JP5743350B2 - Data transmission apparatus, forward error correction method, and program - Google Patents

Description

  The present invention relates to a data transmission device, a forward error correction method, and a program.

  In recent years, with the broadbandization of communication networks, the use of IPTV services that distribute high-quality video content over an IP network is increasing. In the IPTV service, since a large number of streaming transmissions are simultaneously performed in order to realize a multi-program, it is desirable to improve the efficiency of the transmission itself, and a multicast technique is adopted. By adopting multicast technology, it is possible to prevent a plurality of identical programs from flowing on the same route, leading to efficient use of the network.

  In recent years, in IPTV services, it has been urgent to be able to provide high-quality video content to various terminals such as mobile terminals, tablet terminals, PCs, and high-resolution 4K televisions. However, in an IP network, it is difficult to completely guarantee transmission quality, and IP packet loss leads to deterioration in viewing quality. As a function for preventing quality degradation due to packet loss, there are a method of retransmitting a lost packet, a method of transmitting a redundant packet in advance and correcting an error on the receiving side (forward error correction), and the like. Since forward error correction (FEC) does not involve retransmission of a packet, application to real-time video content distribution is expected.

  When error correction is performed in units of packets, an FEC encoded block (hereinafter referred to as a block) is configured by a plurality of media packets (video, etc.) on the transmission side, and a redundant code is generated for each block. Send out as FEC packet. When the packet is lost, the lost packet is recovered by using the FEC packet by a calculation on the receiving side alone.

  Since adding excessive redundant data may hinder the efficiency of network use, it is desirable to adaptively change it according to the error occurrence frequency of the communication path and the allowable delay time. As a method for adaptively selecting an optimal encoding parameter, a method for acquiring network quality information in real time and adaptively changing the encoding parameter in accordance with the quality information has been proposed (see Patent Document 1). ).

JP 2007-74152 A

  However, when the conventional technology is applied to a service that distributes a stream using multicast, several problems occur as described below. In the following, it is assumed that the data receiving device that receives the service is a smartphone or a tablet.

  First, the basic operation of the conventional forward error correction system will be described with reference to FIG. As shown in this figure, the data receiving apparatus measures the burstiness value of the packet loss and the packet loss rate as the network quality information, and transmits them to the data transmitting apparatus (steps S101 → S102). The data transmitting apparatus determines FEC parameters (hereinafter referred to as parameters) such as code length and redundancy based on the burstiness value and the packet loss rate received from the data receiving apparatus, performs encoding based on the parameters, The forward error correction information (hereinafter referred to as FEC information) for decoding including the parameters is packetized and transmitted to the data receiving apparatus (steps S103 → S104 → S105 → S106). The data receiving apparatus decodes the data of the encoded block using the FEC information received from the data transmitting apparatus (step S107).

  In such a case, in step S101, quality information is measured for a long time during multicast reception, leading to an increase in resource consumption. In particular, smartphones, tablets, and the like have a limited amount of data that can be stored compared to a PC. In step S102, since the wireless section is half-duplex communication, transmitting quality information during multicast reception may lead to packet loss of multicast reception data. Furthermore, in step S103, quality information is uploaded from a large number of data receiving apparatuses, which hinders network efficiency and leads to an increase in equipment cost for accumulating quality information. In addition, it is necessary to determine the parameters based on the operational evaluation of whether or not the operational target is passed and the performance evaluation of the FEC parameters themselves. In addition, burst loss due to network switching work, network equipment failure, and the like is likely not to be compensated by FEC, and may add excessive redundant data. Further, in step S105, it is not sufficient to determine the optimum parameter for one data receiving apparatus. That is, it is necessary to determine parameters that are optimal for a large number of data receiving apparatuses that may receive the same stream.

  As described above, when the conventional technology is applied to a service for distributing a stream using multicast, several problems occur. Even when multicast is used, it is desirable to determine optimum parameters for encoding from network quality information and perform forward error correction while adaptively changing the parameters, but such a method has been established. The current situation is not.

  An object of the present invention is to provide a data transmission apparatus, a forward error correction method, and a program capable of determining an optimum parameter even when multicast is used.

In order to achieve the above object, the invention according to the first aspect is a data transmitting apparatus that transmits forward error correction information corresponding to media data, and collects quality information indicating communication quality from a plurality of data receiving apparatuses. Based on the quality information collected by the quality information collection unit and the quality information collected by the quality information collection unit, a section in which a loss equal to or greater than a predetermined threshold commonly seen in the same time zone occurs is excluded. A parameter determination unit that determines parameters for encoding, and error correction that generates forward error correction information based on the parameters determined by the parameter determination unit and transmits the forward error correction information to the data reception device The gist is to include an encoding unit.

The invention according to a second aspect, means in the invention according to the first aspect, as the quality information, the packet of the media data that there is loss loss pattern of packet indicating where to have lost in the encoding block The gist of the present invention is to use a loss list represented by a bit string obtained by concatenating a bit string indicating that there is no packet loss and a bit string corresponding to a position in the coding block .
The invention according to the third aspect is that, in the invention according to the first or second aspect, the media data is streamed using multicast, and the plurality of data receiving apparatuses receive the same stream. To do.

The invention according to the fourth aspect is a forward error correction method for correcting a code error on the data receiving device side by transmitting forward error correction information corresponding to media data from the data transmitting device to the data receiving device, A quality information collecting step in which a data transmitting device collects quality information indicating communication quality from a plurality of data receiving devices, and a quality information collected in the quality information collecting step, which is commonly found in the same time zone A parameter determining step for determining a parameter for encoding based on the remaining quality information , excluding a section in which a loss exceeding the threshold occurs, and forward error correction based on the parameter determined in the parameter determining step Error correction coding step of generating information and transmitting the forward error correction information to the data receiving device. The the gist.

The gist of the invention according to the fifth aspect is a program for causing a computer to function as the data transmission device according to any one of the first to third aspects.

  According to the present invention, it is possible to provide a data transmission apparatus, a forward error correction method, and a program that can determine an optimum parameter even when multicast is used.

It is a block diagram of the forward error correction system in embodiment of this invention. It is a figure for demonstrating the conventional forward error correction system. It is a figure for demonstrating the forward error correction system in embodiment of this invention. It is a figure for demonstrating the lottery probability in embodiment of this invention. It is a figure for demonstrating the problem of the determination process of the conventional parameter. It is a figure for demonstrating the determination process of the parameter in embodiment of this invention. It is a figure which shows the relationship between redundancy and recovery. It is a figure for demonstrating the conventional quality measurement unit. It is a figure for demonstrating the quality measurement unit in embodiment of this invention. It is a flowchart which shows operation | movement of the control-information acquisition part in embodiment of this invention. It is a flowchart which shows operation | movement of the quality measurement part in embodiment of this invention. It is a flowchart which shows operation | movement of the quality measurement part in embodiment of this invention. It is a figure which shows the quality measurement image in embodiment of this invention. It is a flowchart which shows operation | movement of the quality measurement part in embodiment of this invention. It is a flowchart which shows operation | movement of the error correction decoding part in embodiment of this invention. It is a flowchart which shows operation | movement of the quality information collection part in embodiment of this invention. It is a flowchart which shows operation | movement of the parameter determination part in embodiment of this invention. It is a figure which shows a burst property and target BER determination result (primary determination result). It is a figure which shows the result (final determination result) which compared the primary determination result with the service condition value. It is a flowchart which shows operation | movement of the error correction encoding part in embodiment of this invention. It is a flowchart which shows operation | movement of the control information transmission part in embodiment of this invention. It is a figure which shows the example of a loss list in embodiment of this invention. It is a figure which shows another example of a loss list in embodiment of this invention. It is a flowchart which shows the basic operation | movement of the conventional forward error correction system.

  Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.

(System configuration)
FIG. 1 is a configuration diagram of a forward error correction system in an embodiment of the present invention. As shown in this figure, the data transmitting device 10 and the data receiving device 20 are connected via a communication network 30. The data transmitting apparatus 10 is a server or the like that receives media data, generates FEC information (forward error correction information) corresponding to the received media data, and transmits the generated FEC information. On the other hand, the data reception device 20 is various terminals such as a smartphone, a tablet, and a PC that receive FEC information, correct a code error of media data based on the received FEC information, and output repaired media data.

  The basic operation of the forward error correction system in the embodiment of the present invention is the same as in FIG. Although not mentioned in the description of FIG. 24, media data to the data reception device 20 may be transmitted from a different device different from the data transmission device 10 or may be transmitted from the data transmission device 10. In this embodiment, it is assumed that the media data to the data receiving device 20 is transmitted from a different device different from the data transmitting device 10. The data receiving apparatus 20 receives media packets and FEC packets. Of course, the unit of media data does not necessarily have to be a packet, so it may be referred to as “media data” instead of “media packet”.

  Here, the data transmission device 10 includes a control information transmission unit 11, a quality information collection unit 12, a parameter determination unit 13, and an error correction coding unit 14. The control information transmission unit 11 generates control information for controlling the quality measurement process and transmits the control information to the data receiving device 20. The quality information collection unit 12 collects quality information indicating communication quality from the plurality of data reception devices 20. As the quality information, for example, it is possible to use a loss list in which a bit pattern represents a loss pattern indicating the number of consecutive missing packets of media data and where in the encoded block it is lost. The parameter determination unit 13 determines a parameter (FEC parameter) for encoding based on quality information that satisfies a predetermined condition among the quality information collected by the quality information collection unit 12. This parameter is, for example, code length, redundancy, and the like. The error correction encoding unit 14 performs encoding based on the parameter determined by the parameter determination unit 13, packetizes FEC information for decoding including the parameter, and transmits the packetized data to the data reception device 20. The FEC information may include a column weight, the number of subpacket divisions, a coding block number, a sequence number, and the like.

  The header of the FEC packet transmitted from the data transmitting apparatus 10 to the data receiving apparatus 20 includes an FEC parameter and related information (encoded block number, sequence number, column weight, subpacket division number, etc.) related to the parameter. The information generated by encoding is stored in the payload of the FEC packet. The control information may be stored in the header of the FEC packet, or may be stored in a control information packet different from the FEC packet. A part of the control information may be set in the data receiving device 20 in advance.

  On the other hand, the data reception device 20 includes a control information acquisition unit 21, a quality measurement unit 22, and an error correction decoding unit 23. The control information acquisition unit 21 acquires control information from the data transmission device 10. The quality measurement unit 22 performs a quality measurement process based on the control information acquired by the control information acquisition unit 21. The error correction decoding unit 23 corrects the code error of the media data based on the FEC information stored in the FEC packet, and outputs the repaired media data. Specifically, the error correction decoding unit 23 uses the FEC packet to repair a missing packet in the block, and outputs the repaired data.

  In FIG. 1, the data transmission device 10 includes the control information transmission unit 11, the quality information collection unit 12, the parameter determination unit 13, and the error correction coding unit 14, but is not limited thereto. . For example, the quality information collection unit 12 and the parameter determination unit 13 can be provided as a set in a separate device independent of the data transmission device 10. Even in this case, the relationship between the processing units is the same as in FIG.

(Unicast)
FIG. 2 is a diagram for explaining a conventional forward error correction system. Here, it is premised on a service for delivering a stream from the data transmitting apparatus 10 to each of the data receiving apparatuses 20A to 20Z using unicast. The data transmitting apparatus 10 distributes initial parameters to the data receiving apparatuses 20A to 20Z, and then acquires quality information from the data receiving apparatuses 20A to 20Z, and each data receiving apparatus 20A to 20Z is based on the quality information. The optimal parameters are distributed.

(Multicast)
FIG. 3 is a diagram for explaining the forward error correction system in the embodiment of the present invention. Here, it is assumed that the data transmission device 10 delivers a stream to each of the data reception devices 20A to 20Z using multicast. The data transmitting apparatus 10 distributes initial parameters to the data receiving apparatuses 20A to 20Z, and then acquires quality information from the data receiving apparatuses 20A to 20Z, and based on the quality information, the data transmitting apparatuses 20A to 20Z The optimal parameters are distributed.

  Specifically, the data transmitting apparatus 10 collects quality information from each of the data receiving apparatuses 20A to 20Z that have received the multicast, and determines an optimum parameter according to each same stream / usage matrix. As a result, it is possible to determine the most suitable encoding parameters for each of the data reception devices 20A to 20Z that have received the multicast.

  In addition, the data transmission device 10 includes a lottery probability in a part of data at the time of multicast distribution. The lottery probability is one piece of control information and has a very small amount of data. The data reception devices 20A to 20Z are targeted for quality measurement / quality information transmission probabilistically according to the probability specified by the lottery probability. As a result, it is possible to suppress control associated with quality information measurement and communication associated with transmission of quality information.

  Furthermore, the data transmission device 10 adaptively changes the lottery probability according to all the data reception devices 20A to 20Z that receive the service. Thereby, the quality information transmitted from the data receiving devices 20A to 20Z to the data transmitting device 10 can be suppressed to the minimum necessary.

(Lottery probability)
FIG. 4 is a diagram for explaining the lottery probability in the embodiment of the present invention. The value of the lottery probability of quality measurement is determined by the throughput of the data transmitting apparatus 10 that collects quality information, the number of channels (ch), the number of viewing users, and the distribution end time in each channel, as in the following equation.

Lottery probability = Quality information throughput per unit time on the server
÷ Number of channels ending at the same time as the corresponding channel (including own channel)
÷ Number of users viewing the channel
× 100
For example, the throughput of quality information reception per unit time in the data transmission device 10 is 90 cases / second (the size of quality information sent at one time from one of the data reception devices 20A to 20Z is 61 KB, the data transmission device 10 Is assumed to be 5.5 MB / s). In this case, when the number of channels is 5, as shown in FIG. 4, the lottery probability set for each channel is 90 ÷ 2 ÷ 500 × 100 = 9% for 1ch, 90 ÷ 2 ÷ 1000 × 2ch. 100 = 4.5%, 3ch 90 ÷ 2 ÷ 1000 × 100 = 4.5%, 4ch 90 ÷ 2 ÷ 2000 × 100 = 2.25%, 5ch 90 ÷ 1 ÷ 5000 × 100 = 1.8 %.

(Conventional parameter determination process)
A general FEC coding parameter determination process is as follows. That is, when the burstiness value is larger than a predetermined burstiness value (threshold value), the encoded block length is increased. On the other hand, if the burstiness value is smaller than the threshold value, the encoded block length is reduced. Further, when the loss rate value is larger than a predetermined loss rate value (threshold value), the redundancy is increased. On the other hand, if the loss rate value is smaller than the threshold value, the redundancy is reduced.

  FIG. 5 is a diagram for explaining the problem of the conventional parameter determination process. The horizontal axis represents time t, and the vertical axis represents quality information (the number of consecutive missing data packets). As shown in this figure, there is quality information Q1 including a burst loss period that cannot be repaired by FEC. Conventionally, analysis has been performed based on quality information including a section having strong burst characteristics. For this reason, there is a problem in that the redundancy is excessively given, and there is a problem in that the analysis result is inclined in a direction in which the encoding block length is set to be excessively long (a direction in which the delay time is excessively extended).

(Parameter determination process in the embodiment of the present invention)
FIG. 6 is a diagram for explaining a parameter determination process in the embodiment of the present invention. 6 (a) and 6 (b), the horizontal axis indicates time t, and the vertical axis indicates quality information (the number of consecutive missing media data packets). Moreover, the horizontal axis of FIG.6 (c) shows the number of block continuous defect | deletions, and the vertical axis | shaft has shown the number of blocks. In the figure, reference symbol T1 indicates a threshold for excluding quality information, and reference symbol T2 indicates a guaranteeable threshold by FEC.

  As shown in FIGS. 6 (a) and 6 (b), the data transmitting device 10 has a burst loss equal to or higher than a threshold T1 commonly found in the same time zone in the quality information collected from the data receiving devices 20A to 20Z. The section is excluded from the analysis target section, and the optimum parameter is determined from the remaining quality information. As described above, if the quality information of the section equal to or higher than the threshold value T1 is not used for the analysis for parameter optimization, an appropriate redundancy and coding block length are given as shown in FIG. 6 (c). be able to. In other words, it is possible to avoid giving an excessive redundancy or encoded block length based on parameters determined based on quality information including burst loss that cannot be repaired by FEC.

  In addition, the data reception devices 20A to 20Z are configured to send the quality information after the multicast reception is completed. Thereby, it is possible to avoid a collision between communication at the time of multicast reception and communication at the time of transmission of quality information.

(Conventional quality measurement unit)
FIG. 7 is a diagram illustrating the relationship between redundancy and recovery. FIGS. 7A and 7B are examples in which eight media packets are transmitted. A, B, ... indicate FEC packets, 1, 2, ... indicate media packets. The redundancy amount is the same in both FIGS. 7A and 7B (redundancy 20%), but the longer block length (in this example, FIG. 7B) can be restored depending on the loss location. There is.

  FIG. 8 is a diagram for explaining a conventional quality measurement unit. Here, it is assumed that the quality measurement unit is a predetermined number of packets (eight in this example). The unit of the FEC encoded block and the quality information acquisition unit do not match. Whether the redundancy is increased or decreased is determined from the ratio of the number of lost packets to the total number of packets to be received at the quality information measurement interval.

[Loss rate] (excluding FEC) = 2/8 = 25% (first quality measurement)
When the service quality target is BER (Block Error Rate) = 10 ^ -5 (general target value), it is determined whether the target has been achieved from the quality information for 10 ^ 5 blocks, and the necessary redundancy is determined. There is a need. It is not preferable to optimize the redundancy based only on the result of the loss rate. In addition, depending on the FEC method, the number of source packets in a coding block is fluid with the variable length correspondence of the packet size, and therefore the total number of packets to be received is not always constant in the quality measurement unit time. For this reason, erroneous determination of appropriate redundancy and delay from the loss rate and burst value may lead to a reduction in repair capability, provision of extra redundancy, and extension of delay time.

(Quality measurement unit in the embodiment of the present invention)
FIG. 9 is a diagram for explaining quality measurement units in the embodiment of the present invention. As shown in this figure, the quality measurement unit is synchronized with the unit of the encoded block, and the loss rate and burst property for each encoded block are acquired as quality information. As the loss rate, the average loss rate for each coding block during the period specified by the control information is obtained as in the following equation.

Loss rate = (0/4 + 2/4 + 1/3 + 1/3 + ...) / N = 29% (excluding FEC)
It is determined whether or not the target BER has been achieved from the quality information accumulated for the total number of blocks equal to or greater than the number of blocks that can be confirmed by the target BER (operational evaluation). Even if the total block is not acquired from one data receiving device 20, the quality information collected from the plurality of data receiving devices 20 may be the total block (10 ^ 5).

  As the loss list, a loss list indicating not only the number of consecutive defects but also a loss pattern indicating where the loss is in the encoded block in a bit string is acquired. In the following example, “01” indicates no defect and “00” indicates a defect.

01,00,00,00,01, 1 block length 5, number of sources 5, 20120815110900, ...
01,01,00,00,00, 1 block length 5, 5 sources, 20120815110901, ...
01,01,01,00,01, 1 block length 5, number of sources 3, 20120815110902, ...
...
It is possible to apply not only the optimum redundancy as the FEC performance but also an appropriate redundancy according to the service quality target. In addition, for the FEC method in which whether error correction is possible or not is determined by the loss position in the block even with the same number of consecutive losses, the criterion for determining whether or not to change the block length has increased compared to the prior art, and the block length has become more appropriate It is possible to adapt.

  Thus, in the forward error correction system in the embodiment of the present invention, the quality information is acquired by synchronizing the quality measurement unit with the coding block. This evaluates whether the parameters are functioning properly even if the number of packets in the block fluctuates (if the number of lost packets in the block is up to X, it should be recovered, but the result is) Is possible.

  Further, in addition to the evaluation result described above, the redundancy is determined based on the determination result of whether or not the service quality target BER is reached. This makes it possible to optimize the required redundancy and block length according to the service quality target.

  Further, information indicating where and how many in the coding block are lost is acquired as a burstiness value. Accordingly, it is possible to determine a more optimal block length by determining the block length based on the burstiness value.

(Operation of control information acquisition unit)
FIG. 10 is a flowchart showing the operation of the control information acquisition unit 21 in the embodiment of the present invention. First, when an application linked with the apparatus calls an initialization function in response to activation of a viewing player (step S1), the control information acquisition unit 21 reads an opt-in flag and determines a value of the opt-in flag (step S2 → S2a). The opt-in flag indicates whether or not the user of the data receiving device 20 permits acquisition / transmission of quality information. If the opt-in flag is true, the subsequent processing is performed, and if it is false, the subsequent processing is not performed. Next, the control information initial value is read to determine whether or not the transmission date storage file exists (steps S3 → S4). Thereby, when there is a transmission date storage file, the transmission date storage file is read (step S5), and when there is no transmission date storage file, a transmission date storage file is generated (step S6). Thereafter, when multicast reception is started, it is determined whether or not to be a quality measurement target (steps S7 → S8).

  Hereinafter, step S8 for determining whether or not to be a quality measurement target will be described in more detail. That is, the control information acquisition unit 21 determines whether or not to perform quality measurement related to the received media stream from the control information packet transmitted from the server, and a probability corresponding to a target for transmitting quality information. “Quality measurement lottery probability (effective value, index)” is acquired. When all of the following conditions (1) to (4) are satisfied, the quality information is to be transmitted, the quality measurement unit 22 is called, and quality measurement is started. On the other hand, when any of the following conditions (1) to (4) is not satisfied, the quality information is not transmitted and the quality measurement is not performed.

(1) The opt-in flag of the input parameter is true: Agree (2) Whether the control information quality measurement is “Yes”
(3) Current date (system time)-last transmission date of quality information transmission date storage file ≧ quality information transmission freeze days of control information initial value (4) Winning lottery result Specifically, lottery is performed by the following formula.

0 ≦ lottery value <valid value lottery value = random numbers from 0 to 10 exponential power are generated, and the generated random number value is set as the lottery value. For example, if the effective number is 20 and the index is 3, a lottery value = 0 to a random number less than 10 to the third power (1000) is generated. And if 0 ≦ lottery value <20, the winning combination is made.

  Thereby, it is possible to prevent quality information from being sent from all the data receiving apparatuses 20, and to collect the minimum necessary quality information. In addition, it is possible to prevent the data receiving devices 20 that are the quality measurement targets from being biased, and to make all the data receiving devices 20 equally the quality measurement targets.

(Operation of the quality measurement unit)
FIG. 11 is a flowchart showing the operation of the quality measuring unit 22 in the embodiment of the present invention. First, the quality measurement unit 22 generates a loss list and acquires the number of error media packets before error correction (steps S11 → S12). Next, when error correction decoding processing is performed, the number of error media packets after error correction is acquired, and it is determined whether or not to accumulate quality information (steps S13 → S14 → S15). Thereby, when accumulating quality information, the quality information is accumulated, and when the multicast reception is completed, the quality information is transmitted (steps S15 → S16 → S17 → S18). On the other hand, when the quality information is not accumulated, the process returns to the loss list generation step (step S15 → S11).

  Hereinafter, step S15 for determining whether or not to store quality information will be described in more detail. That is, when the following (1) and (3) are satisfied, it is determined as the first block of the quality information acquisition block number, the current time is set as the previous quality information acquisition cycle start time (described later), and the number of acquired blocks ( After setting 0 to (described later), the process proceeds to the quality information accumulation step S16. On the other hand, when the following (1) and (2) are satisfied but (3) is not satisfied, the process proceeds to the quality information accumulation step S16 without doing anything. In other cases, the process proceeds to the loss list generation step S11.

(1) More than the quality information acquisition start time of the control information initial value has elapsed since the start of decoding.

(2) The number of acquisition times is less than the number of quality information acquisition blocks of the control information initial value.

(3) Quality information has never been accumulated, or the quality information acquisition cycle (time) of the control information initial value has elapsed since the previous quality information acquisition cycle start time.

  The previous quality information acquisition cycle start time is the time at which initial quality information acquisition was performed for each cycle. The previous quality information acquisition cycle start time is updated every quality information acquisition cycle of the control information initial value.

  The acquisition block count is the acquisition count of quality information for one block. The number of acquisition blocks is initialized every quality information acquisition cycle of the control information initial value.

  FIG. 12 is a flowchart showing the operation of the quality measuring unit 22 in the embodiment of the present invention. Here, the quality information accumulation (step S16) shown in FIG. 11 will be described in detail. First, the quality measuring unit 22 determines whether “accumulated size + acquired quality information” is less than [maximum accumulated size] of the control information (step S21). If the control information is less than the [maximum accumulation size], it is determined whether the accumulation count is less than [control information [number of quality information acquisition cycles] × [number of quality information acquisition blocks]] (step S22). As a result, if it is less than “[quality information acquisition cycle count] × [quality information acquisition block count] of control information”, the quality information is accumulated, 1 is added to the accumulation count, and the process returns to the caller ( Step S23).

  FIG. 13 is a diagram showing a quality measurement image in the embodiment of the present invention. Here, the quality information acquisition start time is set to 5 seconds, the quality information acquisition block count is set to 2 blocks, the quality information acquisition cycle is set to 10 sec, and the quality information acquisition cycle count is set to 3. Thereby, it is possible to prevent quality information in the same reception elapsed time zone during multicast reception from being acquired and collected. Each setting value of such setting items may be stored in a control information packet and distributed from the data transmitting apparatus 10 to the data receiving apparatus 20, or may be set in the data receiving apparatus 20 in advance. It may be.

  FIG. 14 is a flowchart showing the operation of the quality measuring unit 22 in the embodiment of the present invention. Here, the quality information transmission (step S18) shown in FIG. 11 will be described in detail. First, the quality measurement unit 22 determines whether or not to transmit quality information, and determines the communication protocol when transmitting (step S31 → S32). If the communication protocol is https, the certificate file is read, quality information is transmitted, and the communication result is determined (steps S33 → S34 → S35). On the other hand, if the communication protocol is http or the certificate file does not exist, the quality information is transmitted without reading the certificate file, and the communication result is determined (step S34 → S35). Thereby, when a normal response is received in the communication result determination, the transmission date storage file is updated and the process returns to the caller (step S36).

(Operation of error correction decoder)
FIG. 15 is a flowchart showing the operation of the error correction decoding unit 23 in the embodiment of the present invention. The error correction decoding unit 23 performs FEC decoding processing based on the parameter determined by the parameter determination unit 13 and outputs the repaired media data (step S41). This operation is repeated until there is no input data or there is a stop request (step S42).

(Operation of the quality information collection unit)
FIG. 16 is a flowchart showing the operation of the quality information collection unit 12 in the embodiment of the present invention. When the quality information is received, the quality information collecting unit 12 performs data calculation and registers data indicating the calculation result (steps S51 → S52 → S53). This operation is repeated until no quality information is received.

(Operation of parameter determination unit)
FIG. 17 is a flowchart showing the operation of the parameter determination unit 13 in the embodiment of the present invention. First, the parameter determination unit 13 acquires quality information, determines a non-evaluation section, and performs data calculation for the non-evaluation section (steps S61 → S62 → S63). Next, burstiness / target BER determination (primary determination) is performed, and service condition suitability determination (final determination) is performed (steps S64 → S65). Finally, parameters are determined based on the result of the service condition suitability determination (final determination) (step S66).

  FIG. 18 is a diagram illustrating a burst property / target BER determination result (primary determination result). As shown in this figure, when the burst property is smaller than the threshold value and the target BER determination greatly exceeds the appropriate value (1-1), the code length is decreased and the redundancy is decreased. Further, when the burst property is the same as the threshold value and the target BER determination greatly exceeds the appropriate value (1-2), the redundancy is reduced. When the burst property is larger than the threshold value and the target BER determination greatly exceeds the appropriate value (1-3), the code length is increased and the redundancy is decreased. When the burst property is smaller than the threshold value and the target BER determination is within the appropriate range (2-1), the code length is decreased. When the burst property is the same as the threshold value and the target BER determination is within the appropriate range (2-2), the code length is not changed and the redundancy is not changed. If the burstiness is greater than the threshold and the target BER determination is within the appropriate range (2-3), the code length is increased. When the burst property is smaller than the threshold value and the target BER determination is significantly below the appropriate value (3-1), the code length is reduced and the redundancy is increased. Further, when the burst property is the same as the threshold value and the target BER determination is significantly below the appropriate value (3-2), the redundancy is increased. Further, when the burst property is larger than the threshold value and the target BER determination is significantly below the appropriate value (3-3), the code length is increased and the redundancy is increased.

  FIG. 19 is a diagram illustrating a result (final determination result) obtained by comparing the primary determination result with the service condition value. As shown in this figure, when the primary determination result of the code length is larger than the service condition value and the primary determination result of the redundancy is also larger than the service condition value (A-1), the code length is not changed and the redundancy is also changed. Not going to change. When the primary determination result of the code length is less than or equal to the service condition value and the primary determination result of redundancy is greater than the service condition value (A-2), the code length is increased to the service condition value upper limit. When the primary determination result of the code length is larger than the service condition value and the primary determination result of the redundancy is less than or equal to the service condition value (B-1), the redundancy is increased to the service condition value upper limit. Further, when the primary determination result of the code length is equal to or less than the service condition value and the primary determination result of the redundancy is also equal to or less than the service condition value (B-2), the parameter of the primary determination result is adopted.

  Thus, if the primary determination and the final determination are combined, a more appropriate code length and redundancy can be set. For example, even if the determination result of 1-3 is received, in the case of a service condition in which a delay due to an increase in the code length is not permitted, the code length is maintained, and the redundancy can be increased.

(Operation of error correction encoder)
FIG. 20 is a flowchart showing the operation of the error correction coding unit 14 in the embodiment of the present invention. First, the error correction coding unit 14 performs FEC coding processing based on the parameter determined by the parameter determination unit 13 (step S71). As a result of this encoding process, FEC information is stored in the payload of the FEC packet, and information such as used parameters is stored in the header of the FEC packet (step S72). Finally, this FEC packet is transmitted to the data receiving device 20 (step S73). This operation is repeated until there is no input data or there is a stop request (step S74).

(Operation of control information transmitter)
FIG. 21 is a flowchart showing the operation of the control information transmission unit 11 in the embodiment of the present invention. First, the control information transmission unit 11 sets a transmission control information setting value for each distribution stream (ch) (step S81). The transmission control information setting value is specifically a value indicating whether or not quality measurement is performed and a value indicating the quality measurement lottery probability. Next, the transmission control information setting value is read when the FEC generation process is started or when the control information is changed (step S82). Finally, a value obtained by converting the read information is stored in the control information packet, and multicast transmission is performed (step S83).

(Example of loss list)
FIG. 22 is a diagram showing an example of a loss list in the embodiment of the present invention. Here, the case where the packet size of the input data is different depending on the packet is assumed, and the FEC method for performing padding so that the input packet has a specific packet size is illustrated. “01” means no loss and “00” means that there is a loss. In this case, the loss list is “00010001” when the loss status is indicated by a bit string for each packet unit.

  FIG. 23 is a diagram showing another loss list example according to the embodiment of the present invention. Here, an FEC method is illustrated in which an input packet is divided into a specific packet size (GM) and processed. “01” means no loss, “00” means loss, and “11” means a packet delimiter. In this case, if the loss status is indicated by a bit string in GM units, the loss list is “00000000110101010111000000000000000011”.

  The bit character string of the media packet for one block and the bit character string of the FEC packet payload for one block are combined, and finally a character string of “10” is added so that the data length is a multiple of four. In this example, one packet or 1GM is represented by a 2-bit bit string, and is converted to a hexadecimal number for transmission, so that it is complemented with 10 so as to be a multiple of 4. Of course, any number of bits may be arbitrarily determined.

  As described above, according to the forward error correction system in the embodiment of the present invention, an optimum parameter can be determined even when multicast is used. That is, when determining parameters for generating FEC information according to communication quality (packet loss status, etc.), excluding quality information for sections where loss has occurred that cannot be recovered by forward error correction processing. Since the parameters are determined, it is possible to avoid excessive packet generation.

  Such a forward error correction system is particularly useful when the data receiving device 20 that receives the service is a smartphone, a tablet, or the like. In other words, since a smartphone, a tablet, or the like is used in a wireless environment and has fewer resources than a PC, the present system capable of avoiding excessive packet generation can be said to be a system having a very high practical value.

  The present invention can be realized not only as a forward error correction system (data transmission device 10 and data reception device 20) but also as a forward error with a characteristic processing unit provided in such a forward error correction system as a step. This can be realized as a correction method. Specifically, the forward error correction method according to the embodiment of the present invention corrects a code error on the data receiving apparatus 20 side by transmitting FEC information corresponding to media data from the data transmitting apparatus 10 to the data receiving apparatus 20. This is a forward error correction method. In such a forward error correction method, the data transmitting device 10 collects quality information indicating communication quality from a plurality of data receiving devices 20 and the quality information collected in the quality information collecting step in advance. A parameter determination step for determining a parameter for encoding based on quality information that satisfies a predetermined condition, FEC information is generated based on the parameter determined in the parameter determination step, and the FEC information is transmitted to the data reception device 20. The correction encoding step to be transmitted is executed.

  In addition, each step of the forward error correction method can be realized as a program that causes a computer to execute the steps. It goes without saying that such a program can be stored in the data transmission device 10 or the data reception device 20 and can be distributed via a recording medium such as a CD-ROM or a transmission medium such as a network. Nor.

DESCRIPTION OF SYMBOLS 10 ... Data transmitter 11 ... Control information transmission part 12 ... Quality information collection part 13 ... Parameter determination part 14 ... Error correction encoding part 20 ... Data reception apparatus 21 ... Control information acquisition part 22 ... Quality measurement part 23 ... Error correction decoding Chemical department

Claims (5)

A data transmission device for transmitting forward error correction information corresponding to media data,
A quality information collecting unit for collecting quality information indicating communication quality from a plurality of data receiving devices;
In the quality information collected by the quality information collection unit, a section in which a loss of a predetermined threshold value or more commonly seen in the same time zone is generated is excluded, and encoding is performed based on the remaining quality information. A parameter determination unit for determining parameters;
A data transmission device comprising: an error correction encoding unit that generates forward error correction information based on the parameter determined by the parameter determination unit and transmits the forward error correction information to the data reception device. As the quality information, a bit sequence packets of the media data means without loss of the bit string and a packet, which means there loss loss pattern packet indicating where to have lost in the encoding block in the coded block 2. The data transmission apparatus according to claim 1, wherein a loss list indicated by a bit string concatenated in correspondence with a position is used.   3. The data transmitting apparatus according to claim 1, wherein the media data is stream-distributed using multicast, and the plurality of data receiving apparatuses receive the same stream. A forward error correction method for correcting a code error on the data receiving device side by transmitting forward error correction information corresponding to media data from the data transmitting device to the data receiving device,
The data transmitting device is
A quality information collecting step for collecting quality information indicating communication quality from a plurality of data receiving devices;
In the quality information collected in the quality information collection step, a section where a loss of a predetermined threshold value or more commonly seen in the same time zone is excluded is excluded, and encoding is performed based on the remaining quality information. A parameter determination step for determining parameters;
Forward error correction information is generated based on the parameter determined in the parameter determination step, and error correction coding step of transmitting the forward error correction information to the data receiving device is performed. Method. The program for functioning a computer as a data transmitter of any one of Claim 1 to 3 .

Images