JP2009194507A - Packet communication device, packet communication method, and packet communication program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To achieve distribution control which includes a high effect of suppressing the degradation of quality against reproduction quality degradation factors occurring on a network in video/sound distribution. <P>SOLUTION: A packet communication device 200 includes at least a quality degradation factor extraction part 204 which extracts an occurrence pattern of quality degradation factors causing the degradation in reproduction quality of vide/sound packets, a model analysis/simulation part 205 which obtains occurrence probabilities of quality degradation factors with respect to a network model generated on the basis of behavior information of video/sound packets, a quality influence risk evaluation part 206 which sets priorities to quality degradation factors on the basis of occurrence probabilities of quality degradation factors and degree of influence upon reproduction quality of the quality degradation factors, and a control parameter calculation part 207 which selects a distribution control method of video/sound packets on the basis of the priorities and calculates a control parameter of the distribution control method. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

  The present invention relates to a packet communication device, a packet communication method, and a packet communication program, and more particularly to a packet communication device, a packet communication method, and a packet communication program for realizing data communication that requires real-time performance such as video and audio. .

  An increasing number of communication applications transmit video data and audio data as IP (Internet Protocol) packets from the transmission side to the reception side, and restore and reproduce the video data and audio data on the reception side. For example, IP telephone (VoIP: Voice over IP) on the Internet, video conferencing, video distribution such as movies and live shows.

  In such an application (hereinafter referred to as a video / audio distribution application), the influence of the network (specifically, an increase in delay or packet loss) is ignored as a factor that degrades the quality of video and audio. I can not. In order to avoid or reduce this effect, many video and audio distribution applications are based on statistical information (average delay and packet loss rate) on network quality reported from the receiver side through means such as RTCP (RTP Control Protocol). Thus, control such as changing the transmission bit rate and interleaving is performed.

  For example, Non-Patent Document 1 proposes application of a technique called an interleave method. The interleaving method is a method of increasing the resistance to continuous loss by changing the packet transmission order in the form as shown in FIG. In FIG. 1, 16 consecutive packets A1 to A4, B1 to B4, C1 to C4, and D1 to D4 are assumed. Normally, A1 → A2 → A3 → A4 → B1 → B2 is transmitted in accordance with the order of state 1 in FIG. 1. However, in order to increase resistance to continuous loss, the packet transmission order as in state 2 in FIG. Sort by. By switching the packet transmission order in this way, even if a continuous packet loss occurs as in state 3 in FIG. 1, by reconstructing the packet order, continuous loss as in state 4 in FIG. Rather, it is possible to make it appear that a single loss has occurred. If the packet transmission order is not changed, packets are continuously lost in the form as in state 3 ′ in FIG. 1, and video and audio are interrupted.

  In the interleaving scheme, continuous loss is dealt with by changing the transmission order of packets. However, the longer the interleave length, the higher the resistance to continuous loss and the greater the delay. For example, FIG. 2 shows a packet transmission order in the case of interleave lengths 4 (left diagram) and 6 (right diagram). When the interleave length in the left diagram of FIG. 2 is 4, three packets are inserted between the packets A2 to be reproduced next to A1. When the interleave length is 6, five packets are inserted between the packets A2 to be reproduced next to A1. Since the packets sandwiched between them are recognized as a delay, the time required for recovering the packet loss that occurred once is increased accordingly.

Y.J Liang, J.G Apostolopoules, and B. Girod, “Analysis of packet Loss. For Compressed video: Does Burst-Length Matter?”, IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP2003, Apr. 2003

  However, conventionally, since quality control is performed based on information transmitted from the receiver side during packet transmission, it is difficult to set appropriate control parameters in advance in video and audio network distribution.

  The present invention has been made in view of the above problems, and its main object is to provide a packet communication device, a packet communication method, and a packet communication program capable of preventing deterioration in reproduction quality of video / audio packets. There is to do.

  In order to achieve the above object, a packet communication apparatus according to the present invention includes a quality deterioration factor extraction unit that extracts a generation pattern of a quality deterioration factor that causes deterioration in reproduction quality of a video / audio packet, and behavior information of the video / audio packet. Based on the model analysis / simulation unit for determining the occurrence probability of the quality deterioration factor for the network model generated on the basis of the above, the occurrence probability of the quality deterioration factor and the influence of the quality deterioration factor on the reproduction quality, A quality impact risk evaluation unit that sets priority for the quality degradation factor, and a control parameter calculation that selects a delivery control method of the video / audio packet based on the priority and calculates a control parameter of the delivery control method Part.

  The present invention also includes a plurality of networks including a packet communication device and a relay device, and each of the relay devices includes the above-described means in a packet communication system in which each of the networks is connected by the relay device. The control parameter is calculated for the network to which the relay device belongs and the network between the relay devices, and the video / audio packet transfer process is executed based on the calculated control parameter. This is to notify the packet communication device of the network to which the relay device belongs.

  The packet communication method according to the present invention also includes a first step of extracting an occurrence pattern of a quality deterioration factor causing deterioration in reproduction quality of video / audio packets, and a network generated based on the behavior information of the video / audio packets. Priority is given to the quality degradation factor based on the second step of determining the probability of occurrence of the quality degradation factor for the model, and the occurrence probability of the quality degradation factor and the degree of influence of the quality degradation factor on reproduction quality. And a fourth step of selecting a delivery control method of the video / audio packet based on the priority and calculating a control parameter of the delivery control method. is there.

  Further, the packet communication program of the present invention is based on the quality deterioration factor extraction unit for extracting the occurrence pattern of the quality deterioration factor that causes the deterioration of the reproduction quality of the video / audio packet, and the behavior information of the video / audio packet. Based on the generated network model, a model analysis / simulation unit for determining the occurrence probability of the quality deterioration factor, the occurrence probability of the quality deterioration factor, and the degree of influence of the quality deterioration factor on reproduction quality, the quality deterioration factor And a quality influence risk evaluation unit that sets a priority, and a control parameter calculation unit that selects a delivery control method of the video / audio packet and calculates a control parameter of the delivery control method.

  According to the packet communication device, the packet communication method, and the packet communication program of the present invention, in an application that requires real-time performance such as video and audio distribution, an appropriate countermeasure is taken against the occurrence of quality degradation factors caused by the network. It is possible to suppress the reproduction quality degradation of video and audio.

  The reason is that by determining the probability of occurrence of a possible quality degradation factor from the observed packet behavior and setting the priority to the quality degradation factor that should be addressed from the magnitude of the effect on playback quality, This is because it becomes possible to implement appropriate measures for preventing quality degradation.

  DESCRIPTION OF EMBODIMENTS Hereinafter, embodiments for carrying out the present invention will be described with reference to the drawings. FIG. 3 is a configuration diagram showing a connection relationship of terminals in the embodiment of the present invention. As shown in FIG. 3, the transmission side terminal 110 and the reception side terminal 120 are connected via a network 130.

  The transmission-side terminal 110 includes a packet communication device 111 that transmits a packet to the network 130 and a video / audio distribution application 112 that sends application data to the network 130 through the packet communication device 111.

  The receiving side terminal 120 includes a packet communication device 121 that receives a packet from the network 130 and a video / audio distribution application 122 that receives packet data from the packet communication device 121.

  Further, the network 130 provides a route 131 and a route 132 as routes connecting the transmission-side terminal 110 and the reception-side terminal 120.

  Note that the packet communication device 111 and the packet communication device 121 are the same packet communication device, but the transmitting side terminal 110 and the receiving side terminal 120 play different roles, and thus are distinguished from each other. Further, the roles of the transmitting side and the receiving side of the transmitting side terminal 110 and the receiving side terminal 120 are fixed for convenience of explanation, and there is actually no problem in that the voluntary roles are switched and data is transmitted and received between them. .

[First Embodiment]
FIG. 4 is a block diagram of the first embodiment of the packet communication apparatus according to the present invention. Referring to FIG. 4, a packet communication apparatus 200 according to the present embodiment includes a network data collection unit 201, a network modeling unit 202, a control parameter database unit 203, a quality degradation factor extraction unit 204, a model analysis / simulation unit 205, a quality An impact risk evaluation unit 206 and a control parameter calculation unit 207 are included. These means operate as follows.

  The network data collection unit 201 receives video packets and audio packets transmitted from the opposite packet communication device through the network, and collects behavior information of the received packets. The network modeling unit 202 models the behavior of the currently used network from the behavior information of video packets and audio packets. The control parameter database unit 203 stores information on video and audio distribution control parameters that can be provided by the packet communication device 200, quality degradation factors to which the parameters can be applied, and effects at the time of application. The quality degradation factor extraction unit 204 extracts the generation parameters of quality degradation factors that may be generated from the control parameters for video and audio distribution and the application area. The model analysis / simulation unit 205 calculates the occurrence probability of the occurrence pattern of the quality deterioration factor through the analytical method or the simulation method based on the behavior model of the network. The quality impact risk evaluation unit 206 sets a priority for the occurrence pattern of the quality degradation factor to be dealt with based on the influence of the occurrence pattern of the quality degradation factor on the reproduction quality of video and audio and the probability of the occurrence. The control parameter calculation unit 207 calculates a video and audio distribution control method and control parameters for suppressing the influence of the occurrence pattern of the quality degradation factor having a high priority.

  Each of the above means may be provided as hardware in the packet communication apparatus 200, or a computer is configured as a packet communication program that causes the packet communication apparatus 200 to function as each of the above means, and the packet communication program is operated on the packet communication apparatus 200. It is good also as a structure.

  Next, the overall operation of the present embodiment will be described in detail with reference to the overall configuration diagram of FIG. 3, the configuration diagram of FIG. 4, and the flowchart diagram of FIG.

  The video / audio distribution application 122 that configures the receiving terminal 120 transmits the video / audio that configures the opposing transmitting terminal 110 through the network 130 (assuming that only the path 131 or the path 132 is used as the network path). The video or audio packet transmitted from the audio distribution application 112 is received, and the network data collection unit 201 collects behavior information of the received packet (step S101).

  The network modeling unit 202 uses the received packet behavior information collected by the network data collection unit 201 to model the network behavior or modify the parameters of the already modeled network model (step S102).

  The quality degradation factor extraction unit 204 acquires information on control parameters that can be provided by the video / audio distribution application 122 from the control parameter database unit 203, and extracts occurrence patterns of quality degradation factors to which these control parameters are applied ( Step S103).

  The model analysis / simulation unit 205 calculates the occurrence probability of the occurrence pattern of the quality degradation factor extracted in step S103 by using an analytical method or a simulation method for the network model acquired in step S102 (step S104). .

  The quality impact risk evaluation unit 206 assigns a priority to the quality degradation factor to be taken from the magnitude of the influence of the occurrence pattern of the quality degradation factor extracted in step S103 on the reproduction quality and the occurrence probability of the occurrence pattern obtained in step S104. Setting is performed (step S105).

  The control parameter calculation unit 207 selects a control method for suppressing the influence of the high-priority quality degradation factor set in step S105 from the control methods stored in the control parameter database unit 203, and the required quality The control parameters are calculated so as to satisfy (step S106).

  A method is conceivable in which the processing from step S101 to step S106 is performed periodically or at an arbitrary timing according to the changing network state.

  The effect of this embodiment is that, in applications that require real-time performance such as video and audio distribution, appropriate measures are taken against the occurrence of quality degradation factors caused by the network, and video and audio playback quality degradation Can be suppressed. The reason is that the probability of occurrence of possible quality degradation factors is determined from the observed packet behavior (behavior information), and the priority is set to the quality degradation factor to be dealt with based on the magnitude of the effect on playback quality. This is because it becomes possible to carry out distribution control that suppresses the influence of quality deterioration that occurs frequently.

[Second Embodiment]
Next, a second embodiment of the present invention will be described with reference to FIG. Referring to FIG. 6, a packet communication apparatus 300 according to the present embodiment includes a network data collection unit 301, a model reconstruction determination unit 302, a network modeling unit 303, a control parameter database unit 304, a quality degradation factor extraction unit 305, a model. An analysis / simulation unit 306, a quality impact risk evaluation unit 307, and a control parameter calculation unit 308 are included. These means operate as follows.

  The network data collection unit 301 receives video packets and audio packets transmitted from the opposite packet communication device through the network, and collects behavior information of the received packets. The model reconstruction determination unit 302 determines whether it is necessary to reconstruct the network model from the behavior information of the received packet. The network modeling unit 303 performs modeling of the behavior of the currently used network from the behavior of video packets and audio packets. The control parameter database unit 304 stores information on video and audio distribution control parameters that can be provided by the packet communication device 300, quality degradation factors to which the parameters can be applied, and effects at the time of application. The quality degradation factor extraction unit 305 extracts the generation pattern of quality degradation factors that may occur from the control parameters for video and audio distribution and the application area. The model analysis / simulation unit 306 calculates the occurrence probability of the occurrence pattern of the quality deterioration factor through the analytical method or the simulation method based on the behavior model of the network. The quality impact risk evaluation unit 307 sets the priority for the quality degradation factor generation pattern to be dealt with based on the influence of the occurrence pattern of the quality degradation factor on the reproduction quality of video and audio and the probability of the occurrence. The control parameter calculation unit 308 calculates a video and audio distribution control method and control parameters for suppressing the influence of the occurrence pattern of the quality degradation factor having a high priority.

  Each of the above means may be provided as hardware in the packet communication apparatus 300, or a computer is configured as a packet communication program that functions as each of the above means, and the packet communication program is operated on the packet communication apparatus 300. It is good also as a structure.

  Next, the overall operation of the present embodiment will be described in detail with reference to the overall configuration diagram of FIG. 3, the configuration diagram of FIG. 6, and the flowchart diagram of FIG.

  It is assumed that the transmission side terminal 110 and the reception side terminal 120 are communicating through a network path 131. The video / audio distribution application that configures the reception-side terminal 120 receives video and audio packets transmitted from the video / audio distribution application that configures the opposite transmission-side terminal 110 via the network 130, and the network data collection unit 301. Collects the behavior information of the received packet (step S201).

  The model reconstruction determination unit 302 calculates the average delay, the maximum value of the delay, the minimum value, and the variance indicating the fluctuation width from the behavior information of the received packet, and the calculation result is the past average delay, maximum value, minimum Compare with value and variance. If the numerical value is greatly changed, it is determined that the network model has changed (in the example of FIG. 3, the network route has changed from the route 131 to the route 132), and the network modeling unit indicates that the model needs to be rebuilt. 303 is notified (step S202).

  When receiving a model reconstruction notification from the model reconstruction determination 302, the network modeling unit 303 discards the network model held so far and constructs a new model. If no model reconstruction notification has been received, the parameters of the network model already modeled are corrected (step S203).

  The quality degradation factor extraction unit 305 acquires information on control parameters that can be provided by the video / audio distribution application from the control parameter database unit 304, and extracts occurrence patterns of quality degradation factors to which the control parameters are applied (step). S204).

  The model analysis / simulation unit 306 calculates the occurrence probability of the occurrence pattern of the quality deterioration factor extracted in step S204 by using an analytical method or a simulation method for the network model acquired in step S203 (step S205). .

  The quality impact risk evaluation unit 307 assigns a priority to the quality degradation factor to be taken from the magnitude of the influence of the occurrence pattern of the quality degradation factor extracted in step S204 on the reproduction quality and the occurrence probability of the occurrence pattern obtained in step S205. Setting is made (step S206).

  The control parameter calculation unit 308 selects, from the control methods stored in the control parameter database unit 304, a control method that suppresses the influence of the high-priority quality degradation factor set in step S206, and the required quality. The control parameters are calculated so as to satisfy (step S207).

  The processing from step S201 to step S207 may be performed periodically, or may be performed at an arbitrary timing according to the changing network state.

  The first effect of the present embodiment is that in applications that require real-time performance such as video and audio distribution, appropriate countermeasures are taken against the occurrence of quality degradation factors caused by the network, and video and audio It is possible to suppress the reproduction quality deterioration. The reason is that the probability of occurrence of possible quality degradation factors is determined from the observed packet behavior (behavior information), and the priority is set to the quality degradation factor to be dealt with based on the magnitude of the effect on playback quality. This is because it becomes possible to carry out distribution control that suppresses the influence of quality deterioration that occurs frequently.

  The second effect of the present embodiment is that even when the network that is the target of route change or the like is changed, distribution control that follows the change can be performed. The reason is that by storing the received packet behavior information such as the fluctuation width of the delay from the received packet behavior information, if a large change is detected in the behavior information, the network model is reconstructed, This is because it can follow changes in the target network.

  Next, the operation of the best mode for carrying out the present invention will be described using specific examples. FIG. 8 is a configuration example of a video / audio distribution system configured using the packet communication apparatus including the first embodiment or the second embodiment.

  The video / audio distribution system in FIG. 8 connects these terminals to the video / audio distribution terminal 1100 and the video / audio distribution terminal 1200 which are packet communication apparatuses including the first embodiment or the second embodiment. The network 1000 is configured. Here, for the sake of simplicity, it is assumed that the video / audio distribution terminal 1100 is transmitting audio data to the video / audio distribution terminal 1200. The video / audio distribution terminal 1100 transmits audio packets to the video / audio distribution terminal 1200 at intervals of 20 milliseconds, and the video / audio distribution terminal 1200 reproduces the received audio packets. This processing is similarly performed for video packets, and is performed similarly even when the transmission / reception relationship is reversed.

  The video / audio distribution terminal 1200 collects behavior information of received audio packets and performs network modeling. In this embodiment, the hidden Markov model shown in FIG. 9 is used as a modeling method. The hidden Markov model is a probability model composed of a plurality of internal states that transition according to a Markov process and the output probability distribution of symbols in each internal state. In the network modeling in this embodiment, the hidden Markov model is output from each internal state. Delay or packet loss is adopted as the symbol. 10 and 11 are a state transition probability table and a symbol output probability table in the hidden Markov model generated from the collected behavior information of the voice packet. The video / audio distribution terminal 1200 can update the state transition probability table and the symbol output probability table in the hidden Markov model periodically or at an arbitrary timing.

  Next, the video / audio distribution terminal 1200 extracts a quality deterioration factor to be considered from the executable control method. The quality degradation factors that are extracted include the jitter fluctuation fluctuation width for controlling the jitter buffer length that absorbs fluctuations in the received packet delay fluctuation, called jitter, and the continuous interleaving length that increases the resistance to packet loss. There are packet loss numbers and so on. In this embodiment, the number of consecutive packet losses is considered. The video / audio distribution terminal 1200 obtains the probability of occurrence of each successive packet loss using the hidden Markov model when the number of successive packet losses changes to one, two,. In order to obtain the probability of occurrence using the hidden Markov model, one of the following two methods is used.

  The first method is an analytical method using the state transition probability table of FIG. 10 and the symbol output probability table of FIG. Specifically, if it is assumed that the user stays in the internal state A at a certain stage, the probability that a packet loss will occur is Y (A, L), and any state, for example, the state B at the next time point The probability that a packet loss will occur is obtained by X (A, B) Y (B, L). In this way, by adding the transition probability to each state and the packet loss occurrence probability at the transition destination, it is possible to calculate the probability when two packet losses continue. Similarly, when the packet loss continues three times or more, it can be obtained analytically.

  As another method, using a hidden Markov model and random numbers according to the state transition probability table of FIG. 10 and the symbol output probability table of FIG. 11, a large number of consecutive packet behaviors are obtained by simulation, and packet loss occurs continuously. This is a method of obtaining the occurrence probability from the occurrence frequency of the behavior. This is a method generally known as the Monte Carlo method. FIG. 12 is a table in which the deterioration value and countermeasure priority are calculated as the occurrence probability of the number of consecutive packet losses and the effect of the event on the reproduction quality of the sound obtained in this way. The influence on the voice reproduction quality is calculated using an evaluation formula such as E-Model that expresses the voice quality numerically, taking into account the influence of delay and packet loss, for example. In the present embodiment, the countermeasure priority is a multiplication value of the occurrence probability and the reproduction quality deterioration value.

  The video / audio distribution terminal 1200 calculates the control method and its control parameters using FIG. For example, when taking countermeasures against quality degradation factors up to the second highest priority of countermeasure priority, 2 is selected as the interleave length so as to suppress the effects of single packet loss and two consecutive packet losses. The calculated control method and its control parameters are notified from the video / audio distribution terminal 1200 to the video / audio distribution terminal 1100, and the video / audio distribution terminal 1100 applies to the distribution control, or the video / audio distribution terminal 1200 It is used in the form of being applied to distribution control to the video / audio distribution terminal 1100.

  FIG. 13 is a configuration example of a video / audio distribution system configured using the packet communication apparatus including the first embodiment or the second embodiment.

  In the video / audio distribution system of FIG. 13, video / audio distribution is performed via relay apparatuses 2200 and 2300 which are packet communication apparatuses including the first embodiment or the second embodiment, and relay apparatuses 2200 and 2300. Video / audio distribution terminals 2400, 2500, 2600, and 2700 that perform the above and a network that connects these terminals. This network includes a network 2000 to which the relay apparatus 2200 and video / audio distribution terminals 2400 and 2500 belong, and a network 2100 to which the relay apparatus 2300 and video / audio distribution terminals 2600 and 2700 belong.

  Assume that video / audio distribution has been performed between the video / audio distribution terminal 2400 and the video / audio distribution terminal 2600 connected to the network at a certain point in time. In this case, the relay apparatus 2200 receives video and audio packets transmitted from the video / audio distribution terminal 2400 to the video / audio distribution terminal 2600 and transfers the packets to the relay apparatus 2300. The relay apparatus 2300 transfers the transferred video / audio packet to the video / audio distribution terminal 2600. The video / audio packet transmitted from the video / audio distribution terminal 2600 to the video / audio distribution terminal 2400 is reversely processed.

  The relay device 2200 and the relay device 2300 collect packet behavior information at the same time as performing packet transfer processing. Similarly to the case described in the first embodiment, the relay device 2200 determines the network between the video / audio distribution terminal 2400 and the relay device 2200, the relay device 2200, and the relay device 2300 from the collected behavior information of the received packet. For the two networks, the network between the two, network modeling, extraction of quality degradation factors, priority setting taking into account the probability of occurrence of quality degradation factors and the degree of influence of quality degradation, and calculation of control method and control parameters The calculation result is applied to the distribution control during the transfer process. Also, the calculation result is notified to the video / audio distribution terminal 2400. Similarly, the relay apparatus 2300 performs the same processing for two networks: a network between the video / audio distribution terminal 2600 and the relay apparatus 2300 and a network between the relay apparatus 2300 and the relay apparatus 2200.

  Next, the video / audio distribution terminals 2500 and 2700 start video / audio distribution. At this time, the relay apparatus 2200 determines that a common network model can be used because the video / audio distribution terminal 2400 and the video / audio distribution terminal 2500 belong to the same network, and notifies the video / audio distribution terminal 2400. The control method and control parameters are notified to the video / audio distribution terminal 2500. The relay apparatus 2300 performs the same determination / processing for the video / audio distribution terminal 2700. As a result, the video / audio distribution terminal 2500 and the video / audio distribution terminal 2700 can implement appropriate distribution control at the same time as starting the video / audio distribution service.

  INDUSTRIAL APPLICABILITY The present invention can be used for a packet communication apparatus, a packet communication method, and a packet communication program in data communication that requires real-time performance such as video / audio.

It is a figure which shows the packet transmission by an interleaving system. It is a figure explaining the interleaving length in an interleaving system. It is a connection block diagram of the terminal in 1st and 2nd embodiment. It is a block diagram which shows the structure of the packet communication apparatus of 1st Embodiment. It is a flowchart figure which shows the packet communication method of 1st Embodiment. It is a block diagram which shows the structure of the packet communication apparatus of 2nd Embodiment. It is a flowchart figure which shows the packet communication method of 2nd Embodiment. 1 is a configuration example of a video / audio distribution system according to a first embodiment. It is a figure explaining the hidden Markov model which is a network modeling method. It is a figure which shows the state transition probability table in a hidden Markov model. It is a figure which shows the symbol output probability table | surface in a hidden Markov model. It is a figure which shows the countermeasure priority with respect to the quality degradation factor in a 1st Example. It is a structural example of the video / audio distribution system in the second embodiment.

Explanation of symbols

DESCRIPTION OF SYMBOLS 110 Transmission side terminal 111, 121 Packet communication apparatus 112, 122 Video / audio delivery application 120 Reception side terminal 130 Network 131, 132 Path 200 Packet communication apparatus 201 Network data collection part 202 Network modeling part 203 Control parameter database part 204 Quality degradation factor Extraction unit 205 Model analysis / simulation unit 206 Quality influence risk evaluation unit 207 Control parameter calculation unit 300 Packet communication device 301 Network data collection unit 302 Model reconstruction determination unit 303 Network modeling unit 304 Control parameter database unit 305 Quality degradation factor extraction unit 306 Model analysis / simulation unit 307 Quality impact risk evaluation unit 308 Control parameter calculation unit 1000 Work 1100 and 1200 video and audio terminal 2000 and 2100 network 2200, 2300 relay device 2400,2500,2600,2700 video and audio terminal

Claims (10)

A quality degradation factor extraction unit that extracts occurrence patterns of quality degradation factors that cause degradation of reproduction quality of video and audio packets;
A model analysis / simulation unit for determining the occurrence probability of the quality deterioration factor for the network model generated based on the behavior information of the video / audio packet;
Based on the occurrence probability of the quality degradation factor and the degree of influence of the quality degradation factor on the reproduction quality of the video / audio packet, a quality impact risk evaluation unit that sets a priority for the quality degradation factor;
A packet communication apparatus comprising: a control parameter calculation unit that selects a video / audio packet distribution control method based on the priority and calculates a control parameter of the distribution control method. From a change in statistical information regarding the behavior information of the video / audio packet, further comprising a model reconstruction determination unit that determines the necessity of reconstruction of the model of the network,
The packet communication apparatus according to claim 1, wherein the model analysis / simulation unit obtains an occurrence probability of the quality degradation factor for the reconstructed network model. The model analysis / simulation unit uses a hidden Markov model to determine the probability that a continuous packet loss will occur,
The packet communication device according to claim 1, wherein the control parameter calculation unit calculates an interleave length in an interleave method. In a packet communication system comprising a plurality of networks including a packet communication device and a relay device, each of the networks being connected by the relay device,
Each said relay apparatus is provided with the means as described in any one of Claim 1 thru | or 3, calculates the said control parameter with respect to the network to which the own relay apparatus belongs, and the network between relay apparatuses, and calculated control parameter A packet communication system characterized in that the video / audio packet transfer processing is executed based on the information and the calculated control parameter is notified to the packet communication device of the network to which the relay device belongs. A first step of extracting an occurrence pattern of a quality deterioration factor that causes deterioration in reproduction quality of video / audio packets;
A second step of obtaining an occurrence probability of the quality degradation factor for the network model generated based on the behavior information of the video / audio packet;
A third step of setting a priority for the quality degradation factor based on the occurrence probability of the quality degradation factor and the degree of influence of the quality degradation factor on reproduction quality;
A packet communication method comprising: a fourth step of selecting a delivery control method for the video / audio packet based on the priority and calculating a control parameter of the delivery control method. Further comprising the step of determining the necessity of rebuilding the network model from a change in statistical information regarding the behavior information of the video / audio packet,
6. The packet communication method according to claim 5, wherein, in the second step, an occurrence probability of the quality deterioration factor is obtained for the reconstructed network model. In the second step, a hidden Markov model is used to determine the probability that a continuous packet loss will occur,
The packet communication method according to claim 5 or 6, wherein in the fourth step, an interleave length in an interleave method is calculated. Computer
A quality degradation factor extraction unit that extracts occurrence patterns of quality degradation factors that cause degradation of the playback quality of video and audio packets,
A model analysis / simulation unit for determining the occurrence probability of the quality degradation factor for the network model generated based on the behavior information of the video / audio packet;
A quality impact risk evaluation unit that sets a priority for the quality degradation factor based on the occurrence probability of the quality degradation factor and the degree of influence of the quality degradation factor on reproduction quality;
A packet communication program that functions as a control parameter calculation unit that selects a delivery control method of the video / audio packet based on the priority and calculates a control parameter of the delivery control method. Computer,
From a change in statistical information regarding the behavior information of the video / audio packet, function as a model reconstruction determination unit that determines the necessity of reconstruction of the model of the network,
9. The packet communication program according to claim 8, wherein the model analysis / simulation unit obtains an occurrence probability of the quality degradation factor for the reconstructed network model. The model analysis / simulation unit uses a hidden Markov model to determine the probability that a continuous packet loss will occur,
The packet communication program according to claim 8 or 9, wherein the control parameter calculation unit calculates an interleave length in an interleave method.

Images