JP4800250B2 - Packet receiving apparatus, method and program for determining necessary and sufficient reception buffer size - Google Patents

Description

  The present invention relates to a packet reception apparatus, method, and program for determining a necessary and sufficient reception buffer size. In particular, the present invention relates to a technique for designing or controlling the reception buffer size of a packet receiving apparatus that receives IP (Internet Protocol) packets including multimedia data such as video and audio via the Internet.

  FIG. 1 is a functional configuration diagram of a packet receiving apparatus based on the prior art.

  According to FIG. 1, the packet receiver 1 receives a packet from the packet transmitter 2 via the network. If the network is an IP network such as the Internet, the transferred packet is an IP packet. Media data formats included in the IP packet include file data, video data, audio data, and the like.

  According to FIG. 1, a packet receiver 1 which is a terminal such as a personal computer receives and reproduces streaming data (video data + audio data) from a packet transmitter 2 such as a streaming content distribution server. . At this time, the packet receiving apparatus 1 reproduces the streaming data at the same time interval as the packet transmission time interval, and reproduces the video data and the audio data in synchronization.

  1 includes a reception buffer unit 101, a packet analysis unit 102, a video decoder unit 111, a video buffer unit 112, a display unit 113, an audio decoder unit 121, an audio buffer unit 122, A speaker unit 123 and a synchronization control unit 103 are included. The reception buffer unit 101 temporarily buffers IP packets received via the Internet. The packet analysis unit 102 analyzes the format of the media data included in the IP packet, outputs video data to the video decoder unit 111, and outputs audio data to the audio decoder unit 121. The video decoder unit 111 decodes the video data included in the IP packet and outputs the video signal to the video buffer unit 112. The video buffer unit 112 outputs a video signal to the display unit 113 at a timing according to an instruction from the synchronization control unit 103. The audio decoder unit 121 decodes the audio data included in the IP packet and outputs the audio signal to the audio buffer unit 122. The audio buffer unit 122 outputs an audio signal to the speaker unit 123 at a timing according to an instruction from the synchronization control unit 103. The synchronization control unit 103 instructs the output timing to the video buffer unit 112 and the audio buffer unit 122 so that the video signal and the audio signal are output in synchronization.

  As shown in FIG. 1, jitters caused by the network can be absorbed by temporarily collecting received packets and sequentially reproducing them. Jitter refers to fluctuation of a signal in the time axis direction. In the case of a packet, it means fluctuation (change amount) of transmission delay (= reception time−transmission time).

  There are various conventional techniques for realizing high-quality data transmission by absorbing jitter generated by a communication error. For example, there is a technique for discarding a received packet by controlling a time stamp added to the received packet (see, for example, Patent Document 1). There is also a technique for reducing the influence on the packet receiving device by controlling the fluctuation absorbing buffer considering only the jitter of the network delay for the packet transmitting device (or relay device) (see, for example, Patent Document 2). In addition, there is a technique for transmitting streaming data at a TCP overrate, executing flow control by RTSP (Real Time Streaming Protocol) in the background, and holding the remaining amount of received data within a specified range (see, for example, Patent Document 3). ).

JP 2006-203649 A JP 2006-094130 A JP 2005-303783 A

  Here, there is a problem of how much the size of the reception buffer size of the reception buffer unit is determined in order to absorb jitter. The above-mentioned prior art patent documents try to absorb jitter by fixing the reception buffer size and controlling the part other than the reception buffer. That is, according to the prior art, the reception buffer size taking into consideration the jitter characteristics based on the network has not been sufficiently studied.

  When the network jitter is large, the packet arrival time becomes long, and the reception buffer may become empty. At this time, the video buffer and the audio buffer are also emptied, and the video and audio are temporarily stopped. Thereafter, there are cases where packets are continuously received at a time and the reception buffer overflows (full). At this time, the state is such that video and audio are rapidly reproduced. Further, the reception buffer discards the overflow packet, and an intermittent state occurs in the video and audio.

  When the reception buffer size is set to a fixed value, the packet is often discarded if it is small relative to the jitter of the network. Then, the reception buffer size must be increased, and the cost of the memory increases. On the other hand, if it is large with respect to the jitter of the network, the time for which the packet is temporarily stored in the reception buffer also becomes long, resulting in a delay. This is not appropriate for interactive communication.

  There is also a method of making the reception buffer size variable according to the network state. According to this method, when the reception buffer overflows, playback is stopped once and the reception buffer size is increased. However, every time the reception buffer size is changed, it is necessary to stop reproduction, which requires rebuffering.

  Accordingly, an object of the present invention is to provide a packet receiving apparatus, method, and program capable of determining a necessary and sufficient reception buffer size.

According to the present invention, there is provided a reception buffer that is designed to prevent packet loss and order reversal, is connected to a network that repeatedly increases and decreases packet transmission delay, and receives packets via the network. In the packet receiver,
The network has a tendency that the transmission delay increases as the packet continues to be transmitted, and after reaching the maximum transmission delay, the network has a tendency to repeat the change such that the transmission delay starts to decrease,
A packet transmission time interval receiving means for receiving a packet transmission time interval Δ from the packet transmission device;
Jitter average maximum value calculating means for calculating twice (2Δ) the packet transmission time interval Δ as the maximum value of jitter average;
Receiving buffer size determining means for determining a receiving buffer size based on a maximum value of jitter average.

（分子）Ｍ／Ｄｏ×Ｄｏ＋Ｍ／Δ×Δ：遅延増加及び遅延減少の全伝送遅延
（分母）Ｍ／Ｄｏ＋Ｍ／Δ：送信パケット総数
Ｄｏを∞に大きくした場合であっても高々２×Δにしかならないジッタ平均の最大値は、以下の式に基づくものである

ことも好ましい。 According to another embodiment of the packet receiver of the present invention,
The jitter average value in the jitter average maximum value calculating means is based on the following model formula:
Do: Transmission delay increase for each packet
Δ: Transmission delay decrease for each packet (= packet transmission time interval)
M / Do: Number of transmitted packets with increased delay at maximum transmission delay
M / Do × Do: Maximum transmission delay with increased delay
M / Δ: the number of transmitted packets with reduced delay at the maximum transmission delay
M / Δ × Δ: Maximum transmission delay in delay reduction

(Molecular) M / Do × Do + M / Δ × Δ: total transmission delay of delay increase and delay decrease
(Denominator) M / Do + M / Δ: Total number of transmitted packets
Even when Do is increased to ∞, the maximum jitter average that can only be 2 × Δ is based on the following equation:

It is also preferable.

According to another embodiment of the packet reception apparatus of the present invention, it further comprises jitter measuring means for measuring jitter and deriving a statistical probability distribution composed of the jitter and the number of times (frequency).
The reception buffer size determination means preferably determines the reception buffer size based on the jitter average maximum value and the jitter standard deviation.

According to another embodiment of the packet reception apparatus of the present invention, the reception buffer size determination means is configured to receive the reception buffer when the probability distribution of jitter measured by the jitter measurement means follows a normal distribution of (m, σ ² ). It is also preferable that the size is set based on 2Δ (twice the packet transmission time interval Δ) + 3σ , and 99% or more of received packets are stably received and reproduced.

According to another embodiment of the packet receiving apparatus of the present invention, the reception buffer size determining means is such that the probability distribution of jitter measured by the jitter measuring means follows an exponential distribution of m (average value) = σ (standard deviation). In this case, it is also preferable to set the reception buffer size based on 3σ (6Δ: 6 times the packet transmission time interval Δ) and to stably receive and reproduce 95% or more of received packets.

According to another embodiment of the packet reception apparatus of the present invention, when the standard deviation σ is known, the reception buffer size determining means sets the reception buffer size based on 2Δ + 4σ , and receives 93.8% or more of received packets. It is also preferable to perform stable reception and reproduction.

  According to another embodiment of the packet receiving apparatus of the present invention, the packet transmission time interval receiving means is configured to change the packet transmission time after the change when the video encoding bit rate and the transmission packet size are changed during the transmission. It is also preferable to receive the interval Δ and output the packet transmission time interval Δ to the jitter average maximum value calculating means.

According to the present invention, there is provided a reception buffer that is designed to prevent packet loss and order reversal, is connected to a network that repeatedly increases and decreases packet transmission delay, and receives packets via the network. In the reception buffer size determination method in the packet reception device,
The network has a tendency that the transmission delay increases as the packet continues to be transmitted, and after reaching the maximum transmission delay, the network has a tendency to repeat the change such that the transmission delay starts to decrease,
A first step of receiving a packet transmission time interval Δ from the packet transmission device;
A second step of calculating twice (2Δ) the packet transmission time interval Δ as a maximum value of jitter average ;
Characterized in that it have a <br/> a third step of determining a receive buffer size based on the maximum value of the jitter mean.

（分子）Ｍ／Ｄｏ×Ｄｏ＋Ｍ／Δ×Δ：遅延増加及び遅延減少の全伝送遅延
（分母）Ｍ／Ｄｏ＋Ｍ／Δ：送信パケット総数
Ｄｏを∞に大きくした場合であっても高々２×Δにしかならないジッタ平均の最大値は、以下の式に基づくものである

ことも好ましい。 According to another embodiment of the reception buffer size determination method of the present invention,
The jitter average value in the second step is based on the following model equation:
Do: Transmission delay increase for each packet
Δ: Transmission delay decrease for each packet (= packet transmission time interval)
M / Do: Number of transmitted packets with increased delay at maximum transmission delay
M / Do × Do: Maximum transmission delay with increased delay
M / Δ: the number of transmitted packets with reduced delay at the maximum transmission delay
M / Δ × Δ: Maximum transmission delay in delay reduction

(Molecular) M / Do × Do + M / Δ × Δ: total transmission delay of delay increase and delay decrease
(Denominator) M / Do + M / Δ: Total number of transmitted packets
Even when Do is increased to ∞, the maximum jitter average that can only be 2 × Δ is based on the following equation:

It is also preferable.

According to the present invention, there is provided a reception buffer that is designed to prevent packet loss and order reversal, is connected to a network that repeatedly increases and decreases packet transmission delay, and receives packets via the network. In a program for causing a computer mounted on a packet receiving device to function,
The network has a tendency that the transmission delay increases as the packet continues to be transmitted, and after reaching the maximum transmission delay, the network has a tendency to repeat the change such that the transmission delay starts to decrease,
A packet transmission time interval receiving means for receiving a packet transmission time interval Δ from the packet transmission device;
Jitter average maximum value calculating means for calculating twice (2Δ) the packet transmission time interval Δ as the maximum value of jitter average;
Receiving buffer size determining means for determining a receiving buffer size based on a maximum value of jitter average.

  According to another embodiment of the packet reception program of the present invention,
  The jitter average value in the jitter average maximum value calculating means is based on the following model formula:
      Do: Transmission delay increase for each packet
      Δ: Transmission delay decrease for each packet (= packet transmission time interval)
      M / Do: Number of transmitted packets with increased delay at maximum transmission delay
      M / Do × Do: Maximum transmission delay with increased delay
      M / Δ: the number of transmitted packets with reduced delay at the maximum transmission delay
      M / Δ × Δ: Maximum transmission delay in delay reduction

  (Molecular) M / Do × Do + M / Δ × Δ: total transmission delay of delay increase and delay decrease
  (Denominator) M / Do + M / Δ: Total number of transmitted packets
  Even when Do is increased to ∞, the maximum jitter average that can only be 2 × Δ is based on the following equation:

It is also preferable to make the computer function.

  According to the packet reception apparatus, method and program of the present invention, the jitter average maximum value is calculated based only on the packet transmission time interval Δ in the packet transmission apparatus, and a necessary and sufficient reception buffer size covering the jitter average maximum value is obtained. Can be determined.

  Hereinafter, the best mode for carrying out the present invention will be described in detail with reference to the drawings.

  The packet reception apparatus according to the present invention determines the reception buffer size based on the jitter average maximum value. The jitter average maximum value is based on the packet transmission time interval time Δ transmitted from the packet transmission device. Based on the transmission time interval time Δ, the packet reception device calculates that the jitter average maximum value that will be observed by the packet reception device is 2Δ. Thereby, the packet reception apparatus can ensure a reception buffer size corresponding to the jitter average maximum value 2Δ [ms].

  FIG. 2 is an explanatory diagram of jitter.

According to FIG. 2, it is assumed that the packet transmission device transmits the packet (n−1) to the packet reception device, and that a transmission delay S _n−1 occurs for the packet (n−1). Also, then, the packet transmission device transmits packets (n) to the packet reception device, and transmission delay S _n with respect to the packet (n) occurs. At this time, the transmission delay increase amount D _n is as follows.
S _n : Transmission delay amount of the _nth packet S _n−1 : Transmission delay amount of the (n−1) th packet Transmission delay increase amount D _n = S _n −S _n−1
This variation in transmission delay increment D _n is the instantaneous jitter.

Further, the average jitter is expressed by the following formula 1.

  FIG. 3 is a model sequence diagram of jitter according to the present invention.

  The inventor has confirmed the operation shown in FIG. 3 by simulating a network such as the Internet and changing the amount of jitter therein. If there is jitter in the network that does not discard the packet, the packet delay is repeatedly increased and decreased. Therefore, the jitter measured for any network state reaches its peak. Here, when the maximum jitter can be specified, the reception buffer size corresponding to the jitter can be determined.

  The model sequence of FIG. 3 according to the present invention assumes a situation in which packet loss and packet order reversal do not occur. The packet loss is caused by queue overflow in the transmission / reception device and the relay device. In addition, packet reversal occurs when a plurality of paths are set between a transmission apparatus and a reception apparatus, and congestion occurs in one of the paths.

  In such a situation, the present invention is applied after solving by another method. For example, for packet loss, a packet is transmitted after a sufficient bandwidth is secured in the communication path. Further, for packet order reversal, a route is uniquely determined or congestion is suppressed by bandwidth management.

  According to FIG. 3, initially, a packet transmitted from the packet transmitting device is received by the packet receiving device with a certain transmission delay (packet 1). However, if the packet transmission device continues to transmit packets, the transmission delay gradually increases (packets 2 to 5). When the maximum transmission delay M is reached, the transmission delay starts to decrease rapidly (packets 6 to 9). Eventually, it returns to the first constant transmission delay (packet 10).

（分子）Ｍ／Ｄｏ×Ｄｏ＋Ｍ／Δ×Δ：遅延増加及び遅延減少の全伝送遅延
（分母）Ｍ／Ｄｏ＋Ｍ／Δ：送信パケット総数 The jitter average value based on such an operation is expressed by the following model formula.
Do: Transmission delay increase amount for each packet Δ: Transmission delay decrease amount for each packet (= packet transmission time interval)
M / Do: Number of transmission packets with increased delay at maximum transmission delay M / Do × Do: Maximum transmission delay at increased delay M / Δ: Number of transmitted packets with decreased delay at maximum transmission delay M / Δ × Δ: Maximum at decreased delay Transmission delay

(Numerator) M / Do × Do + M / Δ × Δ: total transmission delay of delay increase and delay decrease (denominator) M / Do + M / Δ: total number of transmitted packets

  According to Equation 2, when Δ = 10 ms and Do = 30 ms, the jitter average value Jave = 15 ms.

Here, it can be understood from Equation 3 that even if Do increases to ∞ in Equation 2, it is only 2 × Δ at most.

  Then, when Δ = 10 ms and Do → ∞ that gives the maximum value, the jitter average maximum value (upper limit value) Jave is 20 ms (2 · Δ) at most. Therefore, the reception buffer size in the packet reception apparatus can be calculated based on a probability distribution with a jitter average maximum value Jave = 20 ms.

  FIG. 4 is a graph showing a frequency distribution with respect to jitter.

  According to FIG. 4, the maximum jitter average value is 2Δ, and 2Δ + α taking into account the relief range from the jitter probability distribution is the reception buffer size. With respect to received packets up to the rescue range, a sufficient reception buffer size is ensured without falling into buffer empty, and smooth and stable reproduction is ensured without causing extra delay.

For example, when the probability distribution of jitter follows the normal distribution of (m, σ ² ) whose probability density function is shown in Equation 4 (m: average, σ ² : variance), m = 2Δ.

  At this time, by setting the reception buffer size to 2Δ + 3σ, 99% or more of received packets can be stably received and reproduced. Further, by setting 2Δ + 2σ, it is possible to stably receive and reproduce 95% or more of received packets. Thus, the reception buffer size can be determined according to the relief range. For example, when the strength of FEC (Frame Error Correct) is strong, it is considered that even if the ratio of the relief range is lowered to some extent, the reproduction is not hindered.

Further, when the probability distribution of jitter follows the exponential distribution indicating the probability density function in Equation 5, m (average value) = 2Δ.

  At this time, m (average value) = σ (standard deviation) = 2Δ. Therefore, by setting the reception buffer to 3σ (6Δ), it is possible to stably receive and reproduce 95% or more of received packets. Also, it is possible to stably receive and reproduce received packets of 86% or more at 2σ (4Δ) or 63% or more at σ (2Δ).

Expression 6 is an expression obtained by integrating the probability distribution of Expression 5 from 0 to t. The probability of relief is obtained by substituting t = 3σ, 2σ, and σ.

Ｘ：確率変数、ｋ：ｋ＞０の任意の数 Even when the probability distribution of jitter is not known, if the standard deviation σ of the distribution is known, the Chebyshev inequality in Equation 7 can be used to determine how many σ should be considered in order to obtain a desired relief range. .

X: random variable, k: any number with k> 0

  For example, if 8/9 (88.8%) relief is desired, K = 3 and 3σ may be set. That is, the reception buffer amount may be set to m + 3σ = 2Δ + 3σ. On the other hand, if 15/16 (93.8%) relief is desired, K = 4 and 4σ may be set. That is, the reception buffer amount may be set to m + 4σ = 2Δ + 4σ.

  Here, a specific numerical value is substituted for explanation.

When the video encoding bit rate is 4 Mpbs and the payload size of the packet to be transmitted is 1400 bytes, 357 packets are transmitted per second. At that time, the packet transmission time interval is 2.8 msec. Therefore, 2Δ = 5.6 msec.
Video coding bit rate: 4Mpbs
Packet payload size: 1400 bytes
Number of packets per second = (4Mbit / 8bit) / 1400 bytes
= 357 (pieces)
Packet transmission time interval Δ = 1000 (ms) / 357 (pieces) = 2.8 msec
2Δ = 2 × 2.8 msec = 5.6 msec

Here, when the probability distribution of jitter follows the exponential distribution, it is as follows.
m (average value) = σ (standard deviation) = 2Δ = 5.6 msec
Receiving buffer size 3σ (6Δ) = 16.8 msec-> 95% or more stable reproduction
2σ (4Δ) = 11.2 msec-> 86% or more stable reproduction
σ (2Δ) = 5.6 msec-> stable reproduction of 63% or more

Further, when the probability distribution of jitter is not known and the standard deviation σ is known (for example, when σ = 8 msec), the following occurs.
m (average value) = 2Δ = 5.6 msec
m + 4σ = 5.6 msec + 4 × 8 msec = 37.6 msec
-> 93.8% or more of stable reproduction m + 3σ = 29.6 msec-> 88.8% or more of stable reproduction

  FIG. 5 is a functional configuration diagram of the packet receiving apparatus according to the present invention.

  According to FIG. 5, the packet reception device 1 includes a jitter average maximum value calculation unit 104, a reception buffer size determination unit 105, a jitter measurement unit 106, and a transmission time interval reception unit 107, as compared with FIG. Have. These functional units can also be realized by a program that causes a computer mounted on the packet receiving apparatus to function.

  The jitter average maximum value calculation unit 104 calculates twice the Δ (2Δ) as the jitter average maximum value based on the packet transmission time interval Δ. The average jitter maximum value is output to reception buffer size determination section 105.

  The reception buffer size determination unit 105 determines the reception buffer size based on the jitter average maximum value. Further, it is possible to determine the reception buffer size by expanding the relief range based on the jitter standard deviation σ.

The reception buffer size determination unit 105 sets the reception buffer size to 2Δ + 3σ when the probability distribution of jitter measured by the jitter measurement unit 106 follows the normal distribution of (m, σ ² ), and more than 99% of received packets Can be received and reproduced stably. Also, the reception buffer size determination unit 105 sets the reception buffer size to 3σ (6Δ) when the probability distribution of jitter measured by the jitter measurement unit 106 follows an exponential distribution of m (average value) = σ (standard deviation). It is also possible to stably receive and reproduce 95% or more of received packets. Further, when the standard deviation σ is known, the reception buffer size determining unit 105 can set the reception buffer size to 2Δ + 4σ and stably receive and reproduce 93.8% or more of received packets.

  Note that when the combination of the video encoding bit rate and the transmission packet size is one type, the reception buffer size of the packet reception device may be fixed. In addition, when there are a plurality of limited combinations of video encoding bit rates and transmission packet sizes, a plurality of reception buffer sizes corresponding to these combinations are prepared, and the reception buffer size is selected according to the communication state. You can also.

  The jitter measuring unit 106 measures jitter and derives a statistical probability distribution composed of the jitter and the number of times (frequency). For example, the jitter measuring unit 106 creates a histogram as shown in FIG. Thereby, it can be detected whether the probability distribution of jitter follows a normal distribution or an exponential distribution. Information on “normal distribution” and “exponential distribution” is output to the reception buffer size determination unit 105.

  The transmission time interval receiving unit 107 receives the packet transmission time interval Δ from the packet transmission device 2. The packet transmitting device calculates a transmission time interval time Δ [ms] of the transmission packet calculated from the video encoding bit rate and the transmission packet size, and transmits the transmission time interval time Δ to the packet receiving device.

  The transmission time interval receiving unit 107 outputs the received packet transmission time interval Δ to the jitter average maximum value calculating unit 104. Also, the transmission time interval receiving unit 107 can receive the changed packet transmission time interval Δ when the video encoding bit rate and the transmission packet size are changed during transmission. The received packet transmission time interval Δ is output to the jitter average maximum value calculation unit 104, and the reception buffer size can be reconfigured.

  As described above in detail, according to the packet reception device, method and program of the present invention, the jitter average maximum value is calculated based only on the packet transmission time interval Δ in the packet transmission device, and the jitter average maximum value is calculated. It is possible to determine a necessary and sufficient reception buffer size to cover.

  According to the various embodiments of the present invention described above, those skilled in the art can easily make various changes, modifications and omissions within the scope of the technical idea and the viewpoint of the present invention. The above description is merely an example, and is not intended to be restrictive. The invention is limited only as defined in the following claims and the equivalents thereto.

It is a function block diagram of the packet receiver in a prior art. It is explanatory drawing of a jitter. It is a model sequence diagram of jitter in the present invention. It is a histogram of jitter versus frequency. It is a functional block diagram of the packet receiver in this invention.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Packet receiver 101 Reception buffer part 102 Packet analysis part 103 Synchronization control part 104 Jitter average maximum value calculation part 105 Reception buffer size determination part 106 Jitter measurement part 107 Transmission time interval reception part 111 Video decoder part 112 Video buffer part 113 Display part 121 audio decoder unit 122 audio buffer unit 123 speaker unit 2 packet transmission device

Claims (11)

Designed so that packet loss and order reversal do not occur and connected to a network that has a tendency to repeat changes such that the transmission delay begins to decrease after the transmission delay increases by continuing to transmit the packet, In a packet reception device having a reception buffer for receiving packets via
A packet transmission time interval receiving means for receiving a packet transmission time interval Δ from the packet transmission device;
Jitter average maximum value calculating means for calculating twice (2Δ) the packet transmission time interval Δ as the maximum value of the jitter average;
And a reception buffer size determining means for determining a reception buffer size based on the maximum value of the jitter average. The jitter average value in the jitter average maximum value calculating means is based on the following model formula:
Do: Transmission delay increase for each packet
Δ: Transmission delay decrease for each packet (= packet transmission time interval)
M / Do: Number of transmitted packets with increased delay at maximum transmission delay
M / Do × Do: Maximum transmission delay with increased delay
M / Δ: the number of transmitted packets with reduced delay at the maximum transmission delay
M / Δ × Δ: Maximum transmission delay in delay reduction

(Molecular) M / Do × Do + M / Δ × Δ: total transmission delay of delay increase and delay decrease
(Denominator) M / Do + M / Δ: Total number of transmitted packets
Even when Do is increased to ∞, the maximum jitter average that can only be 2 × Δ is based on the following equation:

The packet receiving apparatus according to claim 1, wherein: Jitter measurement means for measuring the jitter and deriving a statistical probability distribution consisting of the jitter and its frequency (frequency),
The receive buffer size determination means, based on the maximum value and the jitter standard deviation of the jitter mean, packet receiving apparatus according to claim 1 or 2, characterized in that to determine the receive buffer size. The reception buffer size determining means sets the reception buffer size to 2Δ (2 of the packet transmission time interval Δ) when the probability distribution of the jitter measured by the jitter measurement means follows a normal distribution of (m, σ ² ). 4. The packet receiving apparatus according to claim 3, wherein the packet receiving apparatus is set based on ( multiple) + 3σ and stably receives and reproduces 99% or more of received packets. The reception buffer size determination means sets the reception buffer size to 3σ (6Δ: when the probability distribution of the jitter measured by the jitter measurement means follows an exponential distribution of m (average value) = σ (standard deviation). 4. The packet receiver according to claim 3, wherein the packet receiver is set based on a packet transmission time interval [Delta] of 6) and receives and reproduces 95% or more of received packets stably. The reception buffer size determining means sets the reception buffer size based on 2Δ + 4σ when the standard deviation σ is known, and stably receives and reproduces a reception packet of 93.8% or more. 3. The packet receiving device according to 1 or 2 .   The packet transmission time interval receiving means receives the changed packet transmission time interval Δ when the video encoding bit rate and the transmission packet size are changed during transmission, and sets the packet transmission time interval Δ to The packet receiving apparatus according to claim 1, wherein the packet receiving apparatus outputs to the jitter average maximum value calculating means. Designed so that packet loss and order reversal do not occur and connected to a network that has a tendency to repeat changes such that the transmission delay begins to decrease after the transmission delay increases by continuing to transmit the packet, In a method of determining a reception buffer size in a packet reception device having a reception buffer for receiving a packet via
The network has a tendency that the transmission delay increases as the packet continues to be transmitted, and after reaching the maximum transmission delay, the network has a tendency to repeat such a change that the transmission delay starts to decrease,
A first step of receiving a packet transmission time interval Δ from the packet transmission device;
A second step of calculating twice (2Δ) the packet transmission time interval Δ as the maximum value of the jitter average ;
Receive buffer size determination method characterized by have a <br/> a third step of determining a receive buffer size based on the maximum value of the jitter mean. The jitter average value in the second step is based on the following model equation:
Do: Transmission delay increase for each packet
Δ: Transmission delay decrease for each packet (= packet transmission time interval)
M / Do: Number of transmitted packets with increased delay at maximum transmission delay
M / Do × Do: Maximum transmission delay with increased delay
M / Δ: the number of transmitted packets with reduced delay at the maximum transmission delay
M / Δ × Δ: Maximum transmission delay in delay reduction

(Molecular) M / Do × Do + M / Δ × Δ: total transmission delay of delay increase and delay decrease
(Denominator) M / Do + M / Δ: Total number of transmitted packets
Even when Do is increased to ∞, the maximum jitter average that can only be 2 × Δ is based on the following equation:

The reception buffer size determination method according to claim 8. Designed so that packet loss and order reversal do not occur and connected to a network that has a tendency to repeat changes such that the transmission delay begins to decrease after the transmission delay increases by continuing to transmit the packet, In a program for causing a computer mounted on a packet receiving apparatus having a reception buffer to receive a packet via
The network has a tendency that the transmission delay increases as the packet continues to be transmitted, and after reaching the maximum transmission delay, the network has a tendency to repeat such a change that the transmission delay starts to decrease,
A packet transmission time interval receiving means for receiving a packet transmission time interval Δ from the packet transmission device;
Jitter average maximum value calculating means for calculating twice (2Δ) the packet transmission time interval Δ as the maximum value of the jitter average;
A program for a packet receiving apparatus , comprising: a reception buffer size determining unit that determines a reception buffer size based on a maximum value of the jitter average. The jitter average value in the jitter average maximum value calculating means is based on the following model formula:
Do: Transmission delay increase for each packet
Δ: Transmission delay decrease for each packet (= packet transmission time interval)
M / Do: Number of transmitted packets with increased delay at maximum transmission delay
M / Do × Do: Maximum transmission delay with increased delay
M / Δ: the number of transmitted packets with reduced delay at the maximum transmission delay
M / Δ × Δ: Maximum transmission delay in delay reduction

(Molecular) M / Do × Do + M / Δ × Δ: total transmission delay of delay increase and delay decrease
(Denominator) M / Do + M / Δ: Total number of transmitted packets
Even when Do is increased to ∞, the maximum jitter average that can only be 2 × Δ is based on the following equation:

The program for a packet receiving device according to claim 10 , wherein the computer functions as described above .

Images