JP2013062738A - Method for determining frame type of video packet, and monitoring server - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a method for determining a frame type of a video packet, and a monitoring server, capable of securely determining a frame type of a video packet subjected to high efficiency compression encoding from a packet string being transmitted on a network.SOLUTION: A method for determining a frame type of a video packet, includes the steps of: detecting frame start information from a header of a video packet being transmitted on a network (detecting step); when the frame start information detected at the detecting step indicates a frame start portion, counting the video packet being transmitted on the network to generate a count value until the frame start information indicating the frame start portion is detected next at the detecting step (counting step); storing the count value of the video packet in a storage unit each time the count value of the video packet is generated by the counting step (storing step); and determining a frame type according to an increase/decrease pattern of the count value of the video packet stored in the storage unit, and the count value of the video packet each time the count value of the video packet is generated by the counting step (determining step).

Description

  The present invention relates to a frame type determination method and a monitoring server for determining the frame type of a video packet transmitted to a user terminal via a network.

  There is a large-scale video distribution service that distributes video such as TV video to user terminals using a network such as the Internet. In such a video distribution service, a video signal is subjected to high-efficiency compression coding, and the high-efficiency compression-encoded video signal is converted into a packet sequence and transmitted to a user terminal via a network.

  In high-efficiency compression coding of video signals, for example, an I (Intra Picture) frame that performs coding processing within a frame without using inter-frame prediction, P that performs coding processing by predicting the present from the past using motion compensation prediction (Predictive Picture) frames, forward and backward prediction using bi-directional prediction and encoding processing (Bidirectionally Predictive Picture) Three types of frames called B frames are usually 15 frames as one group (GOP: Group of Picture) Generated. 1 GOP is a group of pictures that can be decoded on the basis of an I frame.

  When a packet including video data in such a plurality of types of frames is transmitted over the network, packet loss may occur in the network. When packet loss occurs, the frame information before and after is used for decoding at the user terminal, so that the deterioration of the video quality due to the loss of one packet may extend not only to one video frame but also to a plurality of video frames.

  Thus, it has been proposed to estimate video quality due to packet loss. As an estimation method, for each video packet sent to the user terminal, the frame type (I, P, B) and the frame generation rule (I frame repetition period and P frame including the I frame) at the time of high efficiency compression encoding to which the packet belongs. In the user terminal, either directly from the frame type and the frame generation rule embedded in the lost packet or indirectly from the frame type and the frame generation rule embedded in the video packet before and after the lost packet. In particular, the quality of the decoded video (quality corresponding to the user's experience quality) is estimated by detecting the frame type and the frame generation rule at the time of high-efficiency compression encoding to which the lost packet belongs (Patent Literature). 1).

JP 2006-33722 A

  DRM (Digital Rights Management) or the like in the payload in an environment where video packets including a video signal that includes high-efficiency compression-encoded video signals in an IP network such as the Internet are transmitted, such as IP (Internet Protocol) retransmission of terrestrial digital broadcasting Therefore, it is impossible to collect and analyze information such as a frame type from the payload by capturing a packet on the IP network for the above-described video quality. Also, the key for decrypting the encryption is managed by the content provider or broadcaster that provides the video signal source, and decryption is performed on the IP network where packets having various contents are transmitted. It is difficult.

  Accordingly, an object of the present invention is to provide a frame type determination method and a monitoring server that can reliably perform the frame type of a video packet that has been subjected to high-efficiency compression encoding from a packet sequence being transmitted over a network. is there.

  In the frame type determination method of the present invention, a video signal is subjected to high-efficiency compression coding to be a sequence of frames of a plurality of types in accordance with a specific frame generation rule, and the payload includes the data of each frame in the sequence of frames. And a frame type for determining a frame type of data in the video packet from the network in a communication system in which video packets including frame start information in the header indicating whether or not the data is a frame start portion are sequentially transmitted over the network. A detection method for detecting the frame start information from a header of the video packet transmitted through the network, and when the frame start information detected in the detection step indicates the frame start portion, Indicates the frame start part in the detection step A counting step of counting the video packets transmitted through the network until the next frame start information is detected and generating as a count value, and each time a count value of the video packet is generated by the counting step A storage step of storing the count value of the video packet in a storage unit, an increase / decrease pattern of the count value of the video packet stored in the storage unit, and a count value of the video packet generated by the counting step And a determination step of determining the frame type according to the count value of the video packet.

  According to the monitoring server of the present invention, the video signal is subjected to high-efficiency compression coding to be a sequence of frames of a plurality of types in accordance with a specific frame generation rule, and the payload includes the data of each frame of the sequence of frames. A monitoring server that determines a frame type of data in the video packet from the network in a communication system in which video packets including frame start information in the header indicating whether the data is a frame start portion are sequentially transmitted over the network. Detecting means for detecting the frame start information from a header of the video packet transmitted through the network, and when the frame start information detected by the detection means indicates the frame start portion, The frame start information indicating the frame start portion is next detected. Measuring means for counting the video packets transmitted through the network and generating a count value, and storing the count value of the video packet every time the count value of the video packet is generated by the measuring means. Storage management means for storing in the storage unit, an increase / decrease pattern of the count value of the video packet stored in the storage unit, and a count value of the video packet each time the count value of the video packet is generated by the measurement means, And determining means for determining the frame type according to the above.

  According to the frame type determination method of the present invention and the monitoring server of the present invention, the frame start information is detected from the header of the video packet transmitted over the network, and the detected frame start information indicates the frame start portion. The video packet transmitted over the network is counted until the next frame start information indicating the portion is detected, and the count value of the video packet is stored in the storage unit every time the count value of the video packet is generated. The frame type is determined according to the increase / decrease pattern of the count value of the video packet stored in the unit and the count value of the video packet every time the count value of the video packet is generated. Therefore, even if the payload data of the video packet is encrypted, the frame type of the data in the transmitted video packet can be reliably determined. Further, there is an effect that it is not necessary to embed information on the frame type and the frame generation rule at the time of high efficiency compression encoding to which each packet belongs, as in the prior art.

1 is a block diagram showing a communication system to which a frame type determination method of the present invention is applied. It is a figure which shows the structure of a RTP packet. It is a block diagram which shows the structure of the monitoring server in the communication system of FIG. It is a figure explaining from an elementary stream to TS packet. FIG. 5 is a diagram illustrating a structure of each of the elementary stream, the PES packet, and the TS packet in FIG. 4. It is a figure which shows the specific structure of TS packet. It is a flowchart which shows operation | movement of the measurement part in the monitoring server of FIG. It is a flowchart which shows operation | movement of the frame classification determination part in the monitoring server of FIG. It is a figure which shows the count packet number of I frame, P frame, and each B frame. It is a figure which shows the change of the count value of the count value sequence preserve | saved at the database part.

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

  FIG. 1 shows a communication system to which the frame type determination method of the present invention is applied. This communication system is a system that performs video distribution including a video distribution server 1, nodes 2 to 5, a user terminal 6, and a monitoring server 7. The video distribution server 1 and the user terminal 6 are connected to the IP network 8, and the nodes 2 to 5 exist in the IP network 8. The video distribution server 1 is a server that distributes at least user terminal 6 with RTP packets including video content in TS (Transport Stream) packets. As shown in FIG. 2, the RTP packet includes an IP header, a UDP header, an RTP header, and an RTP payload, and is defined by RFC1889. The TS packet is included as a packet string in the RTP payload.

  The video content is provided by the content provider 9 via a dedicated line 10 or a recording medium (not shown) such as a DVD. Nodes 2 to 5 are computers, routers, or hubs that enable communication between the video distribution server 1 and the user terminal 6. The user terminal 6 is a television receiver or personal computer that has a communication function and can decode and display data in the TS packet of the payload of the RTP packet. The monitoring server 7 is connected to the node 4 and monitors packets passing through the node 4.

  As shown in FIG. 3, the monitoring server 7 includes an information collection unit 11, a measurement unit 12, a DB management unit 13, a frame type determination unit 14, and a database unit 15.

  The information collection unit 11 takes in the TS packet in the RTP packet from the node 4 using, for example, ssh, and collects information on the TS header of the TS packet. The measurement unit 12 determines whether PUSI (Payload Unit Start Indicator) of the TS header indicates 0 or 1, and if PUSI = 1, then the same PID (packet identifier) is used as the header. The number of TS packets having the same PID is counted until the PUSI of the TS packet having 1 becomes 1. PUSI = 1 indicates that the TS packet includes data from the start position of one frame, and PUSI = 0 indicates that the TS packet includes data other than the start position of the frame. The measurement unit 12 supplies the count value to the DB management unit 13. The DB management unit 13 writes the count value in the database unit 15. The count value is written together with the PID, and the order of writing is made known.

  When the number of samples of the count value written in the database unit 15 exceeds a predetermined number of samples (for example, 50), the frame type determination unit 14 determines the number of frames of 1 GOP based on the increase / decrease pattern of those count values. After determining the number of frames of 1 GOP, the frame type determination unit 14 classifies the frame according to the count value every time the measurement unit 12 generates the count value.

  When MPEG2 is used as the high-efficiency compression encoding method, as shown in FIG. 4, the elementary stream of the MPEG2 video signal is first converted into a variable-length PES packet, and then the PES packet is divided into a fixed length. It is created as a (188 byte) TS packet. In the RTP packet, a plurality of TS packets are arranged in a row in the RTP payload.

  An elementary stream of a video signal of MPEG2 is arranged together with SH (sequence header) as shown in FIG. 5 in one GOP unit consisting of 15 encoded frames. One GOP starts from an I frame (I picture frame), followed by two B frames (B picture frame) and one P frame (P picture frame).

  Each frame of the elementary stream is converted into a PES packet for each frame. The PES packet is a variable length packet including a PES header and PES data. The PES header includes information such as packet length, time stamp, scramble information, copyright information, CRC, and the like. PES data is data corresponding to one frame (I frame, B frame, or P frame).

  As shown in FIG. 6, the TS packet includes a 4-byte TS header and a 184-byte payload. The TS header includes a synchronization byte, a transport error indicator, PUSI (payload unit start indicator), transport priority, PID (packet identifier), transport scramble control, adaptation field control, and continuity index from the beginning of the packet. The numbers in parentheses in FIG. 6 are the number of bytes. The 184-byte payload is actually an adaptation field and / or payload, which may be either the adaptation field or the payload, or both the adaptation field and the payload.

  In the monitoring server 7 shown in FIG. 3, the information of the header of the TS packet captured by the information collecting unit 11 is supplied to the measuring unit 12 every time the TS packet is captured (detection step). As shown in FIG. 7, the measurement unit 12 determines whether the PUSI of the header having the PID indicating the video among the headers of the captured TS packet indicates logic 1 or 0 (step S11), and PUSI = If 1, the counting of TS packets having the same PID as the TS packet in which PUSI = 1 is detected starts from an initial value (for example, 0) (step S12). The TS packet can be counted for each unit information of the header supplied from the information collecting unit 11 to the measuring unit 12. After starting TS packet counting, the measuring unit 12 determines whether PUSI = 1 in the TS packet having the same PID in the header according to the header information sequentially supplied from the information collecting unit 11 (step S13). . Even for TS packets having the same PID in the header, if PUSI = 0, only counting of TS packets is continued. On the other hand, if the TS packets have the same PID in the header and PUSI = 1, the measurement unit 12 provides the count value of the TS packet at that time to the DB management unit 13 together with the PID (step S14). Steps S11 to S12 correspond to a counting step.

  The count value of the TS packet provided to the DB management unit 13 corresponds to the length of one frame. Then, returning to step S12, counting of TS packets having the same PID is started from the initial value. If the determination result in step S13 is PUSI = 0, TS packet counting is continued.

  When the count value and the PID are provided from the measurement unit 12 to the DB management unit 13 in step S14, the count value and the PID are written into the database unit 15 by the DB management unit 13 (storage step). By this writing, the counter value of the same PID is sequentially stored in the database unit 15.

  On the other hand, when the number of samples of the count value of the same PID written in the database unit 15 exceeds a predetermined number of samples, the frame type determination unit 14 obtains the count value from the database unit 15 in the order of writing as shown in FIG. Reading (step S21), the frame structure of 1 GOP is determined from the increment / decrement pattern of the count value (step S22). The number of samples of the same PID count value written in the database unit 15 can be obtained from the DB management unit 13. For example, when the count value peak occurs every 15 samples, 15 frames are set to 1 GOP. When the count value peak occurs every 30 samples, 30 frames are determined. Is determined to be 1 GOP.

  More specifically, when 1 GOP consists of 15 frames, the frames are arranged in the order of frame generation rules, for example, IBBPBBPBBPBBPBB, when 1 GOP consists of 30 frames. For example, the rules are arranged in the order of IBBPBBPBBPBBPBBPBBPBBPBBPBBPBB. Here, I is an I frame, P is a P frame, and B is a B frame. As shown in FIG. 9, the number of TS buckets carrying data for one I frame is sufficiently larger than the number of TS packets carrying P frame data, and the number of TS buckets carrying P frame data is B frames. Since it is larger than the number of TS packets carrying data, the count value for the I frame appears as a peak. FIG. 10 exemplifies increase / decrease of the count value, that is, the number of TS packets when 1 GOP consists of 15 frames. As can be seen from FIG. 9, the count value for the I frame is generated at intervals of 15 count values. The peak of appears. Therefore, the frame structure including the number of frames of 1 GOP can be determined from the increase / decrease pattern of the count value, and the position of the I frame can be determined at the same time, so that the type of each subsequent frame can be determined.

  In addition, after the frame type determination unit 14 determines the number of frames of 1 GOP, when the measurement unit 12 generates a count value and the count value is supplied (step S23), the frame type determination unit 14 determines the number of frames according to the count value. The type is determined (step S24). Steps S21 to S24 correspond to determination steps.

  When it is determined that one GOP is composed of 15 frames or 30 frames, the count value supplied from the measurement unit 12 to the frame type determination unit 14 can be predicted. For example, the count value supplied by setting thresholds for discriminating each frame (threshold for discriminating between I frames and P frames, threshold for discriminating between P frames and B frames) from the increase / decrease pattern of the count value is I It is possible to determine whether the frame is a frame, a P frame, or a B frame and determine whether the determination result matches the predicted frame type. As the threshold setting, for example, an intermediate value between the average value of the count value for the I frame and the average value of the count value for the P frame is set as a threshold value for discriminating between the I frame and the P frame. An intermediate value between the average value of the count values and the average value of the count values for the B frame can be used as a threshold value for discriminating between the P frame and the B frame.

  As described above, when the frame determined by the count value newly supplied from the measurement unit 12 in step S24 does not match the predicted frame, it can be determined that the loss of the TS packet has occurred.

  Here, the influence of video quality deterioration when a loss of TS packets occurs will be described. As described above, there are three types of frames, i.e., I frame, B frame, and P frame, in the TS frame that carries video signal data. When a packet containing I frame data is lost, quality degradation propagates until a packet containing the next I frame arrives. When a packet containing P frame data is lost, quality degradation propagates until a packet containing the next I frame arrives. When one GOP is composed of 15 frames, the frame repeats the frame generation rule such as IBBPBBPBBPBBPBB as described above. Therefore, propagation of quality deterioration is particularly large when a packet including I frame data is lost. When a packet containing B frame data is lost, quality degradation does not propagate. Therefore, when a packet containing I-frame data is lost, the influence on the viewer due to quality degradation is the greatest, and when a packet containing P-frame data is lost, the influence on the viewer is the I-frame data. It is smaller than the case where the containing packet is lost.

  As described above, according to the frame type determination method shown in the embodiment, PUSI is detected as frame start information from the header of a TS packet transmitted over the network, and the detected PUSI is a logic 1 indicating a frame start portion. At a certain time, TS packets transmitted through the network are counted until the next PUSI = 1 is detected, and each time the TS packet count value is generated, the TS packet count value is stored in the database unit 15 as a storage unit. The frame type is determined in accordance with the increase / decrease pattern of the TS packet count value stored in the database unit 15 and the TS packet count value each time the TS packet count value is generated. Therefore, even if the data of the payload of the TS packet is encrypted, the frame type of the data in the transmitted TS packet can be reliably determined. Further, there is an effect that it is not necessary to embed information on the frame type and the frame generation rule at the time of high-efficiency compression encoding to which each packet belongs as in the prior art.

  In the above-described embodiment, the information collecting unit 11 constitutes detection means for detecting frame start information from the header of a video packet (TS packet) transmitted through the network. Further, the measurement unit 12 constitutes a measurement unit that counts video packets, and the DB management unit 13 corresponds to a storage management unit that stores a count value of video packets in the database unit 15 that is a storage unit. Corresponds to determination means for determining the frame type.

  In the above-described embodiment, the information collection unit 11 of the monitoring server 7 is configured by an input / output interface, and the measurement unit 12, the DB management unit 13, and the frame type determination unit 14 are configured by a computer. The operations of the measurement unit 12, the DB management unit 13, and the frame type determination unit 14 may be executed according to a program.

DESCRIPTION OF SYMBOLS 1 Video distribution server 2-5 Node 6 User terminal 7 Monitoring server 11 Information collection part 12 Measurement part 13 DB management part 14 Frame classification determination part 15 Database part

Claims (6)

The video signal is subjected to high-efficiency compression encoding to form a sequence of frames of a plurality of types according to a specific frame generation rule, each frame of the frame sequence includes data in the payload, and the data is a frame start portion. A frame type determination method for determining a frame type of data in the video packet from the network in a communication system in which video packets including a frame start information indicating whether or not the header is sequentially transmitted through the network,
A detection step of detecting the frame start information from a header of the video packet transmitted through the network;
When the frame start information detected in the detection step indicates the frame start portion, the video transmitted through the network until the frame start information indicating the frame start portion is next detected in the detection step A counting step of counting packets and generating a count value;
A storage step of storing the count value of the video packet in a storage unit each time the count value of the video packet is generated by the counting step;
The frame type is determined according to the increase / decrease pattern of the count value of the video packet stored in the storage unit and the count value of the video packet every time the count value of the video packet is generated by the counting step. A frame type determining method comprising: a determining step. The video packet is a TS (transport stream) packet,
The detection step detects a payload unit start indicator as the frame start information from the header of the TS packet,
2. The frame type determination method according to claim 1, wherein when the payload unit start indicator indicates logic 1, it corresponds to indicating the frame start portion.   3. The frame type determination method according to claim 1, wherein the determination step determines the frame type when a count value of the video packet stored in the storage unit exceeds a predetermined number of samples. The frame type at the time of the high-efficiency compression encoding is an I frame that performs encoding processing within a frame without using inter-frame prediction, a P frame that predicts current from the past using motion compensation prediction, and bidirectional processing. There are three types of B frames that use the prediction to perform forward and backward prediction and perform encoding processing.
The frame type determination method according to claim 1, wherein the frame generation rule is a generation rule based on the I frame, the P frame, and the B frame.   The determining step determines a frame structure of 1 GOP (group of pictures) based on the frame generation rule according to an increase / decrease pattern of the count value of the video packet, and after the determination, the count value of the video packet is determined by the counting step. 5. The frame type determination method according to claim 4, wherein the frame type is determined according to a count value of the video packet each time it is generated. The video signal is subjected to high-efficiency compression encoding to form a sequence of frames of a plurality of types according to a specific frame generation rule, each frame of the frame sequence includes data in the payload, and the data is a frame start portion. A monitoring server for determining a frame type of data in the video packet from the network in a communication system in which video packets including a frame start information indicating whether or not the header is sequentially transmitted through the network;
Detecting means for detecting the frame start information from a header of the video packet transmitted through the network;
When the frame start information detected by the detecting means indicates the frame start portion, the video transmitted through the network until the frame start information indicating the frame start portion is next detected in the detection step A measuring means for counting packets and generating a count value;
A storage management unit that stores the count value of the video packet in a storage unit every time the count value of the video packet is generated by the measurement unit;
The frame type is determined according to the increase / decrease pattern of the count value of the video packet stored in the storage unit and the count value of the video packet every time the count value of the video packet is generated by the measuring unit. And a determination unit.

Images