JPH0993133A - Coding form converter - Google Patents

Abstract

PROBLEM TO BE SOLVED: To reduce the conversion processing time by generating block data of conversion destination through direct calculation of block data subjected to orthogonal transformation in the video information re-configuration process of the conversion destination. SOLUTION: The converter is made up of a conversion source video information 101, a block data extract section 102, an inverse quantization section 103, a quantization section 107, a re-configuration section 108 for conversion destination video information, a conversion destination video information section 109, and an inter-frame arithmetic section 110 for coded block data. The block data extract section 102 extracts desired coded block data and the inverse quantization section 103 applies inverse quantization to the data and the block data subjected to orthogonal transformation are calculated directly by the inter- frame arithmetic section 110 to generate coded block data of a conversion destination, the quantization section 107 makes quantization and the re- configuration section 108 re-configures the video information of the conversion destination to obtain the conversion destination video information 109 and the orthogonal transformation is omitted.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a frame rate, a bit rate, an encoding format from the same video information in a video information providing system such as video-on-demand, depending on the processing capacity and transmission speed of a terminal device. The present invention relates to a coding format conversion device that provides high speed multiplex conversion of different video information.

[0002]

2. Description of the Related Art FIG. 3 shows a conventional coding format conversion apparatus, in which reference numeral 101 is conversion source video information, 102 is a block data extraction unit, 103 is an inverse quantization unit, and 104 is an inverse orthogonal transformation unit. , 105 is an inter-frame operation unit of uncompressed coded block data, 106 is an orthogonal transformation unit, 107 is a quantization unit, 108 is a reconstruction unit of conversion destination video information, and 109 is conversion destination video information.

In a conventional coding format conversion apparatus, as shown in FIG. 3, block data is extracted from video information of a conversion source, subjected to inverse quantization, and then subjected to orthogonal transformation (for example, inverse transformation) opposite to that at encoding. DCT) is performed to generate difference block data between frames before compression encoding, and then calculation between frames is performed to generate raw frame data. Next, the inter-frame difference data of the raw frame data is calculated at a bit rate according to the transmission speed to generate difference block data, which is subjected to orthogonal transformation (for example, DCT transformation) and then quantized to obtain a transformation destination. Reconstruct the data stream of video information according to the format.

[0004]

In the conventional apparatus, the encoding format was converted by using a dedicated apparatus called a codec used in a TV phone or a video CD. These dedicated devices include an encoder that orthogonally transforms and quantizes uncompressed video information to generate coded video information, and an inverse quantization and inverse orthogonal transform of the coded video information to uncoded video information. It is a combination with a decoder that restores information. In general, since the codec is black-boxed, neither the encoder nor the decoder provides a function of inputting / outputting block data that is only orthogonally transformed and is temporarily generated during the process.

Furthermore, it is necessary to perform real-time processing from the capture of video signals to the storage (transmission) of encoded information, and the reading (reception) of encoded video information from the screen display. In general, in order to speed up the processing in these codecs, the process from extraction of block data to quantization, orthogonal transformation, and generation of uncompressed coded block data is processed as a series of processes, so in the middle of processing It has been difficult to add a function to take out orthogonally transformed block data that is temporarily generated.

It is an object of the present invention to make it possible to economically realize a bit rate and a coding format conversion according to a transmission speed and a terminal processing capacity by a high speed multiplexing process.

[0007]

FIG. 1 shows the configuration of an embodiment of the present invention. Reference numeral 101 in the figure is conversion source video information, 1
02 is a block data extraction unit, 103 is an inverse quantization unit,
Reference numeral 107 denotes a quantizer, 108 denotes a reconstructing unit for conversion destination video information, 109 denotes conversion destination video information, and 110 denotes an interframe arithmetic unit for encoded block data.

In the present invention, as shown in FIG. 1, the block data extraction unit 102 extracts desired encoded block data from the frame data of the video information 101 of the conversion source, and the inverse quantization unit 103. After dequantizing at
The inter-frame operation unit 110 directly operates the block data that has been subjected to the orthogonal transformation to generate the encoded block data of the conversion destination, which is quantized by the quantization unit 107, and then the reconstruction unit 108 The conversion for obtaining the conversion-destination video information 109 is realized by reconstructing the video information, and the orthogonal conversion, which requires a large amount of processing, is omitted.

[0009]

DESCRIPTION OF THE PREFERRED EMBODIMENTS The operation of the present invention will be described below with reference to the drawings. FIG. 2 is a block diagram showing an embodiment of the present invention. Hereinafter, with reference to FIG. The process of converting to 261 data will be described.

First, the conversion source MPEG1 data is read by the input control unit 12 by the function of the control unit 11, and the analysis control unit 13 makes the block data in the first I picture (in DCT conversion, it becomes a DCT coefficient. After that DCT
(Referred to as a coefficient) is extracted, inversely quantized by the quantization control unit 14, and recorded in the first frame data storage unit 151 of the frame data storage control unit 15.

At the same time, by the operation of the control unit 11, the quantization is performed by the quantization control unit 14 and then sent to the data generation control unit 16 to be reconstructed as the video information of the conversion destination, and stored in the file by the output control unit 17. It

Also for the second I picture, the extracted DCT coefficient is dequantized in the same manner as the first I picture, and is stored in the first frame data storage unit 151 of the frame data storage control unit 15 by the function of the control unit 11. The difference between the calculated DCT coefficient and the DCT coefficient is obtained, and the result is the third frame data storage unit 153.
Memorize to The frame data accumulated in the third frame data storage unit 153 is sent to the data generation control unit 16 by the function of the control unit 11, reconfigured as video information of a conversion destination, and additionally stored in a file by the output control unit 17. To be done. At the same time, the extracted DCT coefficient is overwritten and stored in the first frame data storage unit 151.

In the case of the present invention, the frame rate and the bit rate can be adjusted by adjusting the intervals of the I pictures extracted by the input control unit 12 and the analysis control unit 13. In addition, the quantization control unit 14 can control the bit rate by changing the quantization characteristic value.

Next, a conversion process for P and B pictures of MPEG1 will be described. First, the source MPEG
One data is read from the input control unit 12 by the function of the control unit 11, the DCT coefficient in the first I picture is extracted by the analysis control unit 13, is dequantized by the quantization control unit 14, and is stored as frame data. It is stored in the first frame data storage unit 151 of the control unit 15. At the same time, the control unit 1
1 is sent to the data generation control unit 16 and the H. It is reconstructed as 261 video information and stored in a file by the output control unit 17. Subsequent processing differs depending on the type of the extracted frame.

The method of calculating the DCT coefficient when the extracted frame is an I or P picture differs depending on the macro block type. When the extracted frame is an I picture and P
In the case of a block for intra-frame predictive coding of a picture, D stored in the first frame data storage unit 151
The difference between the CT coefficient and the extracted DCT coefficient is calculated and stored in the third frame data storage unit 153. At the same time, the extracted DCT coefficient is stored in the second frame data storage unit 152.

In the case of a block for forward interframe predictive coding of a P picture, the first frame data storage unit 1
The extracted DCT coefficient is added to the DCT coefficient stored in 51 and stored in the second frame data storage unit 152. At the same time, the extracted block data is stored in the third frame data storage unit 153.

When the processing of all blocks is completed, the H.264 data from the block data of the third frame data storage unit 153 is processed. 261 data is generated and additionally stored in the file. After that, the extracted I and P pictures are stored in the first frame data storage unit 151 and the second frame data storage unit 152.
Alternately and are processed. That is, the processed I,
If the P picture (including the leading I picture) is even, it is as described above, but if it is odd, it is the first
The frame data storage unit 151 and the second frame data storage unit 152 are replaced.

For example, when the extracted frame is an I picture and a block of intra-picture predictive coding of a P picture, the DCT coefficient stored in the second frame data storage unit 152 instead of the first Difference of the third
Frame data storage unit 153. At the same time, the extracted DCT coefficient is stored in the first frame data storage unit 151 instead of the second.

D when the extracted frame is a B picture
The method of calculating the CT coefficient differs depending on the macro block type and is as follows. If the extracted block is a macroblock for intra-frame predictive coding, the difference between the DCT coefficient stored in the first frame data storage unit 151 and the extracted DCT coefficient is calculated and stored in the third frame data storage unit 153. Remember.

In the case of a forward interframe prediction macroblock, the extracted block data is stored in the third frame data storage unit 153. In the backward inter-frame prediction macroblock, the first frame data storage unit 15
The extracted DCT coefficient is added to the DCT coefficient stored in 1. The result is subtracted from the DCT coefficient stored in the second frame data storage unit 152 and stored in the third frame data storage unit 153.

In the interpolative prediction macroblock, the extracted DCT coefficients are added to the DCT coefficients stored in the first and second frame data storage units and averaged. The result is stored in the third frame data storage unit 153.

When the processing of all the blocks is completed, the H.264 data from the block data of the third frame data storage section 153 is deleted. 261 data is generated and additionally stored in the file. Thereafter, the extracted B picture is processed by alternately changing the first frame data storage unit 151 and the second frame data storage unit 152 according to the number of I and P pictures processed so far. That is, if the number of I and P pictures (including the first I picture) processed so far is odd, the processing is as described above, but if it is even, the first frame data storage unit and the second Replace the frame data storage. For example, in the case of the intraframe predictive coding block, the difference from the DCT coefficient stored in the second frame data storage unit is stored in the third frame data storage unit.

In the above description, MPEG to H.264. 261
Although the example in which the frame rate and the bit rate are reduced by the conversion to the above has been described, the present invention converts between the same encoding formats (for example, MPEG to MPEG conversion), increases the frame rate and the bit rate (for example, 64 kbps). H.
It is also clear that it can be applied to the conversion from H.261 to MPEG of 1.5 Mbps).

[0024]

As described above, according to the present invention,
When decoding the video information of the conversion source, it is possible to omit the inverse orthogonal transform, which requires a large amount of processing, up to the stage of inverse quantization of block data. Furthermore, in the process of reconstructing the video information of the conversion destination, since the block data of the conversion destination is generated by the direct operation of the block data subjected to the orthogonal conversion, it is possible to omit the orthogonal conversion processing and the search time of the motion vector, so that the conversion processing time is significantly increased. Can be shortened.

[Brief description of drawings]

FIG. 1 is a diagram showing a configuration of the present invention.

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

FIG. 3 is a diagram showing a conventional configuration.

[Explanation of symbols]

 Reference Signs List 101 conversion source video information 102 block data extraction unit 103 inverse quantization unit 107 quantization unit 108 conversion destination video information reconstruction unit 109 conversion destination video information 11 control unit 12 input control unit 13 analysis control unit 14 quantization control unit 15 frame data storage control section 151-153 frame data storage section 16 data generation control section 17 output control section

 ─────────────────────────────────────────────────── ─── Continuation of the front page (72) Inventor Akio Kameda 1-1-6 Uchisaiwaicho, Chiyoda-ku, Tokyo Nihon Telegraph and Telephone Corporation

Claims (1)

[Claims] 1. An apparatus for performing mutual conversion of a representation format and a bit rate between different data regarding video information which is compression-encoded by orthogonal conversion such as DCT and quantization of difference data between frames in block units. Means for extracting compression-encoded block data from the original information, means for dequantizing block data, means for computing dequantized block data between frames, and generating new block data, A code having means for quantizing newly generated block data according to a format of a conversion destination, and means for reconstructing a data stream according to a representation format of the conversion destination encoded video data. Format conversion device.