JPH1118051A - I-frame extract method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To prevent a bit rate and a processing load in the case of assembling composed coded video data into packets from being increased in the method for extracting packet stream data, consisting only of I-frames. SOLUTION: A video signal is compressed-coded, based on the MPEG protocol and packet stream data 4 divided data including obtained bit streams into optional packets are stored on a magnetic disk 33A. In the case of processing packets, address information of I-frames which configure one picture of video signals and address information at a head packet of a packet stream, including the I-frames and address information in a final packet of the packet stream including the I frames are described in an access table 5. The access table 5 is recorded on a magnetic disk 33B and packet stream data including a desired I-frame are extracted from the packet stream data 4, based on the access table 5.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

[0001] The present invention relates to, for example, MPEG2
Video-on-demand (VOD) systems that deliver compressed video over networks
The present invention relates to an I-frame extraction method in a video server device and a terminal device used for a library system or the like.

[0002]

2. Description of the Related Art In a video distribution apparatus such as a video server, a video signal encoding method based on the MPEG standard is often used. MPEG basically performs inter-frame predictive coding. However, by periodically performing intra-frame coding, it becomes possible to reproduce a program from the middle of a program. In addition, by selecting and extracting frame data by intra-frame encoding, and connecting these frames for continuous reproduction, high-speed reproduction of pseudo video can be realized.
Hereinafter, the I frame indicates frame data obtained by the intra-frame encoding.

A conventional I frame extraction method is disclosed in
What is described in 195925 gazette is known.
This conventional I frame extraction method, as shown in FIG.
The video signal to be stored is input to the encoding device 31 and compression-encoded. The bit string output from the encoding device 31 is given to the packetizing device 32, distributed and stored in an arbitrary packet, and packetized. The packetized data is the first
Is recorded on the magnetic disk 33A. In the process of packetization, an access table 35 is generated and recorded on the second magnetic disk 33B.

FIG. 6 is a model diagram when a bit string is stored in a packet. In this figure, the bit string is an encoding device.
31 is a bit string to be output. To packetize the data of the bit string, data is cut out from the bit string within a predetermined size and stored in each packet. The packetizer 32 prevents bit strings of different frames from being stored in the same packet. For example, if a bit string is stored only halfway in a packet, the rest of the data has no effect on decoding, such as Stuffing byt
The packet is completed by storing and supplementing e (0xff).

When it is necessary to extract an I-frame such as for special reproduction, the access table management unit 37 reads out the access table 35 managed by itself from the magnetic disk 33B, and reads out the packet sequence data 34 while referring to it. Determine the location of the data to be done. The stream reading unit 36
Packet sequence data 34 stored on the magnetic disk 33A
, The data at the position determined by the access table management unit 37 is read out.

In this device, an input video signal is converted to an MP signal.
The encoding device 31 performs compression encoding based on the EG rules and outputs a bit string. The packetizer 32 that has received the bit string distributes and stores the I frame and other frames in different packets, and records the packets on the magnetic disk 33A. Also,
In the process of packetization, an access table 35 indicating the position of a packet storing a bit string composed of only I frames is recorded on the magnetic disk 33B. When a video stream is transmitted from the video server 41 or when a video is played back on a terminal, when playing back the content (information) from the middle or fast-forwarding such as fast-forwarding, the access table 35 is referred to. The corresponding packet data including the frame can be read efficiently.

[0007]

In the conventional I-frame extraction method, before recording on a magnetic disk, it was necessary to distribute and store the packets for each frame in advance. . In the case of packetization with a fixed length, in order to store and distribute each frame, it is necessary to compensate for the packet length with a bit string that is not related to decoding. It is apt to increase, and solving this problem has been a problem.

[0008] The present invention takes into account such problems of the conventional method, and is more adaptable to general-purpose packetization based on the MPEG standard than the conventional method, and does not cause an increase in bit rate. It is an object of the present invention to provide a method for extracting a frame.

[0009]

According to the first aspect of the present invention, data including a bit sequence obtained by compressing and encoding a video signal based on the MPEG standard is generated, divided into arbitrary packets, packetized, and stored in a recording unit. Record. At the time of packetization, address information of an I frame constituting one screen of the video signal, address information of a first packet of a packet sequence including the I frame, and address information of a last packet of a packet sequence including the I frame Is created, and the access table is recorded on recording means different from the recording means. When reading out the packet sequence data from the recording unit that records the packet sequence, the packet sequence data including the desired I frame is extracted based on the access table, and the bit sequence in the extracted packet sequence data is edited. Thus, packet sequence data composed only of I frames is extracted.

According to the second aspect of the present invention, the video signal is
After generating data including a bit string obtained by compression encoding based on the EG rules and dividing it into arbitrary packets, MP
A packet stream output after being system-encoded based on the EG protocol is recorded in a recording unit. At the time of system encoding, an access table that describes address information of I frames constituting one screen of the video signal and information of bit strings other than I frames mixed with the I frames in a packet stream is created. The access table is recorded on recording means different from the recording means. When the packet stream is read from the recording unit that records the packet stream, a packet stream including a desired I frame is extracted based on the access table, and a bit string in the extracted packet stream is edited. A packet stream composed only of frames is extracted.

According to a third aspect of the present invention, in the method for extracting an I frame according to the second aspect, the packet stream composed of only the extracted I frame is received, and the packet stream is obtained when the packet stream is system-decoded. When the bit string is compression-encoded data based on the MPEG standard, the compression-encoded data is decoded based on the MPEG standard.

[0012]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of an I-frame extraction method in a video server device and a terminal device according to the present invention will be described below in detail with reference to the drawings.

Embodiment 1 FIG. 1 is a block diagram showing a configuration of a video server device 11 for implementing an I-frame extraction method according to Embodiment 1 of the present invention.

As shown in FIG.
Given to. The encoding device 1 is a device that compresses and encodes a video signal based on the MPEG standard. The encoded MPEG data is arranged in a bit string and input to the packetizer 2. The packetizing device 2 is a device that inputs a bit string of video data and performs packetization. At the time of packetization, the packetizer performs packetization without being affected by bit string information, and at the same time, in the packet string, the address information of the I frame constituting one screen of the video signal, and the beginning of the packet string including the I frame. An access table 5 describing address information in a packet and address information in the last packet of a packet sequence including an I frame is created. A packet sequence of video data is recorded for each program in the magnetic disk device 33A, and an access table 5 is recorded in the magnetic disk device 33B.

FIG. 2 is an explanatory diagram showing a method of storing a bit string in a packet and a format of the access table 5. In FIG. 2A, the bit string is a coding device.
1 is a bit string to be output. To packetize the data of the bit string, the data is cut out from the bit string within a predetermined size and stored in each packet, and FIG.
Create the packet sequence shown in. An access table 5 having data indicating the position of a packet sequence created in the process of packetization is described for each I frame. The information in the access table of the n-th I frame includes IFP
(Head address of I frame), IDL (data length of I frame), FP (head packet address), FPH
(Head packet length), FPP (head packet length), LP (last packet address), LPP (last packet length).

Video data encoded based on the MPEG standard can be reproduced from the middle only from a bit string including an I frame. When it is necessary to extract an I frame corresponding to the video data to be read, such as in trick play, the access table management unit 7 reads the access table 5 managed by itself from the magnetic disk 33B, The position to be read in the column data 4 is determined. The stream reading unit 6 reads a bit string in the packet including the corresponding I frame stored on the magnetic disk 33A based on the data at the position determined by the access table management unit 7.

The stream editing unit 8 calculates the following two equations for a packet containing data other than the I frame at the leading end of the I frame based on the information in the access table 5.

[Address of data (1) other than I frame] = FP + FPH

[Length of data (1) other than I frame] =
IFP- (FP + FPH)

The stream editing unit 8 performs the following two calculations on a packet containing data other than the I frame at the end of the I frame.

[Address of data (2) other than I frame] = IFP + IDL

[Length of data (2) other than I frame] =
(LP + LPP)-(IFP + IDL)

When it is necessary to extract an I frame such as trick play based on the above calculation results, the access table management section 7 stores the access table 5 shown in FIG. The position to be read out in the packet sequence data 4 is determined by reading from 33B and referring to it. The stream reading unit 6 is a magnetic disk 33A
The data at the position determined by the access table management unit 7 is read out from the packet sequence data 4 stored in the stream table 4, and is output to the stream editing unit 8. The stream editing unit 8 converts the data other than the I frame in the received data into a Stuffing byte (0xff) that does not affect decoding.
Thus, the output unit 9 outputs a packet sequence. FIG. 2D is a model diagram of a packet sequence output from the output unit 9. By decoding this packet sequence, an I frame corresponding to the video data to be read can be extracted.

According to the first embodiment, at the time of packetization, access information describing the address information of the I-frame, the address information of the first packet of the packet sequence including the I-frame, and the address information of the last packet are described. By providing the table, there is no need to distribute and store the I frame and other frames in different packets, and the I frame can be divided into arbitrary packets and stored. At the time of reading, the I frame can be extracted based on the address information of the I frame, the address information of the first packet, and the address information of the last packet.

Second Embodiment Next, a second embodiment for realizing the I-frame extraction method according to the present invention will be described with reference to FIG.

FIG. 3 shows a configuration of the video server device 13 and the terminal device 14 according to the second embodiment of the present invention.

As shown in the figure, an encoding device 1 is a device for compressing and encoding a video signal based on the MPEG standard.
The encoded MPEG data is arranged in a bit string and input to the packetizer 52. Packetizer 52
Is a device for inputting a bit sequence of video data and performing packetization. At the time of packetization, the packetizer 52 performs packetization without being affected by the information of the bit string,
The packet sequence shown in FIG. The system encoder 510 receives the packet sequence, performs system encoding based on the MPEG standard, and outputs a packet stream. Here, the system coding refers to time-division multiplexing of individually coded video, audio, and other coded bit strings, including synchronization, into one data string. At the time of system encoding, the system encoder 510 stores an access table 55 describing address information of an I frame constituting one screen of the video signal and bit string information other than the I frame mixed with the I frame in the packet stream. create. Magnetic disk unit 33A
Records the video data packet sequence for each program,
The access table 55 is recorded in the magnetic disk device 33B.

As the types of system encoding based on the MPEG2 rules, two types of program stream (PS) and transport stream (TS) are defined.

The TS packet shown in FIG. 4B is an output format in a case where packet sequence data is system-encoded into a transport stream in the system encoder 510. . The packet sequence data generated by the packetizer 52 is stored in a fixed-length (188 byte) TS packet. TS packet shown in FIG.
(A) and the TS packet (B) shown in FIG.
5 shows a stream format in the output unit 59 when delivering from a video server to a terminal. System encoder 51
The format of the access table 55 created with 0 is:
The output unit 59 has a different table format when outputting the TS packet (A) and when outputting the TS packet (B).

When the output unit 59 outputs the TS packet (A), information on the TS packet including the I frame and address information on bit strings other than the I frame mixed in the TS packet are described.

When the output unit 59 outputs the TS packet (B), the information of the TS packet including the I frame and the packet header and the address information of the bit string other than the I frame and the packet header are described. , The information of the packet length described in the packet header based on the MPEG standard is set to '0'.

At the request of the terminal device, etc., I
When frame extraction is necessary, the access table management unit 57
However, the access table 55 managed by the
The position to be read in the packet stream 54 is determined by reading from 33B and referring to it. The stream reading unit 56 reads the access table management unit 57 from the packet sequence data 54 stored in the magnetic disk 33A.
The data at the position determined by is read out and output to the stream editing unit 58. In accordance with the information in the access table 55, the stream editing unit 58 causes the output unit 59 to use a stream format such as the TS packet (A) in FIG. 4C or the TS packet (B) in FIG. 4D. Edit to

The stream output from the video server device is input to the system decoder 511 of the terminal device 14. The system decoder 511 performs system decoding,
The data is output to the switch 512 of the MPEG decoder 500.

If the stream format input to the system decoder 511 is the TS packet (A) shown in FIG. 4C, the stream passes through the path 514 by the switch 512 and is decoded. Input to device 515. In the case of a TS packet (B), it is input to the packet decoding device 513 via the switch 512.

The packet decoding device 513 mainly analyzes and deletes a packet header. Since the data input to the decoding device 515 does not include a bit string other than the I frame, decoding this data enables extraction of an I frame corresponding to the video data to be read.

According to the second embodiment, when a plurality of types of data such as video and audio are system-coded, the address information of the I-frame and the information other than the I-frame mixed with the I-frame in the packet stream are used. By providing an access table in which bit string information is described, packetization can be performed without being affected by bit string information in bucketing. At the time of reading, an I frame can be extracted by selecting a bit sequence of the I frame based on the information in the access table.

[0037]

According to the first embodiment of the present invention, at the time of packetization, access information describing the address information of the I-frame, the address information of the first packet of the packet sequence including the I-frame, and the address information of the last packet are described. By providing the table, there is no need to distribute and store the I frame and other frames in different packets, and the I frame can be divided into arbitrary packets and stored. At the time of reading, the I frame can be extracted based on the address information of the I frame, the address information of the first packet, and the address information of the last packet. As a result, it is possible to reduce the processing load when packetizing the compressed and encoded data of the video data, and to perform the I frame extraction without increasing the bit rate.

According to the second embodiment, when a plurality of types of data such as video and audio are system-encoded, the address information of the I frame and the bit string of the bit stream other than the I frame mixed with the I frame in the packet stream are used. By providing an access table in which information is described, packetization can be performed without being affected by bit string information in bucketing. At the time of reading, an I frame can be extracted by selecting a bit sequence of the I frame based on the information in the access table.

[Brief description of the drawings]

FIG. 1 is a block diagram illustrating a video server device that implements an I-frame extraction method according to a first embodiment of the present invention.

FIG. 2 is a diagram illustrating a format model of a bit sequence, a packet sequence, an access table for trick play, and an edited packet sequence in the first embodiment of the present invention.

FIG. 3 is a block diagram showing a video server device and a terminal device for realizing an I-frame extraction method according to a second embodiment of the present invention;

FIG. 4 is a diagram illustrating a format model of a packet sequence, a TS packet, and a TS packet (A) or (B) transmitted to a terminal according to the second embodiment of the present invention.

FIG. 5 is a block diagram showing a video server device for realizing a conventional I-frame extraction method.

FIG. 6 is a model diagram of a bit sequence and a packet sequence in a conventional method.

[Explanation of symbols]

 1,31 Encoding device 33A, 33B Recording means 2,32,52 Packetizer 4,34 Packet sequence data 5,35,55 Access table 6,36,56 Stream reading unit 7,37,57 Access table management unit 8 , 58 Stream editing unit 9,39,59 Output unit 54 Packet stream 500 MPEG decoder 510 System encoder 511 System decoder 512 Switch 513 Packet decoding device 514 Path 515 Decoding device

Claims (3)

[Claims] A step of generating a video signal into data including a bit string obtained by compression-encoding based on the MPEG standard; a step of dividing the data into arbitrary packets and recording the packets in a first recording unit; An access table describing address information of an I-frame constituting one screen of a video signal, address information in a first packet of a packet sequence including the I-frame, and address information in a last packet of a packet sequence including the I-frame is described. Creating the access table; recording the access table in a second recording device different from the first recording device; and reading a packet sequence data from the recording device in which the packet sequence is recorded. Packet sequence data including a frame is extracted based on the access table, and the extracted packet By editing the bit sequence in the preparative column data, I frame extracting method characterized by comprising the steps of: extracting a packet sequence data composed of only I-frames. 2. A step of generating a video signal including a bit string obtained by compression-encoding a video signal based on the MPEG standard, dividing the data into arbitrary packets, and performing system encoding based on the MPEG standard and outputting the packet. Recording a packet stream in a first recording unit; address information of an I frame constituting one screen of the video signal; and information of a bit string other than the I frame mixed with the I frame in the packet stream. Creating an access table; recording the access table on a second recording device different from the first recording device; and reading a packet stream from the first recording device recording the packet stream. , A packet stream including a desired I frame into the access table. Zui extracted by editing the bit string in said extracted packet stream, I frame extracting method characterized by comprising the steps of: extracting a packet stream consisting only of I-frames. 3. When the bit stream obtained when the packet stream composed of only the extracted I frames is system-decoded is compression-encoded data according to the MPEG standard, the compression-encoded data is converted into the MPEG standard. 3. The I-frame extraction method according to claim 2, further comprising the step of decoding based on the I-frame.