JP6220017B1 - Encoding apparatus, encoding method, and encoding program - Google Patents

Abstract

Data to which an error correction code is added is smoothed with a low delay and transmitted. An encoding device calculates parity data for each predetermined block, and sends out source data and parity data for one block in the same time as the time when source data for one block is input. As a result, the source data and the parity data can be received in a time corresponding to one block after the receiving apparatus receives the source data at the head of the block. Further, the input source data is temporarily held until the time necessary for transmitting the parity data of the previous block elapses, and the transmission of the source data is started after the transmission of the parity data of the previous block is completed. . Thereby, parity data is not transmitted in bursts, and the delay until the start of transmission of source data can be suppressed to an increase in parity data. [Selection] Figure 2

Description

  The present invention relates to a technique for encoding data.

  MMT (MPEG Media Transport) is a new media transmission technology that has been internationally standardized (Non-patent Document 3). In MPEG-2 TS, a unique reference clock called a system clock was used. In MMT, synchronous transmission from a plurality of devices or synchronous transmission using a plurality of transmission paths is performed by using a global coordinated universal time. Became possible. Accordingly, new uses such as a hybrid direction of broadcasting and communication are expected for MMT.

  On the other hand, IP communication is best-effort communication, and it is necessary to deal with data loss due to congestion of the IP communication network. When simultaneous delivery such as broadcasting is performed on an IP communication network, it is effective to provide tolerance against data loss by redundancy using a forward error correction code (FEC).

Tatsuya Fujii, Naohiro Kimura, Takahiro Yamaguchi, Narutetsu Ogawara, "Large-capacity 8K video transmission over a shared IP network using high-efficiency LDGM-FEC", Network Virtualization Timed Research Special Committee, IEICE, 2013 Takayuki Onishi, Mitsuro Ikeda, Jiro Naganuma, Yoshiyuki Yashima, "A Study on FEC Relay Method for Low Delay Video Communication", Information Science and Technology Forum General Proceedings, Information Processing Society of Japan, 2004, 3 (4), pp .87-88 “Information technology-High efficiency coding and media delivery in heterogeneous environments-Part 1: MPEG media transport (MMT)”, ISO / IEC 23008-1: 2014 “Information technology-High efficiency coding and media delivery in heterogeneous environments-Part 10: MPEG media transport forward error correction (FEC) codes”, ISO / IEC 23008-10: 2015

  In a method called a block code among forward error correction codes, parity data is generated for each block of a predetermined size. If the generated parity data is transmitted immediately, the delay until decoding on the receiving side can be reduced, but there is a problem that data is transmitted in bursts as shown in FIG.

  In order to smooth the data transmitted in a burst manner, for example, it is conceivable to transmit the parity data superimposed on the source block section to be transmitted next. In this method, if the source data is constant per time and the parity data to be added is also constant per time, a uniform transmission rate can be obtained. However, if smoothing is performed by this shaping method, as shown in FIG. 10, there is a problem in that the reception side cannot perform decoding until the reception of parity data is completed, resulting in a delay. In addition, for example, a shaping method is used with a shaping device sandwiched, but since the packet is passed to the shaping device at the timing when the packet is generated, the parity data to be sent is inserted at a position across blocks. Thus, as in the case shown in FIG. 10, there is a problem that the FEC decoder may not function as intended depending on the occurrence of delay and the setting of timeout processing. Further, in the MMT, the MMTP packet is transmitted in accordance with the transmission time specified in the Coordinated Universal Time attached to the MMT standard packet (MMTP packet). When the transmission timing is changed by the shaping device, there appears a problem that a jitter in the network appears from the receiving side and a new problem occurs in the buffer control of the receiving terminal.

  In Non-Patent Document 2, a delay in the relay node is avoided by transferring the arrived packet before the FEC calculation, but the method for transmitting the parity data is not considered.

  The present invention has been made in view of the above, and an object of the present invention is to smooth and transmit data with an error correction code added with low delay.

An encoding apparatus according to a first aspect of the present invention includes encoding means for calculating an error correction code for each predetermined block for input data, and a transmission rate of the input data in consideration of an increase in the error correction code. A buffer that holds the buffer amount for the time required for sending out the error correction code at the maximum rate as a lower limit, and all input data of the block are input at a transmission rate that takes into account the increase in the error correction code Sending means for sending all the input data of the block and the error correction code within the time corresponding to the block, and the sending means corrects the error correction from the time when the first input data of the block is inputted. After the input data is held in the buffer for the time required for sending the code, the input data starts to be sent, and after all the input data of the block has been sent, Characterized in that it starts transmission of Ri correction code.

An encoding method according to a second aspect of the present invention is a computer encoding method, comprising: calculating an error correction code for each predetermined block for input data; and using the input data as a buffer for the error correction code. Holding the buffer amount for the time required for sending the error correction code as a lower limit at the maximum rate of the transmission rate in consideration of the increase in the block, and the block at the transmission rate in consideration of the increase in the error correction code Transmitting all the input data of the block and the error correction code within the time of one block in which all the input data of the block are input, and the step of transmitting includes the step of transmitting the first input of the block The input data is held in the buffer for a time required for sending the error correction code from the time when the data is inputted, and then the input data starts to be sent. Characterized in that to start the error correction transmission of codes after finished sending all the input data of the block.

  An encoding program according to a third aspect of the present invention is characterized in that a computer is operated as each unit of the encoding apparatus.

  According to the present invention, data to which an error correction code is added can be smoothed and transmitted with low delay.

It is a functional block diagram which shows the structure of the encoding apparatus in 1st Embodiment. It is a figure which shows the source data which the encoding apparatus in 1st Embodiment receives, the source data and parity data which are sent out. It is a functional block diagram which shows the structure of the encoding apparatus in 2nd Embodiment. It is a figure which shows the example of the time stamp of an input packet and an output packet. It is a figure which shows the example of the time stamp of the input packet and output packet containing a packet outside FEC object. It is a figure which shows the example of the time stamp of multiple types of input packet and an output packet. It is a functional block diagram which shows the structure of the encoding apparatus in 3rd Embodiment. It is a figure which shows the source data which the encoding apparatus in 3rd Embodiment receives, the source data and parity data which are sent out. It is a figure which shows the prior art example which transmits parity data. It is a figure which shows another example of the conventional which transmits parity data.

[First Embodiment]
FIG. 1 is a functional block diagram showing the configuration of the encoding apparatus according to the present embodiment.

  The encoding apparatus illustrated in FIG. 1 includes an FEC generation unit 11, a buffer function unit 12, and a transmission unit 13. This encoding device is a device that adds parity data (error correction code) to input source data in units of blocks and transmits the data, and transmits source data and parity data with a transmission amount that takes into account the increase in parity data. . Each unit included in the encoding device may be configured by a computer including an arithmetic processing device, a storage device, and the like, and the processing of each unit may be executed by a program. This program is stored in a storage device included in the encoding device, and can be recorded on a recording medium such as a magnetic disk, an optical disk, or a semiconductor memory, or provided through a network.

  The FEC generation unit 11 calculates the parity data of the input source data for each predetermined block. When the FEC generating unit 11 calculates the parity data from the source data for one block, the FEC generating unit 11 transmits the parity data to the buffer function unit 12.

  The buffer function unit 12 has a function of temporarily holding the source data and parity data sent out by the transmission unit 13, and temporarily stores the source data until the source data can be transmitted with a transmission amount that takes into account the increase in the parity data. Hold on.

  The transmission unit 13 reads the source data and parity data from the buffer function unit 12 and sends out the source data and parity data for one block in the same time as the time when the source data for one block is input. Since the size of the data to be transmitted is increased by the size of the parity data as compared with the input source data, the transmission unit 13 inputs the source data so that the source data and the parity data can be transmitted within the time of one block. Source data and parity data are transmitted at a transmission rate faster than the transmission rate.

  Next, the operation of the coding apparatus according to the present embodiment will be described.

  FIG. 2 is a diagram illustrating source data received by the encoding apparatus according to the present embodiment, source data to be transmitted, and parity data.

  The encoding apparatus starts to input a block of source data A0 from time TA0. The encoding apparatus stores the input source data A0 in the buffer function unit 12 and does not transmit it until time TA1. The encoding apparatus according to the present embodiment stores the input source data A0 in the buffer function unit 12 without immediately transmitting it, thereby creating a gap for transmitting the parity data of the previous block. For example, when the size of parity data is 1 Mbit and the transmission transmission rate is 10 Mbps, the transmission of parity data requires a time of 1 Mbit / 10 bps = 100 msec. Therefore, the buffer function unit 12 has a buffer capable of holding 100 mesc of source data A0. It is 100 msec from the input of the source data A0 to the start of transmission, that is, from time TA0 to time TA1.

  The encoding device starts sending the source data A1 stored in the buffer function unit 12 at the sending transmission speed S1 from time TA1. Although the input source data A0 and the source data A1 to be sent are the same data, different codes A0 and A1 are assigned to the source data in order to distinguish between input and sending.

  The transmission transmission speed S1 is determined such that the transmission of the source data A1 and the parity data Ap is completed within the time for inputting all the blocks of the source data A0, that is, within the time for one block. For example, when the input transmission rate S0 of the source data A0 is 10 Mbps, FEC of 1 block per second is applied (block input time is 1 second), and 10% parity data Ap is added to the source data A0. think of. Since the size of the source data A0 is 10 Mbps × 1 sec = 10 Mbit, the size of the parity data Ap is 10 Mbit × 10% = 1 Mbit. Since the encoding device sends out 11 Mbit of data including the source data A1 and the parity data Ap in one second, the necessary transmission transmission rate S1 is at least 11 Mbps.

  The encoding device inputs the last source data A0 of the block at time TA2. The encoding apparatus according to the present embodiment immediately transmits the last source data A0 of the block without delay. That is, the encoding device has finished sending all of the source data A1 at time TA2.

  When the last source data A0 of the block is input at time TA2, the encoding device calculates the parity data Ap of the block of the source data A0 and starts sending the parity data Ap. Since the encoding device has finished sending all of the source data A1 at time TA2, it can send parity data Ap from time TA2.

  The encoding device starts to input source data B0, which is the next block of source data A0, from time TA2. The encoding apparatus stores the input source data B0 in the buffer function unit 12, and does not send the source data B1 until time TB1.

  The encoding device finishes sending the parity data Ap by time TB1. The time from the time TA1 at which the transmission of the source data A1 is started to the time TB1 is the same as the time for one block. In other words, the encoding device transmits source data A1 and parity data Ap in a time corresponding to one block.

  The encoding device starts sending the source data B1 of the next block from the time TB1 at which the parity data Ap of the previous block has been sent. Similarly, the encoding apparatus transmits source data B1 and parity data Bp within a time corresponding to one block.

  In this embodiment, the transmission of source data is delayed by an increase in parity data (from time TA0 to time TA1), but the receiving side can receive and decode the source data and parity data in one block time. . Therefore, in total, the delay can be suppressed to the increase of the parity data and the time of one block even if the delay until decoding on the receiving side is included.

  On the other hand, when the parity data is transmitted during the transmission of the source data of the next block in order to smooth the data to be transmitted, the receiving side starts receiving the source data and then blocks it until it can be decoded. Minutes of time are required. In this case, if the length of one block is made longer, the delay becomes larger.

  As described above, according to the present embodiment, the encoding device calculates parity data for each predetermined block, and the source for one block is the same as the time when the source data for one block is input. By sending data and parity data, the receiving device can receive the source data and parity data in a time of one block after receiving the source data at the head of the block. Time can be kept down.

  According to the present embodiment, the input source data is temporarily held until the time necessary for sending the parity data of the previous block elapses, and the source data after the sending of the parity data of the previous block is finished. By starting the transmission of data, parity data is not transmitted in bursts, and the delay until the start of transmission of source data can be suppressed to an increase in parity data.

[Second Embodiment]
FIG. 3 is a functional block diagram showing the configuration of the encoding apparatus according to the present embodiment.

  The coding apparatus shown in FIG. 3 is obtained by adding a time stamp assigning unit 14 that gives a time stamp indicating a transmission time to the packet to the coding apparatus according to the first embodiment. The transmission unit 13 transmits a packet according to a time stamp (transmission time).

  When an MMTP packet is input as source data, it is necessary to give the MMTP packet the transmission time of the packet specified by UTC (Coordinated Universal Time).

  The time stamp adding unit 14 adds a time stamp to which an offset based on the position of the packet in the block is added. As described in the first embodiment, the first source data in the block is not transmitted until the transmission of the parity data of the previous block is completed, but the last source data in the block is transmitted immediately. That is, the time from input to transmission of a packet becomes shorter as it goes from the first packet to the last packet in the block. Therefore, the time stamp assigning unit 14 assigns a time stamp obtained by adding the buffer time to the original time stamp to the first packet in the block, and an offset for the buffer time to the last packet in the block. A time stamp that is not added is assigned. A time stamp obtained by adding an offset corresponding to the position in the block of the packet is assigned to the packet between the first and last packets.

  Next, a specific example of giving a time stamp to a packet will be described.

  Here, a method for deriving a time stamp for transmitting a packet from the maximum bit rate of input and transmission will be described. Specifically, a time stamp is added with a time obtained by adding a value obtained by multiplying the packet interval by the ratio of the input transmission rate to the output transmission rate (input transmission rate / sending transmission rate) to the sending time of the previous packet. To do.

  The upper part of FIG. 4 shows the input packet and its time stamp value, and the lower part shows the outgoing packet and the time stamp value after rewriting. One block is composed of ten packets P1 to P10. Assume that the interval between input packets is 10. Since the encoding apparatus transmits 11 packets, which are a combination of the packets P1 to P10 and the parity data packet FEC, within one block time, the interval between the transmission packets is 10 × 10 / 11≈9.

  At time t0 (= 10), the first packet P1 of the block is input. The packet P1 is temporarily held in the buffer function unit 12 until the parity data of the previous block has been transmitted.

  At time t1 (= 19), the head packet P1 of the block is transmitted. The packet P1 is rewritten to 19 by adding 9 to the original time stamp (= 10). The time stamp of the packet P2 next to the packet P1 is rewritten to 28 obtained by adding 9 to the time stamp of the previous packet P1. Hereinafter, similarly, the packets P3 to P10 are transmitted with the packet interval set to 9. In packets P1 to P10, offsets of 9, 8, 7, 6, 5, 4, 3, 2, 1, and 0 are added to the original time stamps based on the positions in the block, respectively.

  At time t2 (= 100), the last packet P10 of the block is transmitted with the original time stamp (= 100). That is, the last packet P10 of the block is transmitted without delay.

  After all the packets P1 to P10 in the block are transmitted, the parity data packet FEC is transmitted. The time stamp of the parity data packet FEC is given with reference to the time stamp of the last packet P10 of the block.

  In the example of FIG. 4, the time for one block to input the packets P1 to P10 is 90, and the time to send the packets P1 to P10 and the packet FEC is 90.

  After sending the packet FEC, the packet P11 of the next block is sent.

  Next, a process for giving a time stamp when a packet that is not subject to FEC is included will be described.

  Packets that are not subject to FEC are sent without changing the interval with the previous packet.

  The upper part of FIG. 5 shows an input packet including a packet not subject to FEC and its time stamp value, and the lower part shows a transmission packet and a time stamp value after rewriting. In the example of FIG. 5, packets P4, P5, P7, P10, and P13 are packets that are not subject to FEC. Assume that the interval between input packets is 10.

  After the time t1 (= 19), the FEC target packets P1, P2, and P3 are transmitted with a packet interval of 9. The encoding apparatus gives a time stamp (= 47) obtained by adding a packet input interval (= 10) to the time stamp (= 37) of the previous packet P3 to the packet P4 not subject to FEC. The encoding device changes the packet interval according to the transmission transmission rate for the FEC target packet, and does not change the packet interval for the non-FEC packet.

  At time t2 (= 150), the last packet P15 of the block is transmitted without delay. Thereafter, parity data is calculated, and a parity data packet FEC is transmitted.

  Next, processing when a plurality of types of packets are included will be described.

  When different types of packets such as a video stream and an audio stream are input to the encoding device, the encoding device calculates parity data in units of blocks in which the different types of packets are combined, not for each type of packet. .

  The upper part of FIG. 6 shows the input video and audio packets and their time stamp values, and the lower part shows the outgoing packets and the time stamp values after rewriting. In FIG. 6, one block is composed of 10 packets.

  At time t0 (= 10), the top video packet V1 of the block is input. The video packet V1 is temporarily held in the buffer function unit 12 until time t1 (= 19).

  At time t1 (= 19), the video packet V1 is sent out. Thereafter, video packets V2 and V3 are transmitted with the packet transmission interval changed. Since the audio packet A1 is input next to the video packet V3, the audio packet A1 is transmitted next to the video packet V3. Thereafter, video packets and audio packets are input, and the packets are transmitted in the input order.

  At time t2 (= 65), the last voice packet A4 of the block is input. The encoding device calculates parity data from the video packets V1 to V6 and the audio packets A1 to A4, and transmits a parity data packet FEC after transmitting the audio packet A4.

  Similarly for the next block, packet data is calculated from the video packets V7 to V11 and the audio packets A5 to A9.

  In the examples described so far, the time stamp is calculated by assuming that the packet is input at the maximum bit rate and shortening the transmission interval of each packet. However, the time stamp to be given to the packet based on the actual bit rate is calculated. You may calculate sequentially. For example, the time stamp may be calculated from the relationship between the first packet of the block and the packet to be processed, or the time stamp may be calculated from the interval between the previous packet and the packet to be processed.

  Further, in the block, when the time stamps assigned to the sequentially input packets are reversed, the same value as the time stamp of the previous packet may be forcibly assigned.

  As described above, according to the present embodiment, the time stamp giving unit 14 adds a value corresponding to the position in the block of the packet to the time stamp given to the packet, so that the transmitting unit 13 The input source data and parity data can be smoothed and transmitted simply by transmitting a packet in accordance with the assigned time stamp.

[Third Embodiment]
FIG. 7 is a functional block diagram showing the configuration of the encoding apparatus according to the present embodiment.

  The encoding apparatus shown in FIG. 7 includes an alignment unit 15 that aligns the time stamps of the source data to be transmitted to the encoding apparatus of the first embodiment and interleaves the source data and the parity data. It is added. For example, when packets with time stamps of 0, 60, 60, 60, 60, 60, and 70 are input, the transmission interval is large from 0 to 60, but packets with a time stamp of 60 are output in bursts. Therefore, after inputting the source data for one block, the aligning unit 15 considers the distribution of the time stamps and, for example, sets the above time stamps to 100, 109, 118 so that the time stamps of the packets to be transmitted are uniform. , 127, 136, 145, 154.

  In the encoding apparatus according to the present embodiment, after the source data for one block is input and the parity data is calculated, the alignment unit 15 executes the process and sends the source data and the parity data in a time corresponding to one block. . In the present embodiment, there is a delay of one block or more from when the encoding apparatus inputs source data until transmission starts. However, in the receiving apparatus, the time from the reception of the source data at the head of the block to the reception and decoding of the data for one block including the parity data is the time for one block, and the delay amount on the receiving side Can be suppressed.

  Next, the operation of the coding apparatus according to the present embodiment will be described.

  FIG. 8 is a diagram illustrating source data received by the encoding apparatus according to the present embodiment, source data to be transmitted, and parity data.

  The encoding apparatus starts to receive the block of the source data A0 from time TA0. The encoding device temporarily holds the source data A0 and does not send it out.

  At time TB0, the encoding apparatus inputs all the blocks of the source data A0 and starts to input the source data B0 of the next block.

  When all the blocks of the source data A0 are input, the encoding device calculates the parity data Ap from the source data A0.

  The aligning unit 15 aligns or interleaves the source data A1 and the parity data Ap, and gives a time stamp that is a transmission time.

  When the sending time TA1 is reached, the transmitting unit 13 starts sending the source data A1 and the parity data Ap according to the assigned time stamp. The source data A1 and the parity data Ap are transmitted in a time corresponding to one block when the source data A0 is input.

  As described above, according to the present embodiment, the aligning unit 15 aligns or interleaves the source data A1 and the parity data Ap, thereby transmitting data more smoothly and increasing resistance to burst errors. Can do.

DESCRIPTION OF SYMBOLS 11 ... FEC production | generation part 12 ... Buffer function part 13 ... Transmission part 14 ... Time stamp provision part 15 ... Alignment part

Claims (5)

Encoding means for calculating an error correction code for each predetermined block with respect to input data;
A buffer for holding the input data as a lower limit of a buffer amount for a time required for sending out the error correction code at a maximum rate of transmission speed considering an increase in the error correction code ;
Sending means for sending out all the input data of the block and the error correction code within a time of one block in which all the input data of the block is inputted at a transmission rate considering the increase of the error correction code Have
The sending means starts sending the input data after holding the input data in the buffer for a time required for sending the error correction code from the time when the first input data of the block is inputted. An encoding apparatus , wherein transmission of the error correction code is started after transmission of all input data is completed . The input data is a packet with a time stamp,
A time stamp giving means for rewriting the time stamp by adding a value corresponding to the position in the block of the packet to the time stamp of the packet;
The encoding apparatus according to claim 1, wherein the sending unit sends the packet according to a time stamp of the packet. The time stamp assigning means assigns, as a time stamp, a time obtained by adding a value obtained by multiplying the input interval of the packet by a ratio of an input transmission rate to an output transmission rate to a time stamp of a previous packet. 2. The encoding device according to 2. A computer encoding method comprising:
Calculating an error correction code for each predetermined block with respect to input data;
Holding a buffer amount as a lower limit for the time required for sending the error correction code at a maximum rate of transmission speed in consideration of an increase in the error correction code in the input data;
Sending all the input data of the block and the error correction code within a time of one block in which all the input data of the block is inputted at a transmission rate considering the increase of the error correction code; Have
The sending step starts sending the input data after holding the input data in the buffer for a time required for sending the error correction code from the time when the first input data of the block is inputted. An encoding method , wherein transmission of the error correction code is started after transmission of all the input data is completed . An encoding program for causing a computer to operate as each unit of the encoding device according to claim 1 .

Images