JPH11355768A - Reproduction method for encoded image, code string converter and recording medium for code string conversion program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To perform conversion of a code string required in the case of transmitting and reproducing encoded images in real time in a reproduction part provided with the transmission line of a transmission speed slower than an encoding speed and the decoder of a low processing speed at a high speed in small circuit constitution. SOLUTION: A code string conversion part 105 converts a first code string for which digital image signals 101 are compression-encoded by a first encoding system to a second code string obtained by a second encoding system without returning it to the digital image signals. It is sent to the decoder 107 by using the transmission line 104 and the decoder 107 decodes the second code string. In the case that the decoder 107 is not provided with a software capable of decoding the second code string in real time, the decoding software 111 is read from a decoding software storage part 108 and sent to the decoder 107 together with the second code string.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for reproducing an encoded image, a code string conversion apparatus, and a recording medium for a code string conversion program, and more particularly, to a method for compressing an image to a high efficiency compression encoding method (hereinafter referred to as an encoding method). The present invention relates to a technique for reproducing a code string (bit stream) obtained by encoding using a combination of a plurality of transmission paths and decoders having different conditions.

[0002]

2. Description of the Related Art Now, a code sequence obtained by coding a digital image signal by a certain high-efficiency coding method is used by using a transmission line having a lower transmission speed. And reproducing the data by transmitting it to a decoder, or reproducing the data by a decoder having a low processing speed even if the transmission speed is not slow.

As described above, when the transmission speed of the transmission path is lower than the encoding speed, or when the processing speed of the decoder is lower than the encoding speed, the code string is transmitted and reproduced in real time. Is difficult. Therefore, it is necessary to create a code sequence obtained by an encoding method that can be transmitted and reproduced in real time for a certain image under the conditions of the respective transmission paths and decoders.

[0004] FIG. 4 shows a conventional configuration for transmitting and reproducing a code string in real time between transmission paths having different transmission rates and decoders connected to the transmission paths having different processing rates. Here is an example.

In FIG. 4, a digital image signal 301
Is an encoder 30 that performs encoding according to the first encoding method.
2 is sent to a decoder 308 via a transmission path 303 as a code string according to the first coding method, and is decoded and reproduced. At this time, the transmission path 303 is separately provided.
When the data is transmitted to the decoder 309 slower than the processing speed of the decoder 308 via the transmission path 304 having a lower transmission speed than the transmission speed of the decoder 308 and is decoded and reproduced, the output of the encoder 302 is output to the decoder 305. Decrypts the frame buffer memory 30
6, and then re-encoded by an encoder 307 based on a second encoding system that performs encoding at an encoding speed that matches the transmission speed of the transmission path 304, to form a code string according to the second encoding system. , To the decoder 309 for decoding and reproduction. Here, the case of two types of transmission paths and decoders has been described as an example, but there may be three or more types.

[0006]

The method of transmitting and reproducing a code string in real time described with reference to FIG. 4 includes a circuit for temporarily decoding and storing one digital image signal on the transmission side, and a transmission method. Since a plurality of encoders are provided for each type of path or decoder, there is a problem that the circuit scale of the entire system becomes large.

[0007] The present invention solves the above-mentioned problems, and performs encoding in real time without increasing the circuit scale on the transmission side even in a situation where various transmission paths having different capabilities and a decoder on the reception side are present. An object is to enable transmission and reproduction of an image.

[0008]

SUMMARY OF THE INVENTION In order to solve the above-mentioned problems, the present invention provides a method for decoding a code stream (bit stream) obtained by encoding using a first encoding method into a digital image signal. , By processing and editing the information contained in the code string to convert it into a code string obtained by high-efficiency compression coding of the second coding method, which has a lower coding rate. The main feature of this method is to transmit and reproduce the encoded code sequence in real time.

FIG. 1 shows a block configuration of the present invention. In FIG. 1, 101 is a digital image signal, and 102 is a first image signal.
, 103 and 104 are transmission paths, 1
05 is a code string converter, 106 is a decoder for decoding the first code string, 107 is a decoder for decoding the second code string, 10
8 denotes a decoding software storage unit, 109 and 110 denote switchers, and 111 denotes decoding software (software decoder).

A digital image signal 101 is encoded by a first encoder 102 into a first code string,
03 that can decode the first code string via
, And are decoded and reproduced. On the other hand, the first code sequence can be transmitted in real-time on the transmission path 104 having a lower transmission speed than the transmission path 103 without simultaneously converting the code sequence into a digital image signal in the code sequence conversion unit 105 and decoding. The signal is converted into a second code string that can be reproduced in real time by the device 107.

The generated second code string is transmitted to the transmission line 104.
Is transmitted to the decoder 107 in real time. At this time, if the decoder 107 does not have the decoding software 111 capable of reproducing the second code string in real time, the switches 109 and 110 connect the decoding software storage unit 108 and the decoder 107. , The decoding software 111 necessary for decoding the second code string is transmitted from the decoding software storage unit 108 through the transmission path 104 to the decoder 107.
Transmitted to

In the decoder 107, the second code string is
It is decoded and reproduced in real time using the decoding software 111. For example, if the first code sequence is a code sequence encoded using DCT coefficients and motion vectors, the code sequence conversion unit 105 performs variable length decoding on the first code sequence,
Take out DCT coefficient and motion vector and take out DC
Of the T coefficients, only a predetermined number or coefficients of low frequency components specified by a parameter or the like are left, and the value of the extracted motion vector is transformed according to the reduction of the DCT coefficient. The motion vector and the motion vector are variable-length coded, and header information is added to form a second code string.

When the above conversion of the code string is realized by a computer using a software program, the program can be stored in an appropriate recording medium such as a portable medium memory, a semiconductor memory, and a hard disk which can be read by the computer. it can.

According to the above-described method, a code sequence obtained by encoding an image signal by the first encoding method is converted into a second code sequence without returning to a digital image signal.
It is possible to perform real-time transmission on the connected transmission path and to perform real-time reproduction by the code string reproducing unit having the connected decoder.

[0015]

DESCRIPTION OF THE PREFERRED EMBODIMENTS In the following embodiments of the present invention,
A code string generated by a first coding method of coding by a DCT (Discrete Cosine Transform) coefficient and a motion vector, such as the MPEG-2 method (ISO / IEC 11172-2), Consider a case in which the data is converted into a second code string having a small code amount, and is transmitted and reproduced.

FIG. 2 shows a block configuration of the present embodiment. The coded sequence conversion unit 105 includes a variable length decoding unit 1051,
DCT coefficient converter 1052, motion vector converter 105
3. It has a code string generation unit 1054.

The first code string encoded by the encoder 102 is sent to the variable length decoding unit 1051. The variable-length decoding unit 1051 performs variable-length decoding on the first code string, extracts DCT coefficients and motion vectors, and sends the DCT coefficients to the DCT coefficient conversion unit 1052 and the motion vectors to the motion vector conversion unit 1053.

The DCT coefficient conversion unit 1052 divides each frame of the digital image signal into blocks (including n × n pixels) having n pixels on each side in both the horizontal and vertical directions according to the first encoding method. among the DCT coefficients of the block ^two blocks n which is obtained by two-dimensional DCT transform, leaving only m ² coefficients which is a low-frequency component (m <
n), and sends the DCT coefficients from which the other coefficients have been deleted to the code string generation unit 1054. M × m D of each block
When the inverse transform of the CT coefficients is performed, blocks of m pixels are obtained both horizontally and vertically. Therefore, the image size of each frame is reduced to m / n in both the vertical and horizontal directions.

The motion vector conversion unit 1053 is adapted to change the image size by the DCT coefficient conversion unit 1052,
The value of the motion vector is reduced and converted, and the converted motion vector is sent to the code string generation unit 1054.

In the code string generation unit 1054, the input D
The CT coefficient and the motion vector are variable-length coded, and header information is added to generate a code sequence of the second coding method.

At this time, the processing of the DCT coefficient conversion unit 1052 and the motion vector conversion unit 1053
The code amount of the coefficient and the motion vector is reduced to such an amount that the second code string can be transmitted / reproduced in real time by the transmission path 104 and the decoder 107.

The process of converting the first code string into the second code string by the code string conversion unit 105 is performed by converting the first code string into a digital image signal and then converting it into a second code string. As compared with the case of performing the processing, there is no need to perform the inverse DCT processing and the recalculation of the motion vector, so that the processing amount is small.

The second code string passes through the transmission path 104
The second code string is sent to the decoder 107. Here, when the decoder 107 does not include the decoding software 111 for decoding the second code string, the decoding software for decoding the second code string stored in the decoding software storage unit 108 in advance is used. , Through the transmission path 104,
It is transmitted to the decoder 107. This decryption software 11
1 has a function of performing two-dimensional inverse DCT on DCT coefficients in an m × m array.

In the decoder 107, the decryption software 11
1 decodes and reproduces the second code string in real time. FIG. 3 is a flowchart of the code string conversion unit according to the embodiment of the present invention. Hereinafter, (a) to (h) shown in FIG.
Will be described with reference to FIG. (a) First, the variable length decoding unit 1051 receives the first code string from the encoder 102 and performs variable length decoding. (b) Separate the motion vector and the DCT coefficient from the decoding result. (c) The DCT coefficient conversion unit 1052
The following processes (d) and (e) are performed depending on whether the extracted DCT coefficient is a coefficient of order (m-1) or less. (d) If the coefficient is less than (m-1) order, multiply it by m / n. This is because the image size is m
/ N. Note that n, m
For example, as a parameter, the code string conversion unit 10
5 is given first. When reducing the transmission rate of the second code string, the value of m (n> m) is reduced. (e) If the coefficient is of the order m or more, discard the DCT coefficient. (f) The motion vector conversion unit 1053 includes the variable length decoding unit 10
51 multiplies the extracted motion vector by m / n. (g) The code string generation unit 1054 includes the DCT coefficient conversion unit 105
2 transforms the new DCT coefficients and the motion vector transform unit 10
53 inputs the converted new motion vector and superimposes them. (h) Next, the result of the superposition of the DCT coefficient and the motion vector is variable-length coded and output as a second code string.

In the above embodiment, the image size is converted on the DCT coefficient, so that the first code string can be converted to a lower-speed transmission path without returning to a digital image signal.
The case where the reproduction unit converts the data into a second code string that can be transmitted and reproduced in real time and transmits and reproduces the data has been described. However, the number of frames per time and the number of quantization steps are converted to convert the code string. What can be done can be easily realized.

In the above-described embodiment, DC-2, which is represented by MPEG-2 as a high-efficiency coding method, is used.
Although the case of the coding method using the T coefficient and the motion vector has been described, other high-efficiency coding methods can be similarly considered.

[0027]

DESCRIPTION OF THE PREFERRED EMBODIMENTS A bit stream (code sequence) encoded by MPEG-2 is not converted back to a digital image signal,
An example of conversion to a low bit rate code string will be described.

The bit rate of the MPEG-2 bit stream before conversion is R, the rate occupied by the header information in the bit stream is R _H , the rate related to the DCT coefficient is R _{DC T} , and the rate related to the motion vector is R _MV . Then
R is R = _RH + _RDCT + _RMV .

The encoding condition is 6 Mbit / s, SP ＠ ML, N
= 15, M = 1 and MPEG-2 encoded standard video "f
When the bit stream of "lower garden" was analyzed, _RH was 7%, _RDCT was 90%, and _RMV was 3%.
It was found that the DCT coefficients accounted for most of the data.

As for the image size, it is necessary to consider the effect of variable-length coding. However, if the DCT coefficient is deleted and the size is reduced to, for example, １ ／, the rate R ′ after conversion becomes R '= (1/4) × 0.9R, and the rate can be reduced to 1/3 to 1/4 at the maximum.

Actually, the standard video "flower gar
den "," mobile & calender ",
“Table tennis” is MPEG-2 encoded, and the high-frequency components of DCT coefficients of 8 × 8 pixels are deleted to obtain 4 bits.
An experiment was conducted to reduce the bit rate by leaving a low frequency component of × 4. At this time, the magnitude of the motion vector is also １ ／.
To form a bit stream. In order to decode an image after code conversion, an IDCT (inverse discrete cosine transform) of 4 × 4 size is required, and therefore, a unique decoder (decoding software) is prepared.

By the above code conversion, "flower
garden "is 3.09 Mbit / s and" mobile
& Calender ”is 2.95 Mbit / s,“ tab
Letennis "is 2.23 Mbit / s. These are about 1/2 to the original bit rate of 6 Mbit / s.
The rate is 1/3, which indicates that a lower bit rate has been achieved.

As described above, for example, the video content is
To realize a video content distribution system in which a bit stream is converted into a bit stream by EG-2 encoding and stored in a server for centralized management, an MPEG-2 bit stream is transmitted in real time over a broadband line and decoded at a receiving terminal. However, it solves the problem that the transmission rate required for real-time decoding cannot be secured on a narrow-band line, and converts the bit stream at a low bit rate by code conversion according to the bandwidth of the transmission path and the processing capability of the terminal. It can be directly converted to video and distributed in real time. In addition, it is possible to easily realize a service of transmitting a video captured in a mobile environment via a telephone line or the like using a portable terminal or the like and distributing a video from a remote place.

[0034]

As described above, according to the present invention, a code sequence obtained by encoding a digital image signal by the first encoding method is encoded at a lower encoding rate. 2 and can be transmitted and reproduced in real time by the connected transmission path and reproduction unit. As a result, the circuit configuration can be reduced when the coded image is transmitted and reproduced in real time by the transmission path and reproduction unit in various states. Further, since there is no need to perform the encoding process again, it is possible to realize a higher processing speed.

[Brief description of the drawings]

FIG. 1 is a block diagram of the present invention.

FIG. 2 is a block diagram showing an embodiment of the present invention.

FIG. 3 is a flowchart of a code string conversion unit according to the embodiment of the present invention.

FIG. 4 is a diagram illustrating a configuration example of a conventional method of code string conversion.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 101 Digital image signal 102 Encoder which performs 1st encoding 103,104 Transmission path 105 Code sequence conversion part 106 Decoder which decodes 1st code sequence 107 Decoder which decodes 2nd code sequence 108 Decoding software Storage unit 109, 110 Switcher 111 Decoding software

Continuing from the front page (72) Inventor Satoshi Ishibashi 3-19-2 Nishi-Shinjuku, Shinjuku-ku, Tokyo Japan Telegraph and Telephone Corporation (72) Inventor Naoki Kobayashi 3-192-1, Nishi-Shinjuku, Shinjuku-ku, Tokyo Nippon Telegraph and Telephone Telephone Co., Ltd. (72) Inventor Tomio Kishimoto 20-1 Sakuragaoka-cho, Shibuya-ku, Tokyo Inside NTT Electronics Corporation

Claims (5)

[Claims] In a method for reproducing an encoded image, a first code string obtained by compressing and encoding a digital image signal by a first encoding method is converted into a second code string without returning to a digital image signal. Converting a second code string obtained by the above-described coding method, sending the converted second code string to a decoder, decoding the second code string, and reproducing a digital image signal. And a method for reproducing an encoded image. 2. A method for reproducing an encoded image, comprising: converting a first code string obtained by compressing and encoding a digital image signal according to a first encoding method into a second image signal without returning it to a digital image signal; A step of converting the second code string obtained by the above-mentioned coding method, a step of sending the converted second code string and a decoding execution unit capable of decoding this code string to a decoder, Decoding the second code string using the decoding execution unit and reproducing a digital image signal. 3. The first code string is a code string encoded using a DCT coefficient and a motion vector.
In the process of converting the first code string into a variable-length code string, the first code string is subjected to variable-length decoding, DCT coefficients and motion vectors are extracted, and a predetermined number or a specified number of low-frequency component coefficients are extracted from the extracted DCT coefficients. Only, and transforms the value of the extracted motion vector according to the reduction of the DCT coefficient.
3. The method for reproducing an encoded image according to claim 1, wherein the transformed DCT coefficients and the motion vector are subjected to variable-length encoding, and header information is added to form a second code string. 4. A code string conversion device for converting a first code string obtained by compressing and coding a digital image signal by a first coding method into a code string obtained by a second coding method. Variable length decoding means for inputting the first code string and performing variable length decoding to extract DCT coefficients and motion vectors; and a predetermined number or a specified number of low frequency components of the extracted DCT coefficients. DCT coefficient transforming means for leaving only the coefficients of the above and discarding other coefficients, a motion vector transforming means for transforming the value of the extracted motion vector in accordance with the reduction of the DCT coefficients, and a transform by the DCT coefficient transforming means. The variable-length coding is performed on the obtained DCT coefficient and the motion vector converted by the motion vector conversion means,
A code string generating unit that adds header information to generate a second code string. 5. A recording medium for recording a program for converting a first code string obtained by compressing and coding a digital image signal by a first coding method into a code string obtained by a second coding method. A medium for inputting the first code string, performing variable-length decoding, extracting DCT coefficients and motion vectors, and extracting a predetermined number or a specified number of low-frequency components from the extracted DCT coefficients. A process of leaving only the coefficients and discarding the other coefficients, a process of transforming the value of the extracted motion vector in accordance with the reduction of the DCT coefficient, and a variable-length encoding of the converted DCT coefficient and the motion vector, A recording medium for a code string conversion program, wherein a program for causing a computer to execute a process of adding header information to form a second code string is recorded.