JP6796234B2 - Video transmission system - Google Patents

Description

The present invention relates to a video transmission system, and more particularly to a video transmission system in which the same RTP sequence number is set for IP packets having the same payload between redundant transmitters.

A broadcasting system that transmits video from a sports broadcast from a stadium to a broadcasting center or a direct broadcasting station, or a video transmission system that transmits a video signal from a (core) broadcasting station to another broadcasting station by multicast is known. ing. When such a video signal is transmitted, the video signal is often converted into an IP packet and transmitted via an IP network according to RTP (Real-time Transport Protocol). RTP is a data communication protocol for delivering data signals such as voice and moving images in real time.

If a failure occurs in the transmission path used to transmit the IP packet described above, the IP packet may not be transmitted. In order to deal with such an event, it is common practice to make the transmission path redundant. By making the transmission path redundant, when a failure occurs in the transmission path of the active system, the IP packet can be transmitted by switching to the IP packet from the transmission path of the backup system. With such a configuration, it is possible to prevent the transmission of the IP packet from being interrupted. Here, in the case of a video signal, if the switching is performed without any control, the disturbance of the video becomes visible to the viewer. Therefore, seamless switching is required when switching as described above.

SMPTE (Society of Motion Picture and Television Engineers) 2022-7 is defined as a standard for a video signal transmission method for switching IP packets as described above. In SMPTE2022-7, a transmitter (transmitter) that transmits an IP packet on an IP network transmits a plurality of IP packets having the same payload data to a destination on a plurality of different transmission paths (Ethernet: registered trademark). It stipulates that. FIG. 1 shows a packet transmission method according to SMPTE2022-7.

As shown in FIG. 1, an IP network 3 exists between the transmitter 1 and the receiver 2, and both are connected via a transmission path 4 (active system) and a transmission path 5 (spare system). The transmitter 1 that has received the input video signal (SDI signal, DVB-ASI signal, or video and audio information contained in the payload in the IP packet received via Ethernet) transmitted from a broadcasting station or the like receives the input video signal. Encapsulate in a series of IP packet groups (hereinafter referred to as "IP packet stream"), and further duplicate each IP packet (this replication produces two IP packet streams having the same payload data) and two. The IP packet stream is transmitted via each of the transmission path 4 and the transmission path 5. The same RTP sequence number is set for the two duplicated IP packets. However, since the two IP packets are transmitted through the transmission path 4 and the transmission path 5, different values are set for some or all of the IP header, MAC header, and VLAN tag if necessary. Will be done. The receiver 2 receives the two IP packet streams transmitted on the transmission path 4 and the transmission path 5, respectively, and usually reconstructs the video signal based on the IP packet stream received from the transmission path 4.

Here, if a failure occurs in the transmission path 4, it is necessary to switch to the IP packet received from the transmission path 5. At the time of this switching, the IP packet stream received from the transmission path 4 and the IP packet stream received from the transmission path 5 are buffered by the buffer memories (delay buffers) 2a and 2b for adjusting the delay time in the receiver 2, respectively. , Adjust so that the delays of both systems at the switching point are the same. This makes it possible to switch the video output at the time of switching without disturbing it.

FIG. 1 shows that a failure has occurred in the transmission path 4 when the receiver 2 receives an IP packet in which the RTP sequence number is set to “2”. In this case, the receiver 2 cannot receive the IP packet in which the RTP sequence number is set to "3" or "4" from the transmission path 4. On the other hand, the receiver 2 receives the IP packets having the RTP sequence number “2” and the IP packets “3” and “4” from the transmission path 5. At the stage of constructing one IP packet stream from the two received IP packet streams, the receiver 2 detects that the IP packet having the RTP sequence number "3" does not exist in the delay buffer of the transmission path 4. , Switching to the transmission path 5. Then, by using the IP packets whose RTP sequence numbers "3" and "4" output by the delay buffer of the transmission path 5 are used for constructing the IP packet stream, no packet loss occurs and the resulting reconstruction is output. It does not disturb the finished video signal. As a result, seamless switching of the video signal is realized.

Application note DT-AN-IP-3, SMPTE 2022-7 “Seamless Protection Switching” for DekTec network adapter, [online], Internet (URL: http://www.dektec.com/products/PCIe/DTA-2162/ downloads / DT-AN-IP-3%20SMPTE%202022-7.pdf)

Even with the above method, of course, if the transmitter fails, it may not be possible to transmit the IP packet stream on either of the two transmission paths. Therefore, if higher reliability is required, the transmitter needs to be redundant (by using multiple enclosures and / or cards). Here, in the method specified by SMPTE2022-7, since the video signal is reconstructed by using the RTP sequence number, it is necessary to match the RTP sequence number between a plurality of redundant transmitters. However, in the prior art, there is no technique for matching RTP sequence numbers between a plurality of transmitters, and the method of reconstructing a video signal based on the RTP sequence numbers cannot make the transmitters redundant.

As a method of matching RTP sequence numbers between redundant transmitters, for example, it is conceivable that a plurality of transmitters exchange RTP sequence numbers with each other. However, in order to use such a method for transmitting high-speed uncompressed video data specified in SMPTE2022-6, it takes a lot of time to exchange sequence numbers, and there is a relative overhead of protocol processing. It has grown and is not realistic with current technology.

The present invention has been made in view of such a problem, and it is not necessary to exchange RTP sequence numbers between a plurality of transmitters, and each transmitter has the same RTP for an IP packet having the same payload. It is an object of the present invention to provide a video transmission system capable of setting a sequence number and transmitting it on an IP network.

In order to solve the above problems, the video transmission system according to the present invention has a first method in which the PTP ground master and the PTP ground master and the PTP are time-synchronized according to PTP, and a video signal distributed from one video signal is input. The first transmitter and the second transmitter are provided with a transmitter, a second transmitter, a receiver connected to the first transmitter and the second transmitter via an IP network, and the first transmitter and the second transmitter. Each of the transmitters of the above divided the distributed video signal into a plurality of IP packets, calculated the initial value of the RTP sequence number based on the PTP time stamp time-synchronized according to the PTP, and calculated the initial value of the RTP sequence number from the initial value. Consecutive sequence numbers are set in the RTP sequence number fields of the plurality of divided IP packets, and the first transmitter sets the plurality of IP packets in which the RTP sequence numbers are set in the first plurality. As the IP packet of the above, the second transmitter transmits the plurality of IP packets in which the RTP sequence number is set to the receiver via the first transmission path according to the RTP. As an IP packet, it is transmitted to the receiver via a second transmission path according to RTP.

In order to solve the above problems, another video transmission system according to the present invention includes a first transmitter, a second transmitter, and the first transmitter, which input a video signal distributed from one video signal. The transmitter includes the transmitter 1 and the second transmitter and a receiver connected via an IP network, and each of the first transmitter and the second transmitter receives the distributed video signal. Divided into a plurality of IP packets, the first transmitter determines a first video frame boundary from the input video signal, and after a predetermined time has elapsed from the determined first video frame boundary. In parallel with issuing the RTP seamens number initialization command to the second transmitter, the Nth video frame boundary is determined from the input video signal, and based on the Nth video frame boundary, the Nth video frame boundary is determined. The RTP sequence number is initialized, and the second transmitter determines the second video frame boundary from the input video signal in response to receiving the initialization command, and the second video It is characterized in that the RTP sequence number is initialized based on the frame boundary.

According to the video transmission system according to the present invention, a plurality of redundant transmitters do not need to exchange RTP sequence numbers with each other, and each independently sets the RTP sequence number according to the RTP and transmits an IP packet. can do.

It is a figure which shows the example of the method of transmitting a video signal according to SMPTE2022-7. It is a figure which shows the processing of PTP used by the video transmission system which concerns on 1st Embodiment of this invention. It is a figure which shows the example of the structure of the video transmission system which concerns on 1st Embodiment of this invention. It is a figure which shows the example of the structure of the transmitter which concerns on 1st Embodiment of this invention. It is a figure which shows the specific setting method of the RTP sequence number which concerns on 1st Embodiment of this invention. It is a figure which shows the example of the structure of the transmitter which concerns on the 2nd Embodiment of this invention. It is a figure which shows the specific setting method of the RTP sequence number which concerns on the 2nd Embodiment of this invention.

IEEE1588v2 defines a highly accurate time distribution method via a LAN (Local Area Network). A network that synchronizes time with PTP is composed of a PTP grand master and a PTP slave. The PTP grand master is usually implemented as hardware that can use a high-precision GPS signal, a radio clock signal, a BB (Black Burst) signal for synchronizing a video signal, or the like as a reference clock source. The PTP slave synchronizes time with the PTP grand master.

As shown in FIG. 2, the relationship between the PTP grand master 201 and the PTP slave 202 is established by the PTP grand master 201 transmitting an Unknown message to the PTP slave 202 (step S201).

Next, the PTP grand master 201 transmits a Sync message to the PTP slave 202 (step S202). The Sync message includes a message transmission time (t1), and when the PTP slave 202 receives it, an arrival time (t2) is recorded.

The PTP slave 202 transmits a Delay_Req message to the PTP grandmaster 201 (step S203). In the Delay_Req message, the actual value (t3) of the message transmission time is recorded.

Next, the PTP grand master 201 transmits a Delay_Resp message to the PTP slave 202 (step S204). The actual arrival time (t4) of the Delay_Req message is recorded in the Delay_Resp message.

The PTP slave 202 uses t1, t2, t3, and t4 to calculate the offset (time difference) between the round-trip delay time, the time of the PTP grand master 201, and the time of the PTP slave 202. Assuming that the one-way delay time is half the round-trip delay time, the offset of the PTP slave 202 is calculated by the following equation.
PTP slave clock offset = (t2-t1)-(one-way delay time)
(One-way delay time) = {(T2-T1) + (T4-T3)} / 2
Equation 1
The PTP slave 202 repeats the above processing a plurality of times per second, smoothes the offset value of the processing result, and feeds it back to the time information, so that the time synchronization with the PTP grand master 201 is maintained.

This description describes the operation of PTP with a 1-step clock (1-step clock) and an E2E delay mechanism (End-to-End delay mechanism). Actually, in PTP, the specifications of the 2-step clock (2-step clock) or the P2P delay mechanism (Peer-to-Peer delay mechanism) are also defined, and further correction and other processing are performed to improve the error accuracy. Although high time synchronization is achieved, no further details will be given herein.

Next, an example of the configuration of the video transmission system according to the first embodiment of the present invention will be described with reference to FIG. The video transmission system according to the present embodiment includes SMPTE2022-2 (which defines the transmission of a constant bit rate MPEG-2 transport stream (TS) on an IP network) and SMPTE2022-6 (a high bit rate medium on an IP network). It is assumed that the protocol specified in SMPTE2022, such as (regulating signal transmission), is compliant. In addition, the transmission of IP packets on the IP network is premised on conforming to RTP / UDP.

As shown in FIG. 3, the video transmission system according to the first embodiment of the present invention includes a distributor 11, a PTP ground master 12, a transmitter 13a, a transmitter 13b, and a receiver 14. The transmitter 13a, the transmitter 13b, and the receiver 14 are connected to each other via the IP network 15. A transmission path 16 exists between the transmitter 13a and the receiver 14, and a transmission path 17 exists between the transmitter 13b and the receiver 14.

When the distributor 11 receives the input video signal transmitted from a broadcasting station or the like, the distributor 11 distributes the input video signal to the transmitters 13a and 13b. That is, the distributor 11 duplicates the input video signal. If the video signal is an SDI signal or DVB-ASI signal, the distributor 11 is a distributor with a coaxial cable or fiber optic cable connector, and if the video signal is received via Ethernet, the distributor 11 is Ethernet. It is a switch of.

The PTP ground master 12 functions as the ground master described with reference to FIG. 2 and synchronizes the PTP time with the transmitters 13a and 13b. The PTP ground master 12 may also perform PTP time synchronization with the receiver 14, but time synchronization with the receiver 14 is not essential.

The transmitters 13a and 13b each divide the input video signal distributed by the distributor 11 into IP packets, and transmit the IP packets to the receiver 14 on the IP network 15. The transmitter 13a transmits the IP packet A divided and IP-packetized from the input video signal to the receiver 14 on the transmission path 16 according to the RTP. The transmitter 13b transmits the input video signal divided and the IP packet B converted into an IP packet to the receiver 14 on the transmission path 17 according to the RTP. That is, each of the transmitters 13a and 13b transmits IP packets having the same payload on separate transmission paths. Here, the same RTP sequence number is set for IP packets having the same payload, the details of which will be described later.

The transmitters 13a and 13b also function as PTP slaves and perform time synchronization with the PTP ground master 12 according to PTP. The time here is a numerical value expressed as the time from the time of SMPTE Epoch (epoch), that is, the time at midnight (0:00:00 am) on January 1, 1970 in Coordinated Universal Time (UTC). That is, the transmitters 13a and 13b have an elapsed time from the epoch at a particular point in time with nanosecond level error accuracy.

In the present embodiment, two transmitters, transmitters 13a and 13b are used for redundancy, but the number of transmitters to be configured is not limited to two, and the transmitters are configured by any number. May be good. In this case, the transmission path is also composed of the number corresponding to the number of transmitters.

Further, the PTP ground master 12 may be mounted on any of the transmitters 13a and 13b, or may be mounted on a third transmitter (not shown) other than the transmitters 13a and 13b. For example, when the PTP ground master 12 is mounted on the transmitter 13a, the transmitter 13b synchronizes the time with the transmitter 13a. When the PTP ground master 12 is mounted on the third transmitter, the transmitters 13a and 13b synchronize the time with the third transmitter.

The receiver 14 receives both the IP packet A transmitted on the transmission path 16 and the IP packet B transmitted on the transmission path 17. Here, under the normal state, the receiver 14 reconstructs the video signal based only on the IP packet A, for example (the IP packet B is discarded).

On the other hand, when any of the transmitter 13a and the transmission path 16 (hereinafter referred to as the active system) fails, the receiver 14 needs to be switched to reconstruct based on the IP packet B from the transmission path 17. Here, in order to switch the output video signal without being disturbed, the IP packets A and B are buffered in the delay buffer in the receiver 14, so that the timing of the IP packets having the same payload is adjusted at the time of switching. Whether or not the IP packet B has the same payload as the IP packet A is determined by referring to the RTP sequence numbers set in both (specified by SMPTE2022-7).

Next, an example of the configuration of the transmitters 13a and 13b (collectively, “transmitter 13” in FIG. 4) described with reference to FIG. 3 will be described with reference to FIG. As shown in FIG. 4, the transmitter 13 includes a video signal receiving unit 131, an IP packet generating unit 132, a PTP control unit 133, a PTP time stamp counter 134, a frame boundary determination unit 135, a sequence number setting unit 136, and a transmitting unit 137. , Data network interface unit 138, and control network interface unit 139.

The video signal receiving unit 131 receives the video signal distributed by the distributor 11. The IP packet generation unit 132 generates an IP packet defined by the RTP / UDP / IP protocol from the received video signal in order to transmit the video signal by IP. An IP packet is generated by adding an RTP header, a UDP header, an IP header, and a MAC header.

The PTP control unit 133 receives a PTP message from the PTP grand master 12 via the control network interface unit 139, and synchronizes the PTP time with the PTP grand master 12. The PTP time stamp counter 134 is counted up based on the time synchronization.

The frame boundary determination unit 135 determines the video frame boundary of the video signal input to the video signal receiving unit 131. When the video signal is an uncompressed video signal, the video frame boundary is determined according to regulations such as SMPTE424MSMPTE292M or SMPTE259M. When the video signal is DVB-ASI, the value of the payload unit start indicator (when the input video signal conforms to SMPTE2022-2) in the TS header is referred to to determine the video frame boundary. If the input interface of the video signal is Ethernet, the video signal is input by the IP packet, and the format of the IP packet conforms to SMPTE2022-6, refer to the M (marker) bit in the RTP header of the IP packet. Then, the video frame boundary is determined. When the input interface of the video signal is Ethernet, the video signal is input by the IP packet, and the format of the IP packet conforms to SMPTE2022-2, the payload unit start indicator in the TS header (the input video signal is SMPTE2022). Refer to the value of (when conforming to -2) to determine the video frame boundary. Since the above-mentioned determination of the video frame boundary is based on the standard and is a conventional technique that is still implemented in many video devices, no further detailed description will be given in this specification.

The sequence number setting unit 136 calculates the RTP sequence number based on the time stamp counted up by the PTP time stamp counter 134. The calculated value is set in the RTP sequence number field (specified by RTP) in the RTP header of the first IP packet of the video frame with reference to the frame boundary determined by the frame boundary determination unit 135. Here, a specific method for setting the RTP sequence number will be described with reference to FIG.

FIG. 5 shows the relationship between the frame boundary of the transmitted IP packet and the time with the horizontal axis as the time axis. As shown in FIG. 5, one video frame has a predetermined number of IP packets.

First, the sequence number setting unit 136 refers to the time stamp of the PTP time stamp counter 134 that has been counted up, and when one second elapses, that is, when the time is updated in seconds (“second boundary””. At the hour (in FIG. 5, "second boundary 1" is shown), the lower 16 bits of the integer part of the value calculated according to Equation 1 are sampled.
Sampling value (SV) = ((32CV-PSV) x VFN x IPN)
Equation 1
32CV is a 32-bit count value of seconds or more of the updated PTP time stamp counter, PSV is a value of seconds or more of a predetermined time set in advance, and VFN is the number of video frames per second. IPN is the number of IP packets per frame of video. Here, the "predetermined time set in advance" is a value common to the transmitters 13a and 13b, and in the embodiment of FIG. 5, it is the time from the SMPTE Epoch defined by SMPTE2057-1. .. When the total number of video frames from the SMPTE Epoch is used as it is for calculating the RTP sequence number, the "predetermined time set in advance" is 0. It is obvious that the above calculation can be easily realized in an ordinary embedded processor or a logic circuit.

Next, the sequence number setting unit 136 calculates the IP packet (frame start IP packet) that is the start of the video frame based on the video frame boundary (frame boundary IP packet) determined by the frame boundary determination unit 135. The RTP sequence number is set as the initial value. Here, the video frame boundary IP packet means the last IP packet of one video frame.

In the example of FIG. 5, the boundary of seconds is reached while a predetermined number of IP packets in the video frame N-1 are generated. At this point, it is assumed that the value of the second counter of the PTP time stamp is 16,890 days (46 years and 100 days), 1 hour, 2 minutes, and 3 seconds ahead of SMPTE Epoch. In this case, 32 CV is 1,459,299,723 seconds. Assuming that the format of the transmitted video signal is 1080P, the size of one video frame is 6,187,500 bytes, and when using an IP packet with a payload length of 1376 bytes specified by SMPTE2022-6, the IPN is 4497. Is. When the frequency of the video frame is 59.94 Hz, the VFN is (60 Hz × 1000/1001). Assuming that PSV is 0, the integer value that is the basis of the sampling value at the second boundary in FIG. 5 is according to Equation 1.
1,459,299,723 x (60 x 1000/1001) x 4497
Therefore, the following value is obtained by rounding down the decimal point from the calculation result of Equation 1.
393,354,896,363,496 (0x165C112DE4FE8 in hexadecimal)
The lower 16 bits of the calculated sampling value are 0x4FE8.

When the boundary between the video frame N and the video frame N-1 is determined, "0x4FE8" is set as an initial value in the RTP sequence number of the frame start IP packet of the video frame N. For the IP packet to be subsequently transmitted, an RTP sequence number in which 1 is sequentially added up to the next "second boundary" (indicated by "second boundary 2" in FIG. 5) is set.

The transmitters 13a and 13b are time-synchronized by PTP having an error accuracy at the nanosecond level, and the difference in the delay time of the video signal delivered from the distributor 11 to both transmitters is the length of the cable to be delivered. It is about the sum of the difference between the delay time and the delay time in the distributor 11. Therefore, if the difference between the "second boundary" time and the timing of the video frame boundary is larger than the above delay time difference, the IP packet A having the same payload transmitted via the transmission paths 16 and 17, respectively, and For IP packet B, the calculation result of Equation 1 is the same, and both IP packets have the same RTP sequence number.

On the other hand, when the difference between the "second boundary" time and the video frame boundary timing is smaller than the above delay time difference, that is, when the "second boundary" time and the video frame boundary timing difference are almost the same. , IP packet A and IP packet B having the same payload transmitted via the transmission paths 16 and 17, respectively, the above calculation results may not be the same. In this case, the sampling value is not set to the RTP sequence number at the next "second boundary" (in FIG. 5, "second boundary 2" with reference to "second boundary 1"), and the sampling value is sequentially 1 to the RTP sequence number. May continue to be added. Instead, by delaying the timing of calculating the sampling value by a certain timing from the "second boundary", the calculation result of Equation 1 becomes the same for the above-mentioned IP packet A and IP packet B, and both IP packets are displayed. They can have the same RTP sequence number.

With such a configuration, the RTP sequence number is calculated at the "second boundary", and the calculated RTP sequence number is set at the start time of the video frame. After that, the RTP sequence number in which 1 is sequentially added until the next "second boundary" is set for the IP packet to be continuously transmitted until the next second boundary. With such a configuration, the transmitters 13a and 13b can independently set the same sequence number for IP packets having the same payload without matching the RTP sequence numbers with each other.

In the present embodiment, the timing for calculating the sampling value and setting the RTP sequence number is set to every second, that is, at the "second boundary", but this is only an example, for example, 0.5. It may be a predetermined time interval such as seconds or 2 seconds. In addition, instead of periodically calculating the sampling value at time intervals and setting the RTP sequence number, it is manually performed via the GUI or command line input via the control network interface unit 139 of the transmitters 13a and 13b. It is also possible to calculate the sampling value based on the command by the software program or the command by the software program and set the RTP sequence number.

The transmission unit 137 transmits the IP packet generated by the IP packet generation unit 132 to the IP network 15 via the data network interface unit 138 in packet units.

In this way, each of the transmitters 13a and 13b sets the same RTP sequence number for the IP packet having the same payload without matching the RTP sequence numbers with each other, and sends the IP packet to the receiver 14. Can be sent.

When the receiver 14 detects that a failure has occurred in the working system, it switches from the IP packet A transmitted via the transmission path 16 to the IP packet B transmitted via the transmission path 17. When this switching is performed, each IP packet having the same payload is determined from the values of the RTP sequence numbers set in IP packet A and IP packet B according to SMPTE 2077-7 as in the conventional case, and the buffer is buffered inside the receiver 14. By ringing, the timing of both IP packets having the same payload is adjusted so that the reconstructed video signal does not experience a momentary interruption (hit) when switching occurs. (It is generally called seamless or hitless that no momentary interruption (hit) occurs.) The details of the process of reconstructing the video signal from the IP packet of the receiver are the same as before according to the specifications of SMPTE2022-7. Since it can be easily realized based on the above technique, detailed description in this specification is omitted.

In the first embodiment, an example in which a failure occurs in the transmission path of the active system and the system is switched to the standby system has been described, but the present invention is not limited to such a case. For example, a case where a failure has not occurred in the transmission path of the active system (that is, the transmission path is not switched) but a part of the IP packet A is lost (lost) is also described in the first embodiment. The above method can be applied. In this case, among the IP packets B transmitted via the transmission path 17, the IP packet B having the same payload as the lost IP packet A can be identified based on the RTP sequence number. The identified IP packet B is used as a substitute for the lost IP packet A. Since it can be identified as an IP packet having the same payload as the lost IP packet, the video signal can be seamlessly reconstructed by delaying a certain period of time by buffering (hereinafter, the second embodiment). The same applies).

<Second embodiment>
Next, a second embodiment of the present invention will be described with reference to FIG. FIG. 6 shows an example of a transmitter according to the second embodiment. In the second embodiment, the transmitter 13a issues an initialization command to the transmitter 13b at a predetermined timing, and the transmitter 13b responds to the initialization command and performs an RTP sequence at the same timing as the transmitter 13a. Initialize the number.

The video transmission system according to the second embodiment is the same as the components according to the first embodiment in terms of components other than the transmitter. As shown in FIG. 6, the transmitter according to the second embodiment includes an initialization command unit 140 in place of the PTP control unit 133 and the PTP time stamp counter 134 included in the transmitter according to the first embodiment. ..

When the video frame boundary is determined by the frame boundary determination unit 135, the initialization command unit 140 issues an initialization command to another transmitter (transmitter 13b in the case of the transmitter 13a) after waiting for a certain period of time. To do. The initialization instruction may be issued using, for example, a dedicated UDP packet or MAC frame. Alternatively, a TCP / IP session may be established between the two transmitters 13a and 13b and issued using a predetermined message. It is self-evident that it can be easily realized by using other conventional communication technologies. Here, a specific method for setting the RTP sequence number will be described with reference to FIG. 7.

FIG. 7 shows the relationship between the frame boundary of the transmitted IP packet and the time with the horizontal axis as the time axis. The upper part of FIG. 7 shows the processing of the transmitter 13a, and the lower part of FIG. 7 shows the processing of the transmitter 13b.

When it is necessary to match the RTP sequence number between the transmitters 13a and 13b, the frame boundary determination unit 135 determines the video frame boundary (the boundary between the video frame 1 and the video frame 2 in FIG. 7) in the transmitter 13a. Then, the initialization command unit 140 waits for a predetermined time (initialization command issuance prohibition time), and then issues the initialization command to the transmitter 13b.

In the transmitter 13a, after the initialization command issuance prohibition period has elapsed (or after the initialization command is issued), when the frame boundary determination unit 135 determines the next video frame boundary, the initialization instruction unit 140 Initializes the RTP sequence number to a predetermined number. On the other hand, when the transmitter 13b receives the initialization command from the transmitter 13a and the frame boundary determination unit 135 determines the next video frame boundary, the initialization command unit 140 of the transmitter 13b has an RTP sequence number. Is initialized to the specified number. Here, the predetermined numbers used for initialization in the transmitter 13a and the transmitter 13b are the same numbers.

In the present embodiment, the reason why the initialization command unit 140 of the transmitter 13a waits for a predetermined time (initialization command issuance prohibition time) until the issuance of the initialization command is to the transmitter 13b immediately after determining the video frame boundary. This is because when the initialization command is issued, the video frame boundary in which the transmitter 13b initializes the RTP sequence number may be erroneous depending on the delay time difference between the video signals input to the transmitter 13a and the transmitter 13b. For example, in FIG. 7, when the transmitter 13b detects the boundary between the video frame 1 and the video frame 2 later than the reception of the initialization command, the target for which the RTP sequence number is initialized is the first IP of the video frame 3. It is not a packet but the first IP packet of the video frame 2, and does not match the transmitter 13a. This event can be avoided by setting the initialization command issuance prohibition time of the transmitter 13a to be larger than the delay time difference between the input video signals of the transmitter 13a and the transmitter 13b.

In this way, the transmitters 13a and 13b can initialize the RTP sequence number at the same timing immediately after determining the next same video frame boundary. After that, the RTP sequence number to which 1 is sequentially added is set for the IP packet to be continuously transmitted. Thereby, the transmitters 13a and 13b can set the same sequence number for the IP packets having the same payload, respectively.

In the second embodiment, the RTP sequence number can be matched without mounting the PTP only by issuing the initialization instruction to the transmitter 13b at a fixed timing by the transmitter 13a. Therefore, the transmitters 13a and 13b can be realized with a simple configuration as compared with the method in the first embodiment.

In the present embodiment, the predetermined period from the determination of the boundary of the video frame to the boundary of the next video frame is set as the initialization command issuance prohibition period, but the present invention is not limited to such an example. For example, after determining the boundary of a video frame, it may wait for an initialization command to be issued until a predetermined number of IP packets are transmitted.

As described above, the video transmission system according to the present invention has been described. It should be noted that the processing performed by each component described above and the order of the processing are exemplary.

A IP packet B IP packet 16 Transmission path 17 Transmission path

Claims (6)

With PTP Grand Master,
A first transmitter that synchronizes time with the PTP ground master according to PTP,
A second transmitter that synchronizes time with the PTP ground master according to PTP,
The first transmitter and the second transmitter are provided with a receiver connected via an IP network.
Each of the first transmitter and the second transmitter receives a video signal distributed from one video signal as an input.
The input video signal is divided into a plurality of IP packets, and the input video signal is divided into a plurality of IP packets.
The initial value of the RTP sequence number is calculated based on the PTP time stamp time-synchronized according to the PTP.
A sequence number consecutive from the initial value is set in each RTP sequence number field of the plurality of divided IP packets.
The first transmitter transmits the plurality of IP packets in which the RTP sequence number is set as the first plurality of IP packets to the receiver via the first transmission path according to the RTP.
Said second transmitter, said plurality of IP packets RTP sequence number is set as the second plurality of IP packets, and transmitted in accordance with RTP, to the receiver via a second transmission path, wherein The first plurality of IP packets and the second plurality of IP packets have the same payload.
A video transmission system characterized by this. Each of the first transmitter and the second transmitter
Obtain a PTP time stamp at a predetermined cycle and
The video transmission system according to claim 1, wherein the initial value of the RTP sequence number is calculated from the acquired time stamp value. Each of the first transmitter and the second transmitter
The video frame boundary is detected from the input video signal,
The video transmission system according to claim 1 or 2, wherein the calculated initial value is set in the RTP sequence number field of the IP packet at a predetermined position based on the detected video frame boundary. The video signal distributed from one video signal is used as an input, the input video signal is divided into a first plurality of IP packets, and the divided first plurality of IP packets are transmitted via a first transmission path. The first transmitter to transmit
The distributed video signal is used as an input, the input video signal is divided into a second plurality of IP packets, and the divided second plurality of IP packets are transmitted via the second transmission path. The second transmitter and the second transmitter , wherein the first plurality of IP packets and the second plurality of IP packets have the same payload .
The first transmitter and the second transmitter are provided with a receiver connected via an IP network.
The first transmitter sets a continuous sequence number as a first RTP sequence number in each RTP sequence number field of the divided first plurality of IP packets.
The first plurality of IP packets in which the first RTP sequence number is set are transmitted to the receiver via the first transmission path according to the RTP.
An initialization command is issued to the second transmitter at a predetermined timing,
Initialize the first RTP sequence number and
The second transmitter sets a continuous sequence number as a second RTP sequence number in each RTP sequence number field of the divided second plurality of IP packets.
The second plurality of IP packets in which the second RTP sequence number is set are transmitted to the receiver via the second transmission path according to the RTP.
Upon receiving the issued initialization instruction,
A video transmission system characterized in that the second RTP sequence number is initialized in response to receiving the initialization instruction. The first transmitter determines the first video frame boundary from the input video signal, and determines the boundary of the first video frame.
The video transmission system according to claim 4, wherein the initialization command is issued after a predetermined time has elapsed from the determined first video frame boundary. The first transmitter determines the Nth video frame boundary from the input video signal by counting from the determined first video frame boundary.
The first RTP sequence number is initialized based on the determined Nth video frame boundary.
The second transmitter is
In response to receiving the initialization command, the second video frame boundary is determined from the input video signal, and the second RTP sequence number is assigned based on the determined second video frame boundary. The video transmission system according to claim 5, wherein the video transmission system is initialized.

Images