JPH09298749A - Moving image distributing method and executing device for the same - Google Patents

Abstract

PROBLEM TO BE SOLVED: To efficiently distribute moving image data by compressing intermediate form moving image data, which are previously stored in a storage device, to the amount of codes corresponding to the load of a network. SOLUTION: Plural pieces of moving image data already stored in a storage medium 107 are extended to a middle stage by a half extending means 1, and plural pieces of intermediate format moving image data 2 are prepared and stored in the storage device in advance. When there is the distribution request of video data from a client terminal, the intermediate format moving image data 2 stored in the storage device are compressed again and the video data are prepared and distributed. When distributing the video data to the client terminal, the transmission velocity of the network or a telephone line is detected by a network performance detecting means 3 and reported to a code amount control means 4. Corresponding to the state of the network or telephone line, the code amount control means 4 determines the optimum amount of codes of the intermediate format moving image data 2.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a moving image distribution method for distributing moving image data to a plurality of client terminals connected via a network, and an apparatus for implementing the same, and more particularly, to an intermediate form moving image data previously stored in a storage device. The present invention relates to a video distribution method for storing video data and distributing the video data created by changing the code amount of the video data of the intermediate form to the client terminal according to the load of the network, and a technique effective when applied to the implementation device. is there.

[0002]

2. Description of the Related Art In recent years, video and audio compression technology has advanced,
MPEG (Moving P), which is an international standard compression method
A video distribution system has been researched and developed, in which digital video data with audio compressed by image Experts Group) or the like is stored in a computer having a large-capacity storage device and distributed to a client terminal via a computer network or a public line. Therefore, a large demand is expected.

As a current technical problem of such a video distribution system, a video-on-demand system which enables quick and stable quality video distribution in response to video distribution requests from a plurality of client terminals. Can be realized.

However, even if the compression technology is advanced, the amount of compressed moving image data is still enormous, and if the moving image distribution requests from a plurality of client terminals are concentrated in a short time, an enormous amount of data stream will be generated. The data is read from the storage device and sent to the network, and the performance of reading the moving image data from the storage device and the limit of the transmission band of the network are reached to cause transmission delay and the like.

When the moving picture data distributed while receiving the transmission delay or the like is reproduced while being received, the moving picture which is interrupted or has awkward motion is reproduced, and the moving picture distribution requests from a plurality of client terminals are concentrated in a short time. In this case, it becomes difficult to convey smooth video information.

As a conventional technique for enabling stable video-on-demand with respect to these problems, a high-speed network is used, and moving image data is distributed and arranged in a plurality of storage devices of a server to improve network and storage device performance. A video distribution system that eliminates the neck has been proposed and developed, and various experiments have been repeated.

[0007] On the other hand, Japanese Patent Application No.
No. 118672 proposes a video delivery system that eliminates network performance bottleneck by re-compressing already-compressed moving image data by increasing the compression rate during delivery.

[0008] The outline is that, when a distribution request for moving image data is issued from a client terminal, the compressed moving image data is expanded into intermediate form moving image data, and the code amount is changed according to the load on the network. The moving image data of the intermediate form in which the code amount is changed is recompressed and delivered to the client terminal.

[0009]

The present inventor has found the following problems as a result of examining the above-mentioned prior art.

That is, of the above-mentioned conventional techniques, the former video distribution system utilizing the high-speed network to disperse and arrange moving image data in a plurality of storage devices of a server eliminates the bottleneck in the performance of the storage device and produces a huge amount of moving images. Although it is effective for stable distribution of data, there is a problem that the introduction cost becomes high because an expensive new network is installed and a plurality of storage devices for distributing and arranging moving image data are required.

Further, the latter video distribution system which eliminates the network performance bottleneck by increasing the compression rate at the time of distribution of the already-compressed data and eliminating the network performance bottleneck, is capable of distributing the moving image corresponding to the traffic even in the existing network. In Although,
A problem that the processing load increases because the process of decompressing the already-compressed data to the intermediate form moving image data and the process of recompressing the extended intermediate form moving image data are performed every time there is a distribution request from the client terminal. was there.

An object of the present invention is to provide a technique capable of efficiently distributing moving image data having a code amount according to the load on the network.

Another object of the present invention is to provide a technique capable of re-compressing moving image data without degrading the quality of the moving image data as much as possible.

The above and other objects and novel features of the present invention will become apparent from the description of the present specification and the accompanying drawings.

[0015]

Of the inventions disclosed by the present application, a representative one will be briefly described below.
It is as follows.

(1) Intermediate-format moving image data created by decompressing already-compressed moving image data to an intermediate decompressed state in an information processing device that distributes moving image data to a plurality of client terminals connected via a network. Of the intermediate form according to the network load detected by the network performance detecting means, the network performance detecting means for detecting the load of the network for transmitting the moving image data, Depending on the code amount determined by the code amount control unit, the code amount control unit that determines the code amount when re-compressing the moving image data and the intermediate form moving image data stored in the storage device in advance by the semi-expansion unit. And a recompression means for recompressing.

In the moving picture distribution server which is the information processing apparatus, first, the half-expansion means expands the already-compressed moving picture data to an intermediate expanded state and stores the intermediate form moving picture data in advance in the storage device. Keep it.

When there is a distribution request for moving image data from the client terminal to the moving image distribution server, the network performance detecting means determines the transmission speed of the network to which the client terminal that has requested the distribution of moving image data is connected. It detects and notifies the code amount control means.

The code amount control means determines the code amount for recompressing the moving image data of the intermediate form according to the state of the network obtained by the network performance detecting means.

The recompression unit changes the code amount of the moving image data of the intermediate form stored in the storage device in advance by the semi-expansion unit according to the code amount determined by the code amount control unit, and the code amount. The re-compressed moving picture data is distributed to the client terminal by re-compressing the moving picture data of the intermediate form with the changed.

In the conventional video distribution system, when the compression ratio of moving picture data is changed in accordance with the load on the network, when the distribution request is made from the client terminal, the already-compressed moving picture data is converted into an intermediate form. After decompressing and changing the code amount, recompression was performed.

As described above, the moving picture distribution server can control the code amount of moving picture data adapted to the state of the network in response to a plurality of distribution requests from a plurality of client terminals connected to the network. , Stable video distribution is possible even on the existing network.

In particular, since the moving picture distribution server stores the already compressed compressed moving picture data in an intermediate form which is expanded halfway, the process of expanding the intermediate form when recompressing is omitted, and the process is relatively light. It is possible to efficiently control the transmission code amount.

Further, since the moving picture distribution server stores the already compressed moving picture data in an intermediate form in which it has been expanded halfway, it is possible to easily perform conversion to another compression method.

As described above, according to the information processing apparatus,
Since the already-compressed moving image data is decompressed into the intermediate form moving image data and stored in the storage device in advance, it is possible to efficiently deliver the moving image data having the code amount according to the load of the network.

(2) In the information processing apparatus described in (1) above, an entropy decoder that creates intermediate form moving image data including a discrete cosine transform coefficient sequence by entropy decoding the compressed moving image data. An entropy encoder that entropy-encodes the discrete cosine transform coefficient in the moving image data of the intermediate form corresponding to the code amount determined by the code amount control means.

In the moving picture distribution server which is the information processing device, when the compressed picture data is expanded to an intermediate expanded state by the half-expansion means, the compressed picture data is entropy-decoded by the entropy decoder to be discrete. Intermediate form moving image data including a cosine transform coefficient sequence is created and stored in a storage device.

When there is a distribution request of moving picture data from the client terminal to the moving picture distribution server, the code amount control means, in accordance with the network state obtained by the network performance detecting means, the intermediate form moving picture. Determine the code amount for recompressing data.

The recompression means changes the number of discrete cosine transform coefficients in the intermediate form moving image data stored in advance in the storage device by the half-expansion means in accordance with the code amount determined by the code amount control means. Then, the changed discrete cosine transform coefficient sequence is recompressed by an entropy encoder, and the recompressed moving image data is delivered to the client terminal.

With the discrete cosine transform coefficient, the higher the frequency component, the less the influence on the quality of moving image data.
In the recompression unit, when reducing the number of discrete cosine transform coefficients in the moving image data of the intermediate form according to the code amount determined by the code amount control unit, reducing the high frequency component of the discrete cosine transform coefficient. By doing so, the code amount of the moving image data can be reduced without degrading the quality of the recompressed moving image data as much as possible.

As described above, according to the information processing apparatus,
Since the previously compressed video data is decompressed into intermediate form video data containing discrete cosine transform coefficients and stored in the storage device, it is possible to recompress the video data without degrading the quality of the video data as much as possible. is there.

[0032]

BEST MODE FOR CARRYING OUT THE INVENTION In the following, in the moving picture distribution method and apparatus for carrying out the invention, moving picture data in an intermediate form is created in advance from moving picture data compressed by MPEG, and a distribution request of the moving picture data is made by the client terminal. A video distribution server of one embodiment that creates and distributes video data recompressed according to the load of the network when is performed.

FIG. 1 is a diagram showing a schematic configuration of the moving image distribution server of this embodiment. In FIG. 1, 1 is half-expansion means, 2 is moving image data in an intermediate form, 3 is network performance detection means, 4 is code amount control means, 5 is recompression means, 6 is code amount measuring means, and 103b is entropy encoder. , 104
Is an encoding buffer, 106 is a system multiplexing encoder, 1
Reference numeral 07 is a storage medium, 108 is a system multiplexing decoder,
110 is a decoding buffer, and 111a is an entropy decoder.

As shown in FIG. 1, the moving image distribution server of this embodiment includes a half-expansion means 1, moving image data 2 of an intermediate form,
Network performance detection means 3, code amount control means 4,
The recompressing unit 5, the code amount measuring unit 6, the entropy encoder 103b, the encoding buffer 104, the system multiplexing encoder 106, the storage medium 107, the system multiplexing decoder 108, and the decoding buffer 110. And an entropy decoder 111a.

Further, as shown in FIG. 1, in the moving picture distribution server of this embodiment, the moving picture data in the storage medium 107 is converted into the moving picture data 2 in the intermediate form by the half-expansion means 1 and stored in the storage device. When a distribution request is made from the client terminal used by the user, the recompression unit 5 recompresses the intermediate form moving image data 2 according to the load on the network and delivers it.

The half-expansion means 1 of the moving picture distribution server of this embodiment comprises a system multiplexing decoder 108, a decoding buffer 110, and an entropy decoder 111a, and expands moving picture data halfway to obtain an intermediate form. This is means for creating the moving image data 2.

The network performance detecting means 3 is a means for detecting the congestion degree of the network between the moving picture distribution server and the client terminal, the transmission speed of the network to which the client terminal is connected, and the like.

The code amount control means 4 determines the moving image data to be distributed according to the transmission performance of the network detected by the network performance detecting means 3 and the code amount for each frame of the moving image measured by the code amount measuring means 6. It is a means for determining the code amount.

The recompression means 5 is an entropy encoder 103.
b, an encoding buffer 104, and a system multiplexing encoder 106, and is means for re-encoding (compressing) the intermediate form moving image data 2 according to the code amount determined by the code amount control means 4. .

The code amount measuring means 6 is a means for measuring the code amount of the moving image data generated by the recompressing means 5 for each frame of the moving image and notifying the measured code amount to the code amount control means 4. .

FIG. 2 is a diagram showing an example of a network configuration and a hardware configuration in which the moving image distribution server of this embodiment is connected. In FIG. 2, 200 is a moving image distribution server, 201 is a CPU, 202 is a main memory, 203 is a display device, 204a is a keyboard, 204b is a mouse,
205 is a storage medium reading device, 206 is a storage device, 207 is a network control device, 208 is a modem,
209 is a clock circuit, 210 is a DSP, 211 is a bus, 2
20 to 240 are client terminals, 250 is an apparatus for reproducing MPEG moving image data, 261 and 262 are networks, and 263 is a telephone line.

As shown in FIG. 2, in the network configuration and hardware configuration in which the moving image distribution server 200 of this embodiment is connected, the moving image distribution server 200 is provided with a CPU 201.
A main memory 202, a display device 203, a keyboard 204a, a mouse 204b, a storage medium reading device 205, a storage device 206, a network control device 207, a modem 208, a clock circuit 209,
It has a DSP 210, a bus 211, client terminals 220 to 240, an MPEG moving picture data reproduction dedicated device 250, networks 261 and 262, and a telephone line 263.

Further, as shown in FIG. 2, in the network configuration and hardware configuration in which the moving image distribution server 200 of the present embodiment is connected, the moving image distribution server 200 and the client terminal 220 are connected via the network 261.
Also, the video distribution server 200 and the client terminal 230
, And 240 via telephone line 263.

In the moving image distribution server 200 of this embodiment,
A CPU 201, which is a central processing unit that controls execution of software processing, a main memory 202 that loads programs and data to be operated, a display device 203 that displays execution results of programs, a keyboard 204a that is an input device that captures instruction input, The mouse 204b, a storage device 206 including a magnetic disk, which is a non-volatile storage medium, a network control device 207 that controls the input / output of data to / from the network 261, and a modem 208 that controls data communication via the telephone line 263. , CPU201
Clock circuit 209 for supplying time data to the CPU 201
A DSP 210, which is a digital signal processor that performs independent data processing, and a storage medium reading device 205 that captures video data stored in a storage medium 107 such as a VideoCD.
Connected with.

The bus 211 of the moving image distribution server 200 is 1
It may be a book bus or a plurality of buses configured via a bridge.

The client terminals 220 to 240 are the network 2 such as a local area network (LAN).
61 or 262 or the telephone line 263 is connected to either one or both, and receives the moving image data distributed from the moving image distribution server 200 to reproduce and display the moving image data.

The hardware configuration of the client terminals 220 to 240 may be the same as the hardware configuration of the moving image distribution server 200, and at least the devices from the CPU 201 to the storage device 206 of the moving image distribution server 200 and the clock circuit 209 are connected to the bus 211. When connecting to the network 261 or 262, the network control device 207 must be provided, and when connecting to the telephone line 263, the modem 208 must be provided.

Further, the client terminals 220-240
Has software for reproducing and displaying MPEG moving image data, and may have a device 250 dedicated for reproducing MPEG moving image data like the client terminal 220.

FIG. 3 shows the moving picture distribution server 20 of this embodiment.
0 is a diagram showing a software configuration example of 0 and client terminals 220 to 240. In FIG. 3, 301 is an operating system, 302 is a communication control driver, and 30.
3 is a disk control driver, 304 is a protocol control program, 305 is a file management program, 400 is a half-expansion program, 500 is a network performance detection program, 510 is a recompression distribution program, and 550 is a moving image reproduction display program.

As shown in FIG. 3, the moving image distribution server 200 and the client terminals 220 to 240 of this embodiment include an operating system 301 and a communication control driver 3.
02, a disk control driver 303, a protocol control program 304, and a file management program 305.
A half decompression program 400, a network performance detection program 500, a recompression distribution program 510,
It has a moving image reproduction display program 550, and FIG.
(A) shows the software configuration of the video distribution server 200,
FIG. 3B shows the software configuration of the client terminals 220 to 240.

The software described below operates on the main memory 202 under the execution of the CPU 201, but a part of the software may be executed by the DSP 210.

First, the moving picture distribution server 200 of the present embodiment.
Also, the common parts of the client terminals 220 to 240 will be described.

The operating system 301 controls the display device 203, the keyboard 204a which is an input device, and the like which form the moving image distribution server 200, and controls the execution of a plurality of programs which will be described later.

The communication control driver 302 controls the network control device 207 and the modem 208 to control data communication processing.

The disk control driver 303 controls the storage device 206 and the storage medium reading device 205.

The protocol control program 304 executes and controls data communication processing with other terminals and programs operating on the other terminals.

A general protocol control program secures a buffer for data transmission and reception on the main memory 202 to perform data communication processing, and the protocol control program 304 is similar to this.

The file management program 305 manages the data stored in the storage device 206 or the storage medium 107 in logical units such as files.

The software from the communication control driver 302 to the file management program 305 may be configured and realized as a part of the functions of the operating system 301, or may be configured independently of the operating system 301.

The software from the operating system 301 to the file management program 305 described above is preferably functionally equivalent in the moving image distribution server 200 and the client terminals 220 to 240. Need not be equal.

The software from the operating system 301 to the file management program 305 includes the half-expansion program 400, the network performance detection program 500, and the recompression distribution program 510, which will be described later.
Also, in response to a processing request from a program such as the moving image reproduction display program 550, processing such as communication and data storage is performed.

Next, the moving image distribution server 200 of the present embodiment.
Describes the software configuration unique to the.

The half-expansion program 400 of the moving image distribution server 200 converts the MPEG moving image data into the intermediate-form moving image data 2 and performs a process corresponding to the half-expansion means 1 for storing in the storage device 206.

The network performance detection program 500 of the moving image distribution server 200 is the network 261 or 2
Communication status such as effective value of transmission speed of 62 and telephone line 263
The process corresponding to the network performance detecting means 3 for detecting the transmission speed of the above is performed.

The recompression distribution program 510 of the moving picture distribution server 200 recompresses the intermediate form moving picture data 2 into a code amount adapted to the state of the network 261 or 262 and the transmission speed of the telephone line 263, and then the client terminal 220.
The processing corresponding to the code amount control means 4, the recompression means 5, and the code amount measuring means 6 to be distributed to the channels 240 to 240 is performed.

In the video distribution server 200, the half-expansion program 400 and the network performance detection program 50
0 and each program of the recompression distribution program 510,
It is desirable to operate independently and in parallel, and therefore, the operating system 301 of the video distribution server 200.
Must be able to provide a multitasking environment, for example.

The processing details of each program of the half-expansion program 400, the network performance detection program 500, and the recompression distribution program 510 of the moving image distribution server 200 will be described later.

Next, the moving picture distribution server 200 of the present embodiment.
A software configuration unique to each of the client terminals 220 to 240 will be described.

The moving picture reproducing / displaying program 550 of the client terminals 220 to 240 controls the MPEG moving picture data reproducing exclusive apparatus 250 when the MPEG moving picture data reproducing exclusive apparatus 250 is connected to the moving picture distribution server 2.
The MPEG moving image data sent from 00 is expanded and displayed on the display device 203.

Alternatively, the client terminals 220-240
The video reproduction display program 550 of the video distribution server 2
The decompression process for decompressing the MPEG moving image data sent from 00 may be performed as a software process operating on the main memory 202 or a software process operating on the DSP 210.

Here, what is called a program, in particular, the half-expansion program 400, the network performance detection program 500, and the recompression distribution program 510, is generally called a task or process.
It is preferably a unit for allocating PU resources.

FIG. 4 shows the moving picture distribution server 20 of this embodiment.
9 is a flowchart showing a processing procedure of a half-decompression program 400 of 0.

FIG. 5 shows the moving picture distribution server 20 of this embodiment.
9 is a flowchart showing a processing procedure of a network performance detection program 500 and a recompression distribution program 510 of 0. In FIG. 5, reference numerals 1001 to 1002 denote data tables.

As shown in FIG. 5, in the network performance detection program 500 and the recompression distribution program 510 of the moving picture distribution server 200 of this embodiment, information is transmitted between the network performance detection program 500 and the recompression distribution program 510. A data table 1001 to be used,
A data table 100 recording the amount of data per frame of a moving image re-encoded by the re-compression distribution program 510.
2 are secured on the main memory 202.

FIG. 6 is a diagram showing a schematic configuration of a conventional MPEG moving picture coding and decoding apparatus. In FIG. 6, 1
01 is an input and preprocessor, 102 is a video resource encoder, 103 is a multiplexing encoder, 103a is a hierarchical encoder, 103b is an entropy encoder, 104 is an encoding buffer, 105 is a video data encoder, and 106 is System multiplex encoder, 107 storage medium, 108 system multiplex decoder, 109 moving image data decoder, 110 decoding buffer, 111 multiplex decoder, 111a entropy decoder, 111b hierarchical decoding , 112 is a video resource decoder, and 113 is a post-processing and output device.

As shown in FIG. 6, the conventional MPEG moving picture coding / decoding apparatus comprises an input / preprocessor 101, a video resource encoder 102, a multiplexing encoder 103, and a hierarchical encoder 103a. , Entropy encoder 103b
, Encoding buffer 104, and moving image data encoder 105
, System multiplexing encoder 106, and storage medium 10
7, system multiplex decoder 108, video data decoder 109, decoding buffer 110, multiplex decoder 1
11, an entropy decoder 111a, a hierarchical structure decoder 111b, a video resource decoder 112, and a post-processing and output unit 113.

Further, as shown in FIG. 6, in the conventional MPEG moving picture coding and decoding apparatus, the input and preprocessor 101 is used.
And video data encoder 105 and system multiplexing encoder 10
6, the storage medium 107 is connected, and the storage medium 107, the system multiplexing decoder 108, the moving image data decoder 109, and the post-processing and output device 113 are connected.

The moving picture data encoder 105 includes a video resource encoder 102, a multiplexing encoder 103, and an encoding buffer 104. The multiplexing encoder 103 forms a layer of compressed data into a hierarchical bit stream. It has a structure coder 103a and an entropy coder 103b that performs entropy coding which is variable length coding using statistical properties.

The moving picture data decoder 109 comprises a decoding buffer 110, a multiplexing decoder 111 and a video resource decoder 112. The multiplexing decoder 111 entropy-decodes a bit stream of encoded moving picture data. Decoder 111a and video resource decoder 11
It has a hierarchical structure decoder 111b that cuts out two units of data to be decoded.

FIG. 7 shows a conventional video resource encoder 10
It is a figure which shows the functional block of 2. In FIG. 7, 70
1 to 707 are blocks.

As shown in FIG. 7, the conventional video resource encoder 102 takes in an input image, quantizes it, and then outputs a quantized transform coefficient.

FIG. 8 shows a conventional video resource decoder 11
It is a figure which shows the functional block of 2. In FIG. 8, 80
1 to 805 are blocks.

As shown in FIG. 8, the conventional video resource decoder 112 performs a process of fetching a quantized transform coefficient, performing inverse quantization, and transmitting an output image.

FIG. 9 is a diagram showing an outline of a conventional MPEG moving picture hierarchical structure. In FIG. 9, 901 is a block layer, 901a is an EOB code, 902 is a macroblock layer, 903 is a slice layer, 903a is SSC, 904 is a picture layer, 905 is a GOP layer, 906 is a video sequence layer, and 910 is a discrete cosine transform. It is a coefficient sequence.

As shown in FIG. 9, the conventional MPEG moving picture hierarchical structure has a block layer 901 and an EOB code 901a.
A macroblock layer 902, a slice layer 903,
SSC 903a, picture layer 904, and GOP layer 90
5, a video sequence layer 906, and a discrete cosine transform coefficient sequence 910.

Further, as shown in FIG. 9, the conventional MPEG moving picture hierarchical structure represents a hierarchical data structure of MPEG moving picture data during the processing of the multiplexing encoder 103 and the multiplexing decoder 111. MPEG movie data is
Before the entropy coding of the multiplexing encoder 103,
Immediately after the entropy decoding of the multiplexing decoder 111, it has such a hierarchical data structure.

Block layer 901 of MPEG moving picture hierarchical structure
Is a unit of data composed of pixels of luminance or color difference of 8 pixels in the vertical and horizontal directions, and the discrete cosine transform is executed in this unit.

The macroblock layer 902 of the MPEG moving picture hierarchical structure includes four luminance blocks which are vertically and horizontally adjacent to each other.
The luminance block and two color difference blocks at the same position on the image are composed of a total of six block layers 901.

Slice layer 903 of MPEG moving picture hierarchical structure
Is composed of one or a plurality of macroblock layers 902 that are consecutive in the scanning order of an image, and at the beginning, an SSC (Slice Start Code) 903a that is a synchronization code.
There is.

Picture layer 904 of MPEG moving picture hierarchical structure
Corresponds to one frame of a moving image, is composed of at least one or a plurality of slice layers 903, and is an intra frame (hereinafter referred to as I frame or I picture) or a forward prediction frame (hereinafter referred to as P frame or P) according to the encoding method. And a bidirectional prediction frame (hereinafter referred to as a B frame or a B picture).

An MPEG video hierarchical structure GOP (Grou
The p Of Picture) layer 905 is composed of one or more I picture layers and 0 or more non-I picture layers.

The video sequence layer 906 of the MPEG moving picture hierarchical structure has the same image size and image rate.
It is composed of one or more GOP layers 905.

Immediately before the entropy coding of the multiplex encoder 103 and immediately after the entropy decoding of the multiplex decoder 111, the block layer 901 encodes the quantized discrete cosine transform coefficient according to run and level. Stored (hereinafter referred to as run-level coding).

The MPEG cosine hierarchical structure discrete cosine transform coefficient sequence 910 is a sequence of 8 × 8 discrete cosine transform coefficients obtained by performing a discrete cosine transform on vertical and horizontal 8 pixels.

The DC component 911 is a DC component of the discrete cosine transform coefficient sequence 910, and a total of 63 AC components 912 to 974 are AC components of the discrete cosine transform coefficient sequence 910.

In the block layer 901, the discrete cosine transform coefficient sequence 910 is arranged in order from the DC component 911 to the AC component 974, and the discrete cosine transform coefficient sequence 91 is obtained.
At the end of 0, EOB indicating the end of the block layer 901
The code 901a is stored.

First, the moving image distribution server 200 of this embodiment.
Prior to the description of FIG. 6, the conventional MPEG will be described with reference to FIGS.
The moving picture encoding and decoding processing will be briefly described.
Regarding the MPEG moving picture coding and decoding processing, "International standard for multimedia coding (6th edition) published by Maruzen Co., Ltd.
Chapter) ”and the like.

As shown in FIG. 6, when the input video signal is encoded in the conventional MPEG moving image encoding process,
First, the input and preprocessor 101 converts input moving image data such as video into a data format required by the moving image data encoder 105, that is, a signal composed of luminance and color difference.

The video resource encoder 102 of the moving picture data encoder 105 compresses the moving picture data by reducing the temporal and spatial redundancy of the moving picture data to reduce the data amount.

That is, as shown in FIG. 7, the video resource encoder 102 of the moving picture data encoder 105 performs the following processing.

The video resource encoder 102 of the moving picture data encoder 105 first fetches moving picture data from the input and preprocessor 101 for each frame in block 701.
In block 702, the vertical and horizontal pixels of the moving image data are 8 pixels (total 6 pixels).
A discrete cosine transform is performed in units of 4 pixels to obtain discrete cosine transform coefficients (64 for the unit).

At block 703, the discrete cosine transform coefficient is quantized. If the frame is an I frame, at block 704 the quantized transform coefficient is multiplexed into the multiplex encoder 103.
Send to

If the frame is a P frame or a B frame, block 705 dequantizes the quantized transform coefficient to obtain a discrete cosine transform coefficient, and block 706 performs an inverse discrete cosine transform on the discrete cosine transform coefficient. A block 707 generates each predicted frame data, the blocks 702 and 703 are subjected to the predicted frame data, and a block 704 sends the quantized transform coefficient to the multiplexing encoder 103.

Hierarchical structure encoder 10 of multiplexing encoder 103
3a configures the data compressed by the video resource encoder 102 into a hierarchical bit stream, and the entropy encoder 103b of the multiplexing encoder 103 performs entropy encoding of the hierarchical bit stream.

The encoding buffer 104 of the moving image data encoder 105 is a buffer for making the bit rate constant when sending entropy encoded data to the system multiplexing encoder 106, and is stored in the encoding buffer 104. The generated data amount is fed back to the video resource encoder 102 and the multiplexing encoder 103 and used to control the bit rate constant.

The system multiplexing encoder 106 multiplexes the moving image data from the moving image data encoder 105 and the audio data so that they are synchronized with each other during reproduction.

The moving image data encoded by the processing from the input / preprocessor 101 to the system multiplexing encoder 106 is recorded on the storage medium 107 such as a CD-ROM.

When the encoded moving picture data recorded in the storage medium 107 is reproduced, the system reproduction decoder 108 separates the reproduced reproduction moving picture and audio data in synchronization with each other.

The moving picture data separated by the system multiplexing decoder 108 is decoded by the moving picture data decoder 109 into moving picture data.

The decoding buffer 110 of the moving picture data decoder 109 is a buffer for guaranteeing the decoding processing time for the data sent from the system multiplexing decoder 108.

Entropy Decoder 1 of Multiplexing Decoder 111
11 a entropy-decodes the bitstream of the entropy-coded moving image data, and the hierarchical decoder 111 b of the multiplexing decoder 111 is the video resource decoder 11
2 cuts out the data unit to be decoded.

The video resource decoder 112 of the moving picture data decoder 109 performs a decompression process for decompressing the compressed information of the data cut out by the hierarchical structure decoder 111b of the multiplexing decoder 111.

That is, as shown in FIG. 8, the video resource decoder 112 of the moving picture data decoder 109 performs the following processing.

The video resource decoder 112 of the moving picture data decoder 109 first takes in the quantized transform coefficient for each frame from the multiplex decoder 111 in block 801, and reverses the quantized transform coefficient in block 802. Quantization processing is performed to obtain discrete cosine transform coefficients.

In block 803, the 64 discrete cosine transform coefficients corresponding to the vertical and horizontal 8 pixels of the original moving image are subjected to inverse discrete cosine transform processing to obtain vertical and horizontal 8 pixels. If the frame is an I frame, block 805 Then, the pixel data decoded in the above procedure is sent to the post-processing and output device 113.

If the frame is a P frame or a B frame, in block 804, each predicted frame data is generated into pixel data of the original moving image, and in block 805, the pixel data decoded by the above procedure is post-processed and output device. Send to 113.

The post-processing and output device 113 performs data conversion and correction of the decoded moving image data in accordance with the specifications of the display device 203 such as a display.

Based on the above, the operation of the moving image distribution server 200 of this embodiment will be described.

As shown in FIG. 1, in the moving picture distribution server 200 of this embodiment, first, the semi-expanding means 1 expands a plurality of compressed moving picture data stored in the storage medium 107 to an intermediate stage. A plurality of intermediate-form moving image data 2 are created in advance and stored in the storage device 206.

That is, the half-expansion means 1 of the moving image distribution server 200 takes in the video data which is the already-compressed moving image data with audio and is stored in the storage medium 107 such as the VideoCD, and performs the system multiplex decoding of the taken video data. The device 108 separates the moving image data and the audio data.

The semi-expansion means 1 of the moving picture distribution server 200 is
The separated moving image data is temporarily stored in the decoding buffer 110, and the moving image data in the decoding buffer 110 is entropy-decoded into a bit stream having a hierarchical structure shown in FIG. Then, the entropy-decoded intermediate form moving image data 2 is stored in the storage device 206.

Next, the moving picture distribution server 200 of the present embodiment.
When there is a video data distribution request from the client terminals 220-240, the intermediate mode moving image data 2 stored in the storage device 206 is compressed again to create video data, and the created video data is distributed. .

When the video data is distributed to the client terminals 220 to 240, the network performance detecting means 3 detects the transmission rate of the network 261 or the telephone line 263 and notifies the code amount controlling means 4 of the code amount controlling means. 4 is a network 261 or a telephone line 263
The optimum code amount of the moving image data 2 of the intermediate form according to the state of is determined.

The adjustment to the optimum code amount is possible by adjusting the number of the quantized discrete cosine transform coefficients which compose the moving image data 2 in the intermediate form corresponding to the high frequency component and entropy-encoding. become.

Therefore, the code amount control means 4 extracts the appropriate number of discrete cosine transform coefficients from the DC component 911 to the AC component 974, as shown in FIG. 9, and inserts the EOB code 901a at the end. Then, the new block layer 901 is configured and input to the entropy encoder 103b.

In the entropy encoder 103b, the bit stream including the discrete cosine transform coefficient sequence 910 is entropy-encoded and input to the encoding buffer 104, and the system multiplexing encoder 106 and the system multiplexing decoder 108
The audio data corresponding to the moving image data separated in (4) and the encoded moving image data in the encoding buffer 104 are multiplexed in synchronization with each other and transmitted to the client terminals 220 to 240 via the network 261 or the telephone line 263.

Code amount measuring means 6 of the moving image distribution server 200
Measures the code amount of the moving image data transmitted from the recompression unit 5 on the network 261 for each frame of the moving image,
The measured code amount is notified to the code amount control means 4.

The code amount control means 4 compares the determined code amount with the code amount notified from the code amount measuring means 6,
It is determined whether or not the recompressed moving image data can be sent out with a desired code amount, that is, the optimum code amount according to the state of the network 261 or the telephone line 263. If the code amount is excessive, the AC components 912 to 974 to be selected are selected. If the number of codes is reduced and the code amount is too small, the AC component 912 to be selected is selected.
~ 974 to increase the number of client terminals 220 ~
The data amount of the moving image data distributed to 240 is controlled.

Next, the moving image distribution server 200 of this embodiment.
In the following, a processing procedure of the half-expansion program 400 for converting the compressed moving image data stored in the storage medium 107 into the intermediate-type moving image data 2 in advance will be described.

The half-expansion program 400 of the moving picture distribution server 200 of this embodiment is the client terminals 220 to 24.
0 MPEG video data to be distributed to the video distribution server 20
It is activated by the operating system 301 triggered by the input of an instruction to store it in 0.

As shown in FIG. 4, the half-expansion program 400 of the moving image distribution server 200 of this embodiment first causes the file management program 305 to execute the storage medium 1 via the storage medium reading device 205 in step 401.
A fixed amount of MPEG moving image data is fetched from 07 and loaded into the main memory 202.

The amount of data that the half-expansion program 400 loads into the main memory 202 at one time may be determined by checking the free space of the main memory 202 at the start of execution of the half-expansion program 400. The determined capacity value or the capacity value determined by the user may be used.

In step 402, it is determined whether the MPEG moving image data loaded in step 401 is the end of the data, for example, by the presence or absence of an end of file (EOF) indicating the end of the data file. If 407 is not the end, step 403
To execute.

At step 403, data other than moving image data such as audio is separated from the MPEG moving image data loaded at step 401 by the system multiplexing decoder 108, and at step 404, the data other than the separated moving image data is converted into a file. It is individually saved in the storage device 206 by the management program 305.

In the following step 405, the separated moving picture data is entropy-decoded by the entropy decoder 111a, and the moving picture data has the hierarchical data structure as shown in FIG. 9 and is quantized to form the block layer 901. The discrete cosine transform coefficient thus converted is converted into moving image data 2 of an intermediate form which is hybrid coded (two-dimensional coded) of run level coding and Huffman coding.

At step 406, the entropy-decoded intermediate form moving image data 2 is transferred to the file management program 3
The data is saved in the storage device 206 by 05, and the process returns to step 401.

If the MPEG moving image data loaded in step 401 is the end of the data, the process proceeds to step 407, and in step 407, end processing such as releasing resources for releasing the data storage area on the main memory 202 is performed, The processing of the half-expansion program 400 is completed.

Next, the moving image distribution server 200 of this embodiment.
In the network, a network performance detection program 500 for detecting the state of the network 261 or the telephone line 263.
Then, the processing procedure of the recompression distribution program 510 for recompressing and distributing the intermediate-format moving image data 2 will be described.

The network performance detection program 500 and the recompression distribution program 510 of the moving picture distribution server 200 of this embodiment each time the moving picture distribution server 200 receives a distribution request from the client terminals 220 to 240.
It is generated and activated by the operating system 301.

That is, the operating system 30
If 1 can accept multiple delivery requests,
A plurality of network performance detection programs 500 and recompression distribution programs 510 are generated and operate in parallel.

Therefore, the data tables 1001 and 10
02 is the generated network performance detection program 5
00 or the same number as the number of recompression distribution programs 510 needs to be generated.

First, the moving picture distribution server 200 of this embodiment.
The processing procedure of the network performance detection program 500 will be described.

As shown in FIG. 5, the network performance detection program 500 of the moving picture distribution server 200 of this embodiment.
Is the protocol control program 30 in step 501.
In the buffer for data transmission used by No. 4, the time from when the data to be transmitted (for example, moving image data) is once stored to when it all flows to the network 261 or the telephone line 263 and is cleared is based on the time information of the clock circuit 209. To measure.

In step 502, the data transmission rate at that time is calculated by dividing the data amount of the transmitted data by the measured measurement time.

In step 503, the data transmission rate calculated in step 502 and the predetermined network 26
1 or the effective value of the data transmission rate of the telephone line 263 is compared, and when the data transmission rate calculated in step 502 is lower than the predetermined effective value, the calculated data transmission rate value of the effective value is calculated. The ratio is recorded in the data table 1001 as the transmission performance.

If the data transmission rate calculated in step 502 is not lower than the predetermined effective value, the ratio of the calculated data transmission rate value to the effective value is 10%.
The data table 100 stores the information of the transmission performance indicating that it is 0%.
Record in 1.

At step 504, it is checked whether or not the recompression distribution program 510 is being executed by inquiring to the operating system 301, for example, and if the recompression distribution program 510 is being executed, step 501 is executed.
If the recompression distribution program 510 is not being executed, the process proceeds to step 505, and the main memory 2
02, the network performance detection program 500
Is completed.

Next, the moving image distribution server 200 of this embodiment.
The processing procedure of the re-compression distribution program 510 will be described.

As shown in FIG. 5, the recompression distribution program 510 of the moving picture distribution server 200 of the present embodiment receives the transmission performance information passed from the network performance detection program 500 via the data table 1001 in step 511. The state of the network 261 or the telephone line 263 is judged by using the clue, the code amount for recompressing and distributing the intermediate form moving image data 2 is determined, and the corresponding amount is dealt with.
The read amount (the number of codes) of the two-dimensionally encoded discrete cosine transform coefficient is determined for each frame of the moving image.

For example, when the transmission performance (transmission speed) of the network 261 or the telephone line 263 is about 50% of the normal effective value (for example, the value measured and empirically determined in advance), The compression distribution program 510 refers to the information that the network performance in the data table 1001 recorded by the network performance detection program 500 is 50%, and the code amount is 50% in step 511.
Make the decision to reduce and deliver.

In step 512, the determined number of discrete cosine transform coefficients is selected from the discrete cosine transform coefficient sequence 910 of the block layer 901 which constitutes the moving image data 2 of the intermediate form, and the file management program 305 is selected.
The EOB code 901a is loaded from the storage device 206 to the main memory 202 by the EOB code 901a, and the EOB code 901a is inserted at the end of all the block layers 901 constituting the frame of the moving image to be distributed. Frame data is reconstructed into a picture layer 904 form.

In order to reduce the processing load of the distribution processing of the moving picture distribution server 200 and to execute the distribution processing at high speed, in step 512, the beginning of each block of the moving picture data 2 in the intermediate form is speeded up. It is necessary to search, but for that purpose, a method of searching the SSC 903a to find the head of the slice layer 903 and tracing the bit stream from there to search the head of the block layer 901 is desirable and effective.

At step 513, whether or not the end of the data of the moving image data 2 in the intermediate form has been processed is determined by, for example, E.
If it is determined based on the presence or absence of OF, etc., and if the end of the data is processed, the process of step 517 is performed, and if the end of the data is not processed, the process of step 514 is performed.

In step 514, the discrete cosine transform coefficient selected in step 513 is set to the entropy encoder 10
Entropy coding by 3b.

In step 515, the system multiplexing encoder 106 re-multiplexes data other than moving image data such as audio, and the re-multiplexed code amount is used in the data table 100.
Record in 2.

In step 516, the recompressed MPEG is used.
The moving image data is transmitted to the client terminals 220 to 240 by the protocol control program 304, and step 51
Return to processing of 1.

From the processing of step 511 for the second time, in addition to the information in the data table 1001, the distribution code amount can also be obtained by using the data amount of the result of re-encoding the previous frame recorded in the data table 1002 as a clue. decide.

For example, the code amount per frame of the moving image data 2 of the intermediate form which is the original data is recorded, and this is compared with the code amount recorded in the data table 1002, and the desired reduction amount is exceeded. If the number has been reduced, a decision is made to increase the number of codes of the two-dimensionally encoded discrete cosine transform coefficient selected in step 512 by one.

At step 513, the intermediate form moving image data 2
If the data up to the end has been processed, step 5
17, and in step 517, the main memory 202
A termination process such as release of resources for releasing the upper data storage area is performed, and the process of the recompression distribution program 510 is terminated.

As described above, in the moving image distribution server 200 of this embodiment, the network 261 or the telephone line 263 is used.
The distribution of the MPEG moving image data in a data amount adapted to the transmission speed of, that is, the control of the code amount of the distributed MPEG moving image data is processed and stored in advance into an intermediate form in which the MPEG moving image data is partially expanded, This can be achieved by simply selecting an appropriate amount of discrete cosine transform coefficients at the time of distribution and performing the latter half step of compression, that is, the processing after entropy coding, so that it can be realized with relatively low load processing.

Therefore, stable moving image distribution without frame loss can be achieved without using a high-speed network, and the moving image distribution server 200 does not require dedicated hardware for compressing and expanding MPEG moving image data. , It is possible to build a relatively inexpensive video distribution system as a whole.

As described above, according to the moving picture distribution server of the present embodiment, since the moving picture data is decompressed into the intermediate form moving picture data and stored in the storage device in advance, the code amount according to the load of the network. It is possible to efficiently distribute the moving image data.

Further, according to the moving image distribution server of the present embodiment, since the moving image data is previously decompressed into the intermediate form moving image data including the discrete cosine transform coefficient and stored in the storage device, the quality of the moving image data can be improved as much as possible. It is possible to recompress the moving image data without loss.

Although the present invention has been specifically described based on the above embodiment, the present invention is not limited to the above embodiment and can be variously modified without departing from the scope of the invention. Is.

For example, the moving image distribution server 20 of this embodiment.
In 0, moving image data of the MPEG system is assumed as moving image data, but moving images for which data compression is performed by discrete cosine transform, for example, motion JPEG (Joint Pho) are used.
It is also applicable to a moving image distribution system using a moving image according to the topographical Experts Group method.

Further, the intermediate form MPEG moving image data in the moving image distribution server of the present embodiment is a moving image for which data compression is performed by discrete cosine transform, for example, Motion JP.
EG and H. It can be relatively easily converted to another compression method such as H.261, and these motion JPEG and H.264. 261
It is also possible to deliver to a client having only a display device of a moving image of another compression method such as.

[0167]

The effects obtained by the typical ones of the inventions disclosed in the present application will be briefly described as follows.

(1) Since the moving picture data is previously decompressed into intermediate form moving picture data and stored in the storage device, it is possible to efficiently distribute the moving picture data of the code amount according to the load of the network. .

(2) Since the moving picture data is previously decompressed into the intermediate form moving picture data including the discrete cosine transform coefficient and stored in the storage device, the moving picture data can be recompressed without degrading the quality of the moving picture data as much as possible. Is possible.

[Brief description of drawings]

FIG. 1 is a diagram showing a schematic configuration of a moving image distribution server 200 of the present embodiment.

FIG. 2 is a diagram showing an example of a network configuration and a hardware configuration in which a moving image distribution server 200 of this embodiment is connected.

FIG. 3 is a diagram showing a software configuration example of a moving image distribution server 200 and client terminals 220 to 240 of the present embodiment.

FIG. 4 is a flowchart showing a processing procedure of a half-expansion program 400 of the moving image distribution server 200 of this embodiment.

FIG. 5 is a flowchart showing a processing procedure of a network performance detection program 500 and a recompression distribution program 510 of the moving image distribution server 200 of this embodiment.

FIG. 6 is a diagram showing a schematic configuration of a conventional MPEG moving image encoding / decoding device.

FIG. 7 is a diagram showing functional blocks of a conventional video resource encoder 102.

FIG. 8 is a diagram showing functional blocks of a conventional video resource decoder 112.

FIG. 9 is a diagram showing an outline of a conventional MPEG moving image hierarchical structure.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 1 ... Half-expansion means, 2 ... Intermediate form moving image data, 3 ... Network performance detection means, 4 ... Code amount control means, 5 ... Recompression means, 6 ... Code amount measuring means, 103b ... Entropy encoder, 104 ... Code Buffer 106, system multiplex encoder 107, storage medium 108, system multiplex decoder 110, decoding buffer 111a, entropy decoder 200, video distribution server 201, CP
U, 202 ... Main memory, 203 ... Display device, 204
a ... Keyboard, 204b ... Mouse, 205 ... Storage medium reading device, 206 ... Storage device, 207 ... Network control device, 208 ... Modem, 209 ... Clock circuit,
210 ... DSP, 211 ... Bus, 220-240 ... Client terminal, 250 ... MPEG moving picture data reproduction dedicated device, 261 and 262 ... Network, 263 ... Telephone line, 301 ... Operating system, 302 ... Communication control driver, 303 ... Disk control Driver 304
... Protocol control program, 305 ... File management program, 400 ... Half-expansion program, 500 ... Network performance detection program, 510 ... Recompression distribution program, 550 ... Movie playback display program, 1001-1
002 ... Data table, 101 ... Input and preprocessor,
102 ... Video resource encoder, 103 ... Multiplex encoder, 103a ... Hierarchical structure encoder, 105 ... Video data encoder, 109 ... Video data decoder, 111 ... Multiplex decoder, 111b ... Hierarchical decoder, 112 ... Video resource decoder, 113 ... Post-processing and output device, 701 to 707
... block, 801 to 805 ... block, 901 ... block layer, 901a ... EOB code, 902 ... macroblock layer, 903 ... slice layer, 903a ... SSC, 90
4 ... Picture layer, 905 ... GOP layer, 906 ... Video sequence layer, 910 ... Discrete cosine transform coefficient sequence.

Claims (4)

[Claims] 1. A moving picture distribution method for distributing moving picture data to a plurality of client terminals connected via a network, wherein intermediate form moving picture data created by expanding already compressed moving picture data to an intermediate expanded state is created. Stored in a storage device in advance,
The load of the network that transmits the moving image data is detected, the code amount for recompressing the moving image data of the intermediate form is determined according to the detected load of the network, and the intermediate form stored in the storage device in advance. The moving image distribution method, characterized in that the moving image data is recompressed and distributed according to the determined code amount. 2. Compressed moving image data is entropy-decoded to create moving image data in an intermediate form including a discrete cosine transform coefficient sequence and stored in a storage device, and the intermediate corresponding to the determined code amount. The moving image distribution method according to claim 1, wherein a discrete cosine transform coefficient in the moving image data of the form is selected, and the selected discrete cosine transform coefficient is entropy-encoded. 3. An information processing apparatus for distributing moving picture data to a plurality of client terminals connected via a network, wherein intermediate moving picture data created by expanding already compressed moving picture data to an intermediate expanded state is created. Semi-expansion means stored in advance in a storage device, network performance detecting means for detecting a load on a network for transmitting moving picture data, and a moving picture of the intermediate form according to the network load detected by the network performance detecting means. Code amount control means for determining the code amount for recompressing the data, and intermediate form moving image data stored in advance in the storage device by the half-expansion means according to the code amount determined by the code amount control means. An information processing apparatus comprising: a recompression unit for recompressing. 4. An entropy decoder for creating moving image data of an intermediate form including a discrete cosine transform coefficient sequence by entropy decoding already compressed moving image data, and a code amount determined by the code amount control means. The information processing apparatus according to claim 3, further comprising an entropy encoder that entropy-encodes the discrete cosine transform coefficient in the moving image data of the intermediate form.