-mentioned problems, the present invention identifies significant signals of signals input from a plurality of input trunk channels, adds a header to the significant signals, and sends out audio as a packet. The packet multiplexing transmitting device includes significant signal detection means for identifying significant signals from input signals, storage means for sequentially accumulating the significant signals detected by the significant signal detection means, and significant signals accumulated in the storage means. storage means usage rate detection means for calculating a storage means usage rate indicating the ratio between the storage capacity of and the maximum storage capacity that can be stored in the storage means at a first fixed time interval; and a first threshold value is set; and processing means for restricting calling of the input trunk channel when the storage means usage rate is equal to or higher than the first threshold value.

(Function) According to the present invention, when the storage means utilization rate is equal to or higher than the first threshold value, the number of input trunk channels is temporarily reduced by performing call control, and the voice operation rate is reduced to the second threshold value. If it is greater than the threshold, the significant signal is discarded without being written into the storage means.

(Embodiment) Next, an embodiment of the voice packet multiplexing transmitter apparatus according to the present invention will be described in detail with reference to the accompanying drawings.

FIG. 4 shows an example of a system configuration of a voice packet multiplexing transmitter device according to the present invention. In the same configuration example, the voice packet multiplexing transmitting side device 30 and the voice packet multiplexing receiving side device 32 are connected to the digital electronic exchange 20/22.
The transmission line 1 is connected to the trunk side of the
02. The electronic exchange 20/22 has telephones 10. Information equipment 12 such as a personal office computer or word processor, and a facsimile 15 are accommodated as terminal devices via a modem 14.

The electronic exchange Ja, 20 encodes the voice signals sent from these terminal devices into, for example, a PCM 8-bit digital signal using a codec, and sends it to the voice packet multiplexing transmitting side device 30 via the input trunk channel 100. The transmitting device 30 multiplexes the digital signal input from the channel 100, performs silent compression using the DSI method, adds a header, and transmits it as an audio packet to the transmission path 102.
The voice packet is multiplexed and sent to the receiving side device 32 via the voice packet. The receiving device 32 decompresses the received voice packet into a PGM 8bi/) digital signal and sends it to the electronic exchange 22 via the output trunk channel 104 indicated by the header of the packet. The electronic exchange 22 converts the digital signal received via the trunk channel 104 into an analog signal using a codec, and sends the analog signal to the terminal device.

Note that in the figure, in order to facilitate understanding, the transmitting device 30 is connected to the exchange 20, and the receiving device 32 is connected to the exchange 22, but normally these exchanges are connected to the transmitting device 30 and the receiving device 32. are connected to each other for bidirectional communication. Although the figure shows an example in which the voice packet multiplexer is connected to an electronic exchange, it is also possible to directly connect the terminal equipment by providing a PCM encoding circuit and a PGM decoding circuit in the voice packet multiplexer. is also possible.

FIG. 1 shows a functional block diagram of the audio packet multiplexing transmitting device 30 in this embodiment. The input side of multiplexer 32 is connected to an input trunk channel 100 consisting of channels CHI-CHn.

The multiplexer 32 is a multiplexing circuit that time-division multiplexes PGM 8-bit digital signals input from each of the channels CHI to CHn, and outputs the resultant signal to an output 200 as a serial input quadratic signal.

An n-channel sound detecting section 42 is connected to the output 200. The sound detecting unit 42 is a detecting unit that detects significant audio signals included in channels CHI-CHn inputted as input idle signals for each channel. The presence detection unit 42 enables silence compression using the DST method,
Audio packetization is performed only on significant audio. When the voice presence detection unit 42 detects a significant voice, it outputs the detected input trunk channel number to the allocation processor 52 via the output 208, and also outputs the detected input trunk channel number to the allocation processor 52 via the output 208.
It is output to the channel audio activity rate detection section 44.

The voice activity rate detection unit 44 is a detection unit that checks the voice activity rate for each input trunk channel. The voice activity rate detection unit 44 calculates the voice activity rate α for each channel from the detection information input through the input 209, for example, every fixed time TA. Then, the calculation result is notified to the allocation processor 52 via the output 210. This fixed time TA is set to a relatively short time.

The buffer memory unit 36 is a storage unit that stores voice information that is more significant than the input 200. Since the inside of the memory unit 36 is modeled as a single queue, the order of the input audio information does not reverse, and the temporal order is guaranteed. Therefore, signal information that is packetized according to the temporal order is output to the packet assembling section 40.

The buffer usage rate detection unit 60 is a usage rate detection unit that calculates the buffer usage rate β stored in the memory unit 38 at every fixed time TO. That is, the detection unit 60 detects
The storage capacity Q in which the audio information is stored in the memory unit 36 for each B is detected, and the ratio of the storage capacity Q to the maximum storage capacity Qmax of the memory unit 3B, that is, the buffer usage rate β is calculated. After calculating the buffer usage rate β, the buffer usage rate detection unit 60 notifies the allocation processor 52 of the calculation. This fixed time TB is also set to a relatively short time.

The packet assembling unit 40 receives signal information from the memory unit 36,
It is also a packet assembling unit that inputs the headers of this signal information from the allocation header encoding unit 56, combines them, and outputs them as audio packets. Packet assembly section 40
From there, voice packets of a plurality of channels are packet-multiplexed and outputted onto the transmission line 102 as an output bearer signal.

The allocation processor 52 is a control device that performs channel allocation control on the significant signals of the multiplexed digital audio information 200 and converts them into audio packets. That is, the allocation processor 52 has a built-in allocation status table, and when a trunk channel number for which voice activity has been detected is input, the allocation processor 52 displays the correspondence relationship between the outgoing trunk channel number, the incoming trunk channel number, and the logical channel number, and the presence of voice activity. Register the presence or absence of detection.

When the voice activity rate α is input from the voice activity rate detection unit 44 for each trunk channel number, the processor 52 compares it with a semi-fixed threshold a and assigns a channel to the significant voice of that trunk channel number. and decide whether to write to the buffer memory 3B. That is, when the voice operation rate α is smaller than the threshold value a, the processor 5
2 notifies the allocation state control unit 54 of the input trunk channel number, and does not notify the allocation state control unit 54 of the input trunk channel number when the voice activity rate α is equal to or higher than the threshold value 3. 6 The allocation state control unit 54 controls the input trunk channel number. When the number is not received, the memory control unit 50 is not instructed to control the accumulation of voice information. Therefore, the voice operation rate α is set to the threshold value 3.
The audio information of the above trunk channel numbers is discarded without being stored in the buffer memory 3B.

Furthermore, the processor 52 has a semi-fixed threshold value for the buffer usage rate β, and when the usage rate β is higher than the threshold value, it is determined that the processor 52 is in a high load or congestion state.
Input restriction control unit 62 through output 310 for calling restriction instructions
to notify. The input restriction section 82 is connected to the multiplexer 32 via an output 312, and upon input of a call restriction instruction, performs call restriction on the corresponding input trunk channel.

Note that the transmitting device 30 employs the DS1S2O method and the discard packet method that allows partial omission of voice information until substantial voice communication quality can be ensured. When the DSI method is used as described above, the gain is, for example, about 2.5 times, so the amount of information can be compressed to less than half. Regarding the method of discarding packets, it is known that if the discard rate is about 1%, it is acceptable in terms of voice communication quality. For this reason, the threshold a
The settings of the threshold value and the threshold value are determined based on the trade-off between the two, taking into consideration channel usage efficiency and speech quality. That is, the threshold value a and the threshold value S are set so as to ensure substantial voice communication quality by setting the channel usage efficiency to 2.5 times and the packet discard rate to 1%.

FIG. 2 conceptually illustrates the operating principle of channel allocation control of the transmitting side device 30.

Therefore, the memory control unit 50 and the allocation state control unit 54
Although the write control unit 50154 shown in FIG. is in control.

Furthermore, although the packet assembling section 40 and the like are omitted in the figure, and the transmission path 102 is shown as being connected to the buffer memory section 36, it is actually connected to the transmission path 102 via the packet assembling section 40. ing.

FIG. 3 shows a channel allocation control flow of the transmitting side device 30. The operation of channel allocation control will be explained using FIG. 1, FIG. 1, and FIG. 2. The multiplexer 32 has an input trunk channel 1 consisting of CHI to CHn.
When a PCM 8-bit audio signal is input from 00, it is time-division multiplexed into a serial audio signal and output to output 200. When the n-channel sound detection section 42 detects a significant audio signal from the input 200 (500), the detection section 42
notifies the allocation processor 52 of the detected trunk channel number via output 208.

Even if the allocation processor 52 receives a notification that voice activity has been detected, if the buffer usage rate β calculated every TB for a certain period of time is greater than or equal to a predetermined threshold value, the allocation processor 52 determines that there is a high load or congestion state ( Ei02). For this reason, the allocation processor 52
does not register the channel allocation in the allocation status table, but outputs a call restriction instruction to the input restriction control section 82. Input regulation section 62
When inputting a call restriction instruction, the call control is performed for the channel number for which voice activity has been detected (E120).Also, when the buffer usage rate β reaches a predetermined value V! If the value is smaller than 4, the allocation processor 52 determines that it is in the normal state (8
02) The input trunk channel number for which voice activity has been detected is registered in the allocation status table (804).

Furthermore, the voice activity rate α of the channel number registered in the allocation status table is input to the allocation processor 52 at fixed time intervals TA. When the processor 52 inputs the voice operation rate α,
It is determined whether the voice activity rate α exceeds a predetermined threshold value a. If the threshold value is 8 or more, it is assumed that there is an overload state, and the channel number is not notified to the write control unit 50154. Therefore, the audio information of this channel is not written into the buffer memory 36 and is discarded by the write control unit 50154 (810). In FIG. 2, discarded discarded packets are indicated by black circles.

Further, when the voice activity rate α is less than the threshold value a, the allocation processor 52 determines that the state is normal and notifies the write control unit 50154 of the channel number. As a result, the write control U4 unit 50154 writes the audio information to the buffer memory 36 (eoa). When the write control to the buffer memory 38 is performed, the new allocation operation is completed. Therefore, when the allocation processor 52 performs write control, it instructs new allocation and ends the operation (e42).
Since this series of channel allocation control is activated in units of audio block information, for example, 2 ms of 1B sample, it is possible to immediately respond to a temporal overload state of audio traffic. Note that the white circles written in the buffer memory 38 indicate newly allocated information as write packets.

The blocks of audio information written in the buffer memory 36 are sequentially stored as write packets in the buffer memory 36 on a first-come, first-served basis, and are sent to the transmission line 10 in predetermined packet units.
This buffer memory 36 moving to the second side is modeled as a single queue, as shown in FIG. does not occur and temporal order is guaranteed. Therefore, if the storage capacity Q stored in the buffer memory 36 does not exceed the transmission capacity C of the transmission line 102 on the output side,
The packets are sequentially sent to the transmission line 102.

Letting the maximum storage capacity of the buffer memory 36 be Qmax, when operating at the maximum buffer processing speed, Qmax=
It becomes C. Therefore, the buffer usage rate β, which is the ratio of the storage capacity Q to the maximum storage capacity Qmax, reflects the processing efficiency of the buffer memory 36. Therefore, it is calculated every fixed time TB, and is notified to the allocation processor as call restriction information as described above.

As described above, in this embodiment, feedback control is performed using the voice activity rate α, the buffer usage rate β, and the threshold values a and b, which define the channel allocation processing operation. That is, by comparing the buffer usage rate β of the buffer memory 36 installed on the output side of the transmitting side device 30 with the threshold value, it is fed back and becomes call restriction information on the input side, and the voice operation rate α is compared with the threshold value. The amount of input to the buffer memory 3B is limited by comparison with a. Therefore, in this embodiment, even in a temporal overload state of voice traffic due to a sudden increase in the number of input ranks, there is a temporary increase in delay time due to an increase in the packet queue in the buffer memory 36, and a temporary increase in the delay time due to the increase in the packet queue in the buffer memory 36. This makes it possible to prevent an increase in packet discards.

Note that this embodiment is applicable to data in the voice band modulated by a modem and the like, and is also applicable to, for example, a data packet multiplexing transmitter.

(Effects of the Invention) As described above, according to the present invention, the storage means usage rate detection means and the voice activity rate detection means are provided, and the information received from each detection means is used as a judgment condition for channel allocation control. Therefore, temporary overload control suitable for the voice traffic situation is possible, and the buffer memory can be used effectively and efficiently. Specifically, the voice activity rate is calculated every second time period, which is relatively short, and writing to the buffer memory can be controlled based on the second threshold value. Therefore, in an overload state due to a temporary increase in voice traffic, discarded packets can be tolerated to ensure substantial voice communication quality, and the efficiency of voice channel usage can be maintained. Furthermore, the storage means usage rate is calculated at each first time period, which is relatively short, and the number of input trunk channels can be controlled based on the first dark value. Therefore, it is possible to prevent the entire transmitting device from becoming inoperable and making it difficult to make a call due to a high load or congestion caused by a sudden increase in voice traffic.

[Brief explanation of the drawing]

FIG. 1 is a functional block diagram showing an embodiment of the voice packet multiplexing transmitting device according to the present invention, FIG. 2 is an operation principle diagram showing the operating principle of channel allocation in the embodiment shown in FIG. 1, and FIG. FIG. 3 is a flow diagram showing an example of a channel allocation control flow in the embodiment shown in FIGS. 1 and 2, and FIG. FIG. 2 is a system configuration diagram shown in FIG. Explanation of the symbols of the main parts 30, . . . Voice packet multiplexing transmitting device 3211. Multiplexer 3B, . . . Buffer memory 40, . . . Packet assembling section 42, . . Allocation processor 54, . . . Allocation status control unit 5B, . . , Allocation header encoding unit 6010. Buffer usage rate detection unit 62, input regulation control unit

Claims (1)

[Scope of Claims] 1. In a voice packet multiplexing transmitting device that identifies a significant signal of a signal input from a plurality of input trunk channels, adds a header to the significant signal, and transmits it as a packet, the transmitting device includes: significant signal detection means for identifying the significant signal from the input signal; storage means for sequentially accumulating the significant signal detected by the significant signal detection means; and the significant signal stored in the storage means. storage means usage rate detection means for calculating at a first fixed time interval a storage means usage rate indicating a ratio between the storage capacity of the storage means and the maximum storage capacity that can be stored in the storage means; and a first threshold value is set. . A voice packet multiplexing transmitting side apparatus, comprising: processing means for restricting calling of the input trunk channel when the storage means usage rate is equal to or higher than a first threshold value. 2. The transmitting side device according to claim 1, wherein the transmitting side device determines the proportion of the significant signal detected by the voice detection means among the input signals by performing voice operation at a second fixed time interval. A second threshold is set in the processing means, and when the voice action rate is equal to or higher than the second threshold, a significant signal equal to or higher than the second threshold is detected. A voice packet multiplexing transmitter device characterized in that it discards voice packets.