JP2005252429A - Ip packetizing unit - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an IP packetizer capable of preventing degradation in voice quality, even if traffic is increased in an IP network. <P>SOLUTION: The IP packetizer comprising a voice compressing block 11 for compressing and IP for packetizing voice data being transmitted to an IP network, a central processor 7 for performing general control and grasping traffic information by communicating with a route selector, e.g. a router, in the IP network using SNMP is further provided with a voice detector 10 for determining the compression ratio in the voice compression block, based on the traffic information grasped. Degradation in voice quality can be prevented or traffic can be reduced on the receiver side by selecting low compression, high compression or silent compression, depending on the traffic state. <P>COPYRIGHT: (C)2005,JPO&NCIPI

Description

  The present invention relates to an IP packetizing apparatus that is used in an IP telephone or an IP-PBX (IP private branch exchange) and compresses voice data into an IP packet and transmits it to an IP network.

  FIG. 7 is a block diagram showing an IP telephone system using a conventional IP packetizing apparatus that uses VoIP technology or the like to convert voice into IP packets and performs voice communication via an IP network.

  In FIG. 7, 1a and 1b have IP packetizing functions, IP telephones having LAN interfaces, and 2a and 2b IP gateway functions (voice compression / decompression, IP addresses) in addition to the conventional PBX (private branch exchange) function. IP-PBX, 3a, 3b equipped with management, LAN interface, etc. are suitable for personal computers (PC), 4a, 4b, 5a-5d with IP packetized voice data (IP packetized voice data) to the destination A router (route selection device) 5 for selecting and transmitting a route is an IP network.

  The operation of the IP telephone system having the conventional IP packetizing apparatus configured as described above will be described with reference to FIG. FIG. 8 is a block diagram showing a network monitoring agent 300 connected to the packet network.

  In FIG. 8, 300 is a network monitoring agent that selects a compression criterion based on the current amount of network traffic, 310 is a processor that extracts, interprets, and executes instructions stored in the data storage device 320, and 330 is a network monitoring agent. A communication port that connects 300 to a packet telephone system (ie, a packet network), thereby linking the network monitoring agent 300 to each connection node. The data storage device 320 includes a connection database 500, a termination database 550, and a dynamic compression adjustment process 600.

  In FIG. 7, the voice data from the IP-PBXs 2a and 2b are integrated with general-purpose data from the PCs 3a and 3b via the routers 4a and 4b, and transmitted through the packet network. This packet network changes dynamically during a call, and packet data does not always pass the same route. Various IP-PBXs and IP telephones are connected to the packet network, and the number of telephone lines fluctuates every moment. Further, there is data from the personal computers 3a and 3b, and the traffic amount (data transfer amount, Packet loss) increases, decreases, or changes.

Under such circumstances, voice communication is required to have real-time properties, and it is necessary to prevent deterioration in sound quality at a telephone terminal (end) in end-to-end transmission. In particular, serious voice quality degradation is degradation due to end-to-end transmission delay time and degradation due to packet loss occurring in the packet network. The loss of an IP packet that does not reach its destination is called packet loss, but there are two main causes of packet loss. The first is based on the routing specification (route selection specification) of the packet network. For example, when there are 15 relay nodes such as routers through which one IP packet has passed, this packet is discarded. Depending on the traffic state, this packet discard phenomenon tends to occur. The second is packet loss related to transmission time. In the packet network, jitter of delay time that varies in transmission delay time occurs. This is mainly caused by the difference in performance of each router (data difference in packet reception time, internal processing time, transmission waiting time, etc.) in data transmitted by being divided into a large number of packets. On the receiving side, the delay time is adjusted so that the time interval between a series of received IP packets is constant. However, IP packets that exceed the adjustable range due to the excessive delay time are discarded. As described above, the voice quality greatly depends on the traffic state in the dynamic packet network which changes every moment.

Here, in (Patent Document 1), for example, in a packet telephone system, network conditions such as traffic volume are monitored with the configuration shown in FIG. 8, and the timing for dynamically adjusting the encoding scheme for the connection is determined. A network monitoring agent 300 is disclosed. The network monitoring agent 300 can select an encoding standard based on, for example, the current network traffic volume, network error characteristics, date and time, or day of the week. In U.S. Patent No. 6,047,089, a coding standard that provides a lower degree of compression and a higher quality level is selected when the network traffic is lighter. The network monitoring agent 300 notifies the devices associated with each connection of changes in the encoding scheme.
JP 2001-57573 A

  However, in the conventional IP packetizing apparatus as described above that communicates voice into IP packets using VoIP technology or the like, if the traffic volume in the IP network increases, packet loss or the like is likely to occur, and sound is cut off. It has the problem of incurring quality degradation.

  This IP packetizing apparatus is required not to deteriorate voice quality even if the traffic amount in the IP network increases.

  In order to satisfy this requirement, an object of the present invention is to provide an IP packetizing apparatus that can prevent voice quality degradation even when the traffic volume in an IP network increases.

  In order to solve this problem, an IP packetizing apparatus according to the present invention includes a voice compression block that compresses voice data, converts the voice data into an IP packet, and transmits the IP packet to the IP network. An IP packetizing apparatus having a central processing unit that communicates with a path selection apparatus and grasps traffic information, and includes a voice detector that determines a compression rate in a voice compression block based on the grasped traffic information have.

  As a result, an IP packetizing apparatus capable of preventing voice quality deterioration even when the traffic amount in the IP network increases can be obtained.

The IP packetizing apparatus of the present invention compresses voice data, converts it into an IP packet and transmits it to an IP network, and controls the whole and communicates with a route selection apparatus such as a router in the IP network using SNMP. An IP packetizing apparatus having a central processing unit for grasping traffic information, and having a voice detector for determining a compression rate in a voice compression block based on the grasped traffic information, for example, there is voice Perform low compression (compression with a low compression ratio) for voiced sections with a low degree of deterioration due to compression, and silence (uncompressed with a high compression ratio) or silence for silent sections that do not matter even if deteriorated, such as background noise Since compression (compression with higher compression ratio) can be performed, depending on the traffic conditions, low compression, that is, superior voice quality High compression or silence compression, that is, priority can be given to traffic, and low compression or high compression / silence compression can be selected according to the traffic conditions to prevent voice quality degradation on the receiver side or to reduce traffic volume. The advantageous effect that it can be achieved is obtained.

  Furthermore, the voice compression block has the previous voice data and the current voice data in one packet, and the voice detector performs low compression on the previous voice data and the current voice data, thereby causing packet loss. In addition, an advantageous effect is obtained that packets lost on the receiving side can be reproduced.

  Further, the voice compression block has the previous voice data and the current voice data in one packet, and the voice detector performs low compression on either the previous voice data or the current voice data, and the previous voice data. By performing high compression on the other audio data of the data and current audio data, lost packets can be reproduced on the receiving side even if packet loss occurs, and both low compression and high compression By using, the amount of traffic can be suppressed, and the advantageous effect is obtained that the lost packet can be reliably reproduced on the receiving side and the voice quality can be maintained.

  Furthermore, the voice compression block has the previous voice data and the current voice data in one packet, and the voice detector performs high compression on the previous voice data and the current voice data, thereby causing packet loss. In addition, it is possible to reproduce lost packets on the receiving side, and to suppress the traffic volume to the maximum by using high compression, and to ensure that voice quality is maintained by reliably reproducing lost packets on the receiving side. The advantageous effect of being able to be obtained is obtained.

  The object of the present invention is to prevent the voice quality from deteriorating even if the traffic volume in the IP network increases, by determining the compression rate in the voice compression block based on the grasped traffic information.

  A first invention made to solve the above problems is a voice compression block that compresses voice data, converts it into an IP packet and transmits it to the IP network, and controls the whole and uses a router in the IP network using SNMP. An IP packetizing device having a central processing unit that communicates with a route selection device and grasps traffic information, and includes a voice detector that determines a compression rate in a voice compression block based on the grasped traffic information For example, low-compression (compression with a low compression ratio) is performed for voiced sections with speech, and low-compression is not compressed. (Compression with high compression ratio) or silence compression (compression with higher compression ratio) can be executed, so depending on traffic conditions Low compression, i.e., voice quality priority, or high compression or silence compression, i.e., traffic priority, can be selected according to traffic conditions to prevent voice quality degradation on the receiver side. Or it has the effect | action and effect that reduction of the amount of traffic can be aimed at.

  According to a second aspect of the present invention for solving the above-described problem, the voice compression block has the previous voice data and the current voice data in one packet, and the voice detector performs the previous voice data and the current voice data. The low-compression is performed, and even if packet loss occurs, the lost packet can be reproduced on the receiving side.

According to a third aspect of the present invention for solving the above-described problems, the audio compression block has the previous audio data and the current audio data in one packet, and the audio detector is either the previous audio data or the current audio data. In this case, low compression is applied to the voice data of the previous time and high compression is applied to the other voice data of the previous voice data and current voice data. Packets can be played back, and the amount of traffic can be reduced by using both low compression and high compression, and lost packets can be played back reliably on the receiving side to maintain voice quality. Has the effect of being able to.

  According to a fourth aspect of the present invention for solving the above-described problem, the voice compression block has the previous voice data and the current voice data in one packet, and the voice detector is used for the previous voice data and the current voice data. It is intended to perform high compression, and even if packet loss occurs, lost packets can be replayed on the receiving side and traffic volume can be suppressed to the maximum by using high compression. Thus, the received packet can be reliably reproduced on the receiving side to maintain the voice quality.

(Embodiment 1)
FIG. 1 is a block diagram showing an IP packetizing apparatus according to Embodiment 1 of the present invention.

  In FIG. 1, 6 is an A / D converter that digitizes a voice analog signal that is a voice of a speaker and outputs voice digital data, 7 is a CPU that controls the control, and 8 is a FIFO that stores and outputs digital data for a certain period of time. Memory, 9 is an electronic volume unit for changing the threshold level of sound and silence, and 10 is audio digital data (also simply referred to as “voice data”) which is data indicating the sound level, and threshold levels of sound and silence A voice detector that detects the presence or absence of voice from two inputs, 11 is a voice compression block that performs various compression on voice data, 12 is a switch that switches between voiced and silent compression (that is, switches the compression rate), Reference numeral 13 denotes a low compression unit used during sound compression, 14 denotes a high compression unit used during background noise compression, and 15 denotes a silence compression unit used during silence compression.

  The operation of the IP packetizing apparatus configured as described above will be described with reference to FIGS. FIG. 2A is a timing chart showing data indicating voice level (voice level data), FIG. 2B is a timing chart showing IP packets, and FIG. 3 is a protocol installed in an IP telephone and a router. It is a protocol block diagram.

  Here, it is assumed that necessary call control is completed via the IP-PBX (switches) 2a and 2b (see FIG. 7) and a call is possible. First, the audio analog signal is digitized by the A / D converter 6 and temporarily stored in the FIFO memory 8. The CPU 7 grasps the background noise level around the IP telephones 1 a and 1 b (see FIG. 7), and controls the electronic volume unit 9 to set a certain threshold value in the voice detector 10. As a result, detection of sound and silence is output from the sound detector 10. In the state where the voice of the speaker before the call is not input, there is no background noise alone, so the changeover switch 12 is switched to the silence compression unit 15 or the high compression unit 14, so long as it is on the silence compression unit 15 side. IP packetization is not performed. That is, no audio data is transmitted to the transmission line in the silent period. The CPU 7 constantly monitors the voice data level and also sets a threshold value, so that the output state of the voice detector 10 is grasped.

  The quality of voice depends on the state of traffic (data transfer amount, packet loss amount) in a dynamic packet network that changes every moment as described in FIG. 7 (background art). . If the traffic state can be detected, the low compression unit 13, the high compression unit 14, and the silence compression unit 15 can be appropriately varied according to the traffic state.

Next, an operation of switching the low compression unit 13, the high compression unit 14, and the silence compression unit 15 according to the traffic state will be described.

  Usually, the IP telephones 1a and 1b are equipped with RTP (Real time Transport Protocol) and RTCP (Real time Control Protocol). RTP can detect packet loss, and RTCP is effective for adjusting the transmission rate of the sender (IP telephone). Further, a management protocol used in an open network environment called SNMP (Simple Net-Work Management Protocol) is intended to collect information in the IP network. The router has the basic functions of SNMP as well as RTP and RTCP, but the IP phone has RTP and RTCP, but SNMP is not currently installed.

  SNMP will be described briefly. A managing side (IP telephone) and a communicating party (management target) are required. In FIG. 7, the management target is all routers 4a, 4b, 5a to 5d in the packet network. As shown in the protocol block diagram of FIG. 3, between the manager (IP telephone) and the agent (router), the manager inquires of the agent for information, and the agent responds to the manager. A collection of information possessed by the agent is referred to as MIB (Management Information Base), and each piece of information defined in the MIB is referred to as an object. When it is desired to confirm the state of an interface (port) of a certain router, the IP telephone makes a request for the MIB information of the router, and the value (return value) returned is judged. Specific traffic information (objects) includes the usage rate of connection lines (input and output data volume and line speed) for each interface of a router, and the load factor of the router (rate of packets discarded because the router could not process them) )and so on. For example, when the load factor of the router increases, there is a possibility that the line capacity of the transfer destination interface (port) is too small for the transfer request traffic or the memory capacity in the router is small. If there are many packets discarded, there is a possibility that retransmission of data may be complicated. By installing SNMP in the IP phone, it is possible to grasp the MIB (traffic information) that the router has, and the IP phone comprehensively determines the traffic status in the IP network. The section performs low compression with a low degree of quality degradation, and the silent section performs high compression. If the section is bad, the speech section performs high compression and the silent section performs silence compression. Low compression in a voiced section gives priority to voice quality, and silent compression in a silent part gives priority to traffic.

  FIG. 2 shows that the IP packet is in a low compression state, a high compression state, and a silence compression state with or without voice. Dotted squares are silence compression, packetization is not performed in this silence period, and data is not transmitted. As described above, using the IP packetizing apparatus of FIG. 1, the IP telephone can grasp traffic information by communicating with the router, and based on this result, a method for compressing packets with and without voice can be selected. Thus, voice quality can be maintained without increasing the traffic volume of the IP network.

  As described above, according to the present embodiment, the voice detector 10 that determines the compression rate in the voice compression block 11 based on the grasped traffic information is provided. Low compression with a low degree of degradation (compression with a low compression rate) is executed by the low compression unit 13, and high compression or silence compression is performed on the high compression unit 14 or silence for silent sections that are not noticeable even if they deteriorate, such as background noise. Since it can be executed by the compression unit 15, it is possible to prioritize low compression, that is, voice quality, or prioritize high compression or silence compression, that is, traffic priority, depending on traffic conditions. It is possible to select according to the state of the receiver and prevent the voice quality deterioration on the receiver side or reduce the traffic volume.

(Embodiment 2)
FIG. 4 shows an operation in the IP packetizing apparatus according to the second embodiment of the present invention. FIG. 4 (a) is a timing chart showing an analog audio signal, and FIG. 4 (b) is a division time of 10 ms (millisecond). 4 (c) is a timing diagram showing IP packetization, FIG. 4 (d) is a packet data diagram showing an IP packet on the transmission path, and FIG. 4 (e) is a packet loss. FIG. 4F is a packet data diagram showing an IP packet when a packet loss occurs. FIG. FIG. 4 shows that the upper transmission side IP telephone (or IP-PBX) captures the voice analog signal, compresses it and transmits it after packetization, and the IP packet from the transmission side passes through the transmission path in the IP network. This shows the time until an IP packet arrives at the receiving side IP telephone (or IP-PBX). The configuration of the IP packetizing apparatus according to this embodiment is the same as that shown in FIG.

  The operation of the IP packetizing apparatus configured as described above will be described.

  After the audio analog signal shown in FIG. 4A is digitized, it is divided at certain intervals (10 ms in FIG. 4B) and sequentially performs low compression with small quality degradation (data of divided times а0 to an). ). This is converted into an IP packet. As shown in FIG. 4C, previous data (previous data) and current data (current data) are put in the same packet in one packet. In FIG. 4C, IP packet data A0 and IP packet data A1 are contained in the same packet. After that, it goes in the same way and continues to the packets of An-1 and An. Conventionally, only An is included in one packet, and An-1 is not included. That is, data that has already been transmitted is not sent twice. This point is different from the prior art, and this embodiment is characterized by the second embodiment. The packet in the figure has a data size of only application data (voice data), and a necessary IP header and the like are omitted. When the traffic state of the IP network is good and there is no large delay or packet loss, packets that arrive at the receiving side are sequentially decomposed and decompressed to reproduce the original audio signal (FIG. 4). (See (e)), the same data An is always included in the previous packet or the next packet. Therefore, all the signals are alternately expanded without being expanded so as not to be redundantly expanded, and the original audio signal is reproduced. Discard unexpanded data. Next, consider a case where packet loss occurs in the IP network. For example, the first packet No. 0 (packets A0 and A1) to the 49th packet (A50 and A49 packets) have been reproduced in the above-described process, but the 50th packet (A51 and A50 packets) 1 packet has not arrived. However, the 51st packet (packets A52 and A51) has arrived. In such a case, the lost A51 and A50 data are present in the 49th and 51st packets, and can be reproduced. As described above, all packet decomposition is performed and data discarding is performed at the expansion stage. Therefore, the 49th (A50 and A49 packet) packet information is not lost, and packet loss is dispersed. If so, it will be possible to complement and regenerate. Further, consider a case where two consecutive packets are lost. The 50th (A51 and A50 packets) and 51st (A52 and A51 packets) are assumed to be packet loss. Although there are no three types of A50, A51, and A52, since the A52 exists in the 49th (A50 and A49 packets) packet and the A52 exists in the 49th (A53 and A52 packets), these data are reproduced. Is possible. Only A51 cannot be reproduced. In this case, since there is no possibility of audio reproduction due to lost packets as in the prior art, there is no choice but to make a retransmission request from the receiving side or execute pseudo-complementation with previous and subsequent packets. Considering this, when two consecutive packets are lost, conventionally, for example, two data of A50 and A51 (or A51 and A52) are used, but in this embodiment, one data of A51 is used. Yes, the degree of quality degradation when pseudo-complementation is applied is also small compared to the prior art. Furthermore, SNMP can be installed as in the first embodiment, and for example, the packet transmission method can be changed to the conventional method or the method of the present embodiment according to the traffic state.

As described above, in this embodiment, by transmitting two pieces of previous data and current data in one packet, even if a packet loss occurs, it can be reproduced on the receiving side, and the deterioration of call quality can be reduced. Yes.

  As described above, according to the present embodiment, the voice compression block 11 has the previous voice data and the current voice data in one packet, and the voice detector 10 is lower than the previous voice data and the current voice data. By performing compression, it is possible to reproduce a lost packet on the receiving side even if packet loss occurs.

(Embodiment 3)
FIG. 5 shows the operation of the IP packetizing apparatus according to Embodiment 3 of the present invention. FIG. 5 (a) is a timing chart showing an audio analog signal, and FIG. 5 (b) is a division time of 10 ms (milliseconds). Second) and 15 ms, FIG. 5 (c) is a timing diagram showing IP packetization, FIG. 5 (d) is a packet data diagram showing IP packets on the transmission path, and FIG. FIG. 5F is a packet data diagram showing an IP packet when a packet loss occurs. FIG. 5F is a packet data diagram showing an IP packet when there is no packet loss. FIG. 5 shows that the upper transmission side IP telephone (or IP-PBX) captures the voice analog signal, compresses it, and transmits it after packetization, and the IP packet from the transmission side passes through the transmission path in the IP network. This shows the time until an IP packet arrives at the receiving side IP telephone (or IP-PBX). The configuration of the IP packetizing apparatus according to this embodiment is the same as that shown in FIG.

  The operation of the IP packetizing apparatus configured as described above will be described with reference to FIG. FIG. 6 is a data configuration diagram showing a schematic configuration of an RTP packet.

  In the second embodiment, fixed compression using low compression is used. In the third embodiment, ann is low compression and bn is high compression. This point is different from the second embodiment. In the second embodiment, since the same data is transmitted in the same low compression with two packets, the total number of packets increases compared to the conventional case, which also increases the traffic volume of the IP network. In order to reduce this even a little, one in the same packet is highly compressed. A high compression ratio means that the original data size can be reduced to a smaller compressed data size. In other words, if the high compression bn data size is the same as the low compression ann data size, the original data in the high compression period can be captured larger, and the total number of packets to be transmitted is consequently reduced. It will be less than 2. In FIG. 5B, for example, this is 15 ms for the high compression period and 10 ms for the low compression period. The important thing here is that there are two types of data with different compression rates in the same packet, so this information is sent on the transmitting side, and the data must be decompressed based on that information on the receiving side. It is not to be. Conventionally, when starting communication, negotiation (negotiation) for matching compression / decompression encoding / decoding schemes is executed to determine the encoding / decoding schemes. In this embodiment, the process is performed as follows.

Matches and recognizes that there are two types of data with different compression rates in the same packet at the time of negotiation. In this matching and recognition, RTP and RTCP are usually used in combination as application protocols for transmitting IP packets, but information in the RTP packet header is used. There is information called PT (payload type) in the header, but the payload indicates actual data and means audio data. That is, the encoding method of actual data (voice data) excluding the header is clearly indicated. Currently, for example, G.M. 728, G.G. 723.1, G.M. 729 and the like, but besides these standardized encoding methods, there are unassigned setting values assuming an encoding method that will be put into practical use. Information (bits) should be specified. The RTP header size is composed of a fixed length portion and an extension portion, and the RTP header size can be found by checking header information indicating the presence or absence of this extension portion. Thereby, the size of the payload can be known by subtracting the RTP header from the RTP header and the overall size of the payload. Here, two different types of compressed data sizes in the payload must be known, but since the start and end are known, it is sufficient to indicate the boundary. This inevitably reveals two different sizes. This information uses CC (Contributing Source Count). This is a contribution source count (4 bits), and is an identifier when a plurality of (individual) data sources are placed on the payload of one RTP packet when a multipoint conference call is performed. Use this. The payload size is approximately 1500 bytes at maximum, and this area can be divided into 16 at maximum by 4 bits. Two kinds of sizes can be understood from this set value. The above is one example executed using the information in the RTP packet header in the third embodiment. As other header information, option (reserve) information can be used. FIG. 6 shows a schematic configuration of the RTP packet.

  In the present embodiment, as in the first embodiment, SNMP is installed, and, for example, the packet transmission method is changed to the conventional method, the method of the second embodiment, or the like according to the present embodiment. It can also be a method.

  In this way, in this embodiment, voice data with different compression ratios is sent and received in the same packet, so that it is possible to reduce deterioration of voice quality while reducing traffic volume.

  As described above, according to the present embodiment, the voice compression block 11 has the previous voice data and the current voice data in one packet, and the voice detector 10 has one of the previous voice data and the current voice data. By performing low compression on audio data and performing high compression on the other audio data of the previous audio data and current audio data, even if a packet loss occurs, the lost packet is played back. In addition, the amount of traffic can be suppressed by using both low compression and high compression, and lost packets can be reliably reproduced on the receiving side to maintain the voice quality.

(Embodiment 4)
The configuration of the IP packetizing apparatus according to Embodiment 4 of the present invention is the configuration shown in FIG. Further, the transmission / reception operation in the IP packetizing apparatus according to the present embodiment is the operation shown in FIG. 6 described in the second embodiment. The difference is that the compression of the second embodiment is a low compression with a small degree of deterioration of the voice quality, but in this embodiment, the compression is high. That is, instead of taking in data every 10 ms and performing low compression, for example, taking data every 15 ms and making it high compression. By doing so, the total number of packets to be transmitted is reduced compared to the low compression of the second embodiment. The amount of traffic is reduced. The method and operation for reproducing audio on the receiving side are the same as those in the second embodiment. The basic operation of reproducing lost packets is the same in all of the second embodiment, the third embodiment, and the fourth embodiment. However, the second embodiment is a low-compression voice that suppresses the degree of deterioration during compression. The priority is on quality, in this embodiment, traffic is prioritized to reduce the number of packets by high compression, and the third embodiment is in the middle. In the present embodiment as well, SNMP can be installed, and the packet transmission method can be the conventional method, the method of the second or third embodiment, or the method of the present embodiment according to the traffic state.

  In this way, in this embodiment, by transmitting and receiving highly compressed audio data in the same packet, it is possible to reduce the degradation of audio quality while reducing the traffic volume to the maximum.

As described above, according to the present embodiment, the voice compression block 11 has the previous voice data and the current voice data in one packet, and the voice detector 10 is higher than the previous voice data and the current voice data. By performing compression, it is possible to reproduce lost packets on the receiving side even if packet loss occurs, and to suppress the traffic volume to the maximum by using high compression, and to reproduce lost packets. It is possible to maintain the voice quality by reliably performing on the receiver side.

  IP packetizing apparatus used for IP telephones or IP-PBXs, which compresses voice data, converts it into IP packets, and transmits the IP packets to an IP network so that voice quality does not deteriorate even if the traffic volume in the IP network increases be able to.

1 is a block diagram showing an IP packetizing apparatus according to Embodiment 1 of the present invention. (A) Timing diagram showing data indicating voice level, (b) Timing diagram showing IP packet Protocol configuration diagram showing protocols installed in IP phone and router (A) Timing diagram showing an audio analog signal, (b) Time division diagram showing that the division time is 10 ms, (c) Timing diagram showing IP packetization, (d) Packet data showing IP packets on the transmission path (E) Packet data diagram showing an IP packet when there is no packet loss, (f) Packet data diagram showing an IP packet when a packet loss occurs (A) Timing diagram showing an audio analog signal, (b) Time division diagram showing that the division times are 10 ms and 15 ms, (c) Timing diagram showing IP packetization, (d) IP packet on the transmission path Packet data diagram, (e) Packet data diagram showing IP packet when there is no packet loss, (f) Packet data diagram showing IP packet when packet loss occurs Data structure diagram showing schematic structure of RTP packet A block diagram showing an IP telephone system using a conventional IP packetizing apparatus for voice communication via an IP network by converting voice into IP packets using VoIP technology or the like Block diagram showing network monitoring agent connected to packet network

Explanation of symbols

1a, 1b IP phone 2a, 2b IP-PBX
3a, 3b PC (PC)
4a, 4b, 5a, 5b, 5c, 5d Router 6 A / D converter 7 CPU
8 FIFO memory 9 Electronic volume 10 Audio detector 11 Audio compression block 12 Switch 13 Low compression unit 14 High compression unit 15 Silent compression unit

Claims (4)

A voice processing block that compresses voice data, converts it into an IP packet and transmits it to the IP network, and a central processing unit that controls the whole and communicates with a route selection device such as a router in the IP network by using SNMP to grasp traffic information An IP packetizing device having:
An IP packetization apparatus comprising: a voice detector that determines a compression rate in the voice compression block based on the grasped traffic information. The speech compression block includes previous speech data and current speech data in one packet, and the speech detector performs low compression on the previous speech data and the current speech data. 2. The IP packetizing apparatus according to 1. The audio compression block has previous audio data and current audio data in one packet, and the audio detector performs low compression on the audio data of either the previous audio data or the current audio data, 2. The IP packetizing apparatus according to claim 1, wherein high compression is performed on the other voice data of the previous voice data and the current voice data. The voice compression block includes previous voice data and current voice data in one packet, and the voice detector performs high compression on the previous voice data and the current voice data. 2. The IP packetizing apparatus according to 1.

Images