JP2001078199A - Video signal coder - Google Patents

Abstract

PROBLEM TO BE SOLVED: To obtain a video signal coder that converts in-frame coded compressed data into inter-farm coded compressed data at high speed. SOLUTION: This coder is provided with a VLD section 2 that applies variable length decoding to received compression data 1 to generate a quantization DCT coefficient, an inverse quantization section 3 that applies inverse quantization to the quantization DCT coefficient to generate a DCT coefficient, a resize IDCT section 4 that change the size of the DCT coefficient and an image format to generate block data, a DC component extract section 5 that extracts only DC components among the DCT coefficients, a DC comport 6 that stores a DC component by one frame as a reference DC component, a moving direction detection section 11 that compares a DC comport of an adjacent frame with the reference DC component to detect a moving direction of the block data, and an encode section 12 that encodes the block data by one frame on the basis of a motion vector detected from the moving direction and generates output compression data 7.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to video signal encoding for converting image data compressed by an intra-frame encoding method into image data of a compression format using both intra-frame encoding and inter-frame predictive encoding. Related to the device.

[0002]

2. Description of the Related Art As a conventional video signal encoding apparatus, an apparatus described in Japanese Patent Application Laid-Open No. 10-164593 is known. Hereinafter, the conventional video signal encoding apparatus will be described with reference to a block diagram of FIG. In the following description, a case where the output compressed data is MPEG image data (hereinafter, referred to as MPEG data) will be described as an example.

In FIG. 10, input compressed data 101
Are rearranged in the image rearranging unit 102 in a compressible order according to an MPEG image compression algorithm. The memory 103 stores the rearranged input compressed data 10
1 is kept compressed. The intra-frame compressed data decoding system 104 decodes the input compressed data 101 in the order rearranged by the image rearrangement unit 102. The inter-frame compressed data encoding system 105 compresses the decoded image data in the order of decoding, and generates output compressed data 106.

For example, if the input compressed data 101 is I0, I0
When the images are input in the order of 1, I2, I3, I4, I5, I6, I7, I8,...
3, I1, I2, I6, I4, I5, I9, I7, I8
.. Are rearranged in an order necessary for MPEG encoding. The rearranged input compressed data 101 is stored in the memory 1
03. The input compressed data 101 held in the memory 103 is decoded by the intra-frame compressed data decoding system 104 in the rearranged order to become image data. This image data is subjected to I0, P3, B1, B2, P6, B4, B
5, P9, B7, B8 ...
G data is generated and output as output compressed data 106. Here, I is intra-frame coded image data (hereinafter, referred to as
P indicates inter-frame forward prediction coded image data (hereinafter referred to as P picture), and B indicates bidirectional prediction coded image data (hereinafter referred to as B picture).

[0005]

In a conventional video signal encoding apparatus, an intra-frame compressed data decoding system for decoding intra-frame encoded compressed data and an inter-frame compressed data encoding system for inter-frame encoding are provided. Are completely separated. Therefore, the processing time of each of the decoding processing and the encoding processing becomes the conversion time of the compressed data as it is, and it is difficult to perform high-speed processing. Further, when the image size and format of the input compressed data and the output compressed data are different from each other, it is necessary to perform a filtering process on the decoded video data and convert the image size and the image format, which further increases the processing time. There's a problem.
Also, like DV (Digital Video) format,
When decoding the compressed data obtained by interleaving and compressing the field data, there is a problem that the boundary of the object is mismatched in a portion where the motion of the video is large, and the boundary becomes unnatural.

SUMMARY OF THE INVENTION It is an object of the present invention to provide a video signal coding apparatus for converting intra-frame compressed data to inter-frame compressed data at a high speed.
It is another object of the present invention to provide a video signal encoding device that decodes only data in one field and performs inter-frame encoding when decoding intra-frame compressed data.

[0007]

SUMMARY OF THE INVENTION A video signal encoding apparatus according to the present invention decodes input compressed data by variable length decoding and performs
VLD means for generating T coefficients, inverse quantization means for inversely quantizing the quantized DCT coefficients to generate DCT coefficients, resize IDCT means for converting the size and image format of the DCT coefficients to generate block data, DC
DC component extracting means for extracting only the DC component of the T coefficient, a DC component memory for holding the DC component for one frame as a reference DC component, comparing the DC component of an adjacent frame with the reference DC component, A motion direction detecting means for detecting a motion direction of the block data; detecting a motion vector from the motion direction;
Encoding means for encoding the block data for a frame and outputting compressed data.

With this configuration, the size and format of the image obtained by decoding the input compressed data are converted by the resize IDCT means into the format required by the encoding means. The filter processing for making the change is unnecessary, and the amount of calculation can be reduced. Also, in order to detect the motion direction of each block or each macroblock by the motion direction detection means, when obtaining an accurate motion vector,
What is necessary is just to search around the previously detected motion direction. Therefore, the search range is narrower than when searching in all directions, so that the amount of calculation can be reduced. As a result, it is possible to obtain a video signal encoding device that converts input compressed data compressed within a frame into output compressed data compressed between frames at a high speed.

A video signal encoding apparatus according to another aspect of the present invention performs variable length decoding of input compressed data and performs quantization DCT.
VLD means for generating coefficients, inverse quantization means for inversely quantizing the quantized DCT coefficients to generate first DCT coefficients,
A DCT coefficient memory for holding the first DCT coefficient as a reference DCT coefficient, and a first difference DC obtained by calculating a difference between a first DCT coefficient of an adjacent frame and the reference DCT coefficient
It is determined whether or not to detect the motion direction from the T coefficient. The first DCT coefficient is output when the motion direction is detected or the intra-frame compression is performed. A motion determining unit that outputs a differential DCT coefficient; a second differential DCT coefficient or a second D that converts the size and image format of the first differential DCT coefficient or the first DCT coefficient that performs intra-frame compression;
DCT coefficient converting means for generating CT coefficients, and resizing I for generating block data by changing the size and image format for the first DCT coefficients for performing motion detection
DCT means, DC component extracting means for extracting only a DC component among the first DCT coefficients, and the D component for one frame.
A DC component memory for holding the C component as a reference DC component,
Comparing a DC component of an adjacent frame with the reference DC component, detecting a motion direction of the block data, detecting a motion vector from the motion direction,
Encoding means for encoding the block data, the second differential DCT coefficient, or the second DCT coefficient as a DCT coefficient based on the motion vector, and outputting output compressed data.

With this configuration, since the size and format of the image obtained by decoding the input compressed data are converted by the resize IDCT means into the format required by the encoding means, the conversion between the image size and the image format is performed. Becomes unnecessary, and the amount of calculation can be reduced. Further, for a block for which the motion direction has not been detected by the motion determining means, the DCT coefficient is converted into an image size and an image format without converting the DCT coefficient back to the image data to generate the DCT coefficient. It becomes possible. Further, in order to detect the motion direction of each block or each macroblock by the motion direction detection means, it is only necessary to search around the previously detected motion direction when obtaining an accurate motion vector. Since the search range is narrower than in the case where the search is performed, the amount of calculation can be reduced. As a result, it is possible to obtain a video signal encoding device that converts input compressed data compressed within a frame into output compressed data compressed between frames at a high speed.

[0011] A video signal encoding apparatus according to still another aspect of the present invention comprises a VLD means for variable-length decoding input compressed data to generate a quantized DCT coefficient, and dequantizes the quantized DCT coefficient to obtain a first signal. An inverse quantization means for generating a DCT coefficient, a DCT coefficient memory for holding the first DCT coefficient as a reference DCT coefficient, and determining whether or not to detect a motion direction according to a mode of the first DCT coefficient. When the direction detection or intra-frame compression is performed, the first DCT coefficient is used. When the motion direction detection is not performed, the first DCT coefficient obtained by taking the difference between the first DCT coefficient of an adjacent frame and the reference DCT coefficient is used. DC that outputs differential DCT coefficients
T mode discriminating means, a DCT coefficient for generating a second differential DCT coefficient or a second DCT coefficient by converting the image format and the size of the first differential DCT coefficient or the first DCT coefficient for performing intra-frame compression Conversion means,
A resize ID for converting the size and image format of the first DCT coefficient determined to detect the direction of motion by the DCT mode determining means to generate block data
CT means, DC component extracting means for extracting only a DC component among the first DCT coefficients, the DC component for one frame
A DC component memory that holds the component as a reference DC component, a motion direction detector that compares a DC component of an adjacent frame with the reference DC component, and detects a motion direction of the block data, and a motion vector from the motion direction. Detecting the block data based on the motion vector;
The encoding apparatus includes an encoding unit that encodes the second differential DCT coefficient or the second DCT coefficient as a DCT coefficient and outputs output compressed data.

With this configuration, since the image size and the image format obtained by decoding the input compressed data are converted into the format required by the encoding unit by the resize IDCT unit, the image size and the image format are converted to the required format. Filter processing for making the change is unnecessary, and the amount of calculation can be reduced. Also, for the blocks for which the motion direction is not detected by the DCT mode determining means, the image size and image format are converted without generating the DCT coefficients into image data to generate DCT coefficients, thereby reducing the amount of calculation. It is possible to do. Further, in order to detect the motion direction of each block or each macroblock by the motion direction detection means, it is only necessary to search around the previously detected motion direction when obtaining an accurate motion vector. Since the search range is narrower than in the case where the search is performed, the amount of calculation can be reduced. As a result, it is possible to obtain a video signal encoding device that converts input compressed data compressed within a frame into output compressed data compressed between frames at a high speed.

[0013]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS A preferred embodiment of a video signal encoding apparatus according to the present invention will be described below with reference to FIGS.

Embodiment 1 FIG. 1 is a block diagram showing a configuration of a video signal encoding device according to Embodiment 1 of the present invention. In the following description, a configuration will be described in which compressed data of the input compressed data 1 compressed in the DV (digital video) format is compressed into the MPEG format and output as output compressed data.

In FIG. 1, in the video signal encoding apparatus of the present embodiment, an output terminal of a VLD unit 2 for performing variable length decoding of input compressed data 1 is connected to an input terminal of an inverse quantization unit 3.
An output terminal of the inverse quantization unit 3 is connected to an input terminal of a resize IDCT (inverse discrete cosine transform) unit 4 and an input terminal of a DC component extraction unit 5, respectively. The output end of the resize IDCT unit 4 is the encoding unit 1
2 input terminals. The output terminal of the DC component extraction unit 5 is connected to the input terminal of the DC component memory 6 and the movement direction detection unit 1.
1 are connected to the first input terminals. An output terminal of the DC component memory 6 is connected to a second input terminal of the movement direction detector 11. The output terminal of the encoding unit 12 is connected to a third input terminal of the movement direction detection unit 11, and the output terminal of the movement direction detection unit 11 is connected to another input terminal of the encoding unit 12. Output compressed data 7 is output from the output end of the encoding unit 12.

The operation of the video signal encoding apparatus according to the first embodiment will be described below with reference to FIGS. FIG.
, The VLD unit 2 performs variable length decoding on the input compressed data 1 and generates a quantized DCT coefficient. The inverse quantization unit 3 calculates V
Inverse quantization of the quantized DCT coefficient decoded by the LD unit 2,
Generate DCT coefficients. The resize IDCT unit 4 performs IDCT operation and deshuffling on the DCT coefficient generated by the inverse quantization unit 3 to generate pixel data.

FIG. 2 is a diagram showing the horizontal and vertical arrangement of the luminance signal Y and the color difference signals Cr and Cb of the pixel data in the image format. FIG. 2A shows a 4: 1: 1 format. , (B) shows a 4: 2: 2 format, and (c) shows a 4: 2: 0 format. In FIG. 2, the luminance signal Y is represented by a circle, and the color difference signals Cr and Cb are represented by a square. The image data in the DV format, which is the input compressed data 1, has the image format of 4: 1, as shown in FIG.
1, and the image size is such that the luminance signal Y is 720 × 48
0, the color difference signal Cr and the color difference signal Cb are 180 × 480.

FIG. 3 is a diagram showing an arrangement of DCT coefficients in a DCT block. FIG.
8D shows the 8-8 DCT mode when the difference between the T coefficients is small;
(B) shows the 2-4-8 DCT mode when the DCT coefficient difference between the fields is large. In the figure, each square represents the position of the DCT coefficient of the determinant.

For the above two modes, 8-8 DC
In the T mode, as shown in FIG.
Of the eight lines in block 20, 1, 3, 5, 7 from the top
Lines (odd lines) are composed of image data of the first field, and lines 0, 2, 4, and 6 (even lines) are second lines.
It consists of field image data. A DCT coefficient is obtained by subjecting this data to a two-dimensional 8 × 8 DCT operation. In the 2-4-8 DCT mode, the two-dimensional 4 ×
The result of the 8DCT operation is the DCT coefficient. As shown in FIG. 3B, two DCT coefficients 21A and 21B
Are obtained by adding the components of
When the difference between the components is obtained, the DCT coefficient of field 2 is obtained.

When this image data is converted into MPEG1 format data, the image format is converted into a 4: 2: 0 format shown in FIG. 2C, and the luminance signal Y data is converted into an image size. 360 × 240
The data of the pixel and the color difference signals Cr and Cb is 180 × 120
It needs to be converted to pixels. When converting the data into MPEG2 format data, the image format is converted into the 4: 2: 2 format shown in FIG.
Regarding the image size, it is necessary to convert the data of the luminance signal Y to 720 × 480 pixels and the data of the color difference signals Cr and Cb to 360 × 480 pixels.

The image format is set to 4: 2: 0 format, the image size is set to a luminance signal Y of 360 × 240 pixels,
A case where the color difference signals Cr and Cb are converted so as to be 180 × 120 pixels will be described with reference to FIG. FIG.
FIG. 3 is a diagram illustrating a method of IDCT calculation of DCT coefficients for generating data of a luminance signal and a color difference signal in each mode. FIG. 4A shows 8-8 DC of the luminance signal data.
A calculation method of a coefficient in the T mode is shown, and (b) shows a calculation method of a coefficient in the 2-4-8 DCT mode of the data of the luminance signal. FIG. 4C shows the 8-8D data of the color difference signal.
A method of calculating coefficients in the CT mode is shown, and (d) shows a method of calculating coefficients in the 2-4-8 DCT mode of data of color difference signals. In FIG. 4, the image size is half the size of the luminance signal Y in both the vertical and horizontal directions, and the color difference signals Cr and Cb are 1/4 in the vertical direction and the horizontal size is the same.

First, generation of data of the luminance signal Y will be described. In the case of the 8-8 DCT mode, FIG.
DCT for four columns on the lower side
The IDCT operation is performed on the coefficients at eight points in each field.
In FIG. 4, the DC at the position of the mesh where the intersection of the thick line exists
An IDCT operation is performed on the T coefficient. To extract the data of field 1, the data of the odd-numbered lines are extracted, and to extract the data of field 2, the data of the even-numbered lines are respectively subjected to four IDCT operations, and the luminance signal Y is obtained.
Generate data for FIG. 4A shows that an operation is performed on data of an even-numbered line. 2-4-8DC
In the case of the T mode, four rows and four columns of the lower rows of the field 1 or the field 2 indicated by the bold line in FIG.
Perform a two-dimensional 4 × 4 IDCT operation on the CT coefficients,
The data of the luminance signal Y is generated.

Next, generation of data of the color difference signals Cr and Cb will be described. In the case of the 8-8 DCT mode, four IDCT operations are performed on the four lower-order points of the entire column of DCT coefficients indicated by the intersections of the vertical thick lines in FIG. Further, an 8-point IDCT operation is performed on the data of line 0 and line 2 or the data of line 1 and line 3,
The data of the color difference signals Cr and Cb is generated. 2-4-8
In the case of the DCT mode, two-dimensional 2 × 8 IDCT is performed on the lower order 2 × 8 DCT coefficient of the DCT coefficient of field 1 or field 2 shown by the bold line in FIG.
An operation is performed to generate data of the color difference signals Cr and Cb.

Next, the image format is set to 4: 2: 2,
The image size is set to 720 × 480 pixels for the luminance signal Y data and 360 × 480 for the color difference signals Cr and Cb.
The case of conversion into pixels will be described with reference to FIG.
5A and 5B are diagrams illustrating a method of IDCT calculation of DCT coefficients for generating data of a luminance signal and a color difference signal in each mode in this case. FIG.
-8 DCT coefficient, (b) is the luminance signal data 2-4-
8DCT coefficient, (c) is an 8-8 DCT coefficient of the color difference signal,
(D) shows each method of calculating the 2-4-8 DCT coefficient of the color difference signal. In this case, the image size is the luminance signal Y
Is the same size as that described above, and the color difference signals Cr and Cb are twice as large as they are vertically.

First, generation of data of the luminance signal Y will be described. In the case of the 8-8 DCT mode, FIG.
As shown in (2), a two-dimensional 8 × 8 IDCT operation is performed to generate data of the luminance signal Y. In the case of the 2-4-8 DCT mode, as shown in FIG. 5B, the DCT coefficient of field 1 or field 2 is set as a low-order coefficient, and 0 is added to the high-order coefficient to obtain a two-dimensional DCT coefficient. An 8 × 8 IDCT operation is performed to generate data of the luminance signal Y. By adding 0 as a higher-order coefficient, the image format can be converted without changing the data.

Next, generation of data of the color difference signals Cr and Cb will be described. In the case of the 8-8 DCT mode, as shown in FIG. 5C, 0 is added to the higher order coefficient of the DCT coefficient, a two-dimensional 8 × 16 IDCT operation is performed, and the data of the color difference signals Cr and Cb are obtained. Generate 2-4-8D
In the case of the CT mode, as shown in FIG. 5D, 0 is added to the higher order coefficient of the DCT coefficient of the field 1 or the field 2, the two-dimensional 8 × 16 IDCT operation is performed, and the color difference signal Cr is obtained. And Cb data are generated. By adding 0 to the higher order side, the image size can be enlarged without changing the data content.

In FIG. 1, the DC component extraction unit 5 extracts only the DC component of the DCT coefficient generated by the inverse quantization unit 3, performs deshuffling, and outputs it to the DC component memory 6 and the motion direction detection unit 11. . The DC component memory 6 holds the DC component extracted by the DC component extraction unit 5 as a reference DC component.

The motion direction detecting section 11 includes an encoding section 12
To obtain the picture type information of the frame to be encoded. If the picture type is a P picture or a B picture, the motion direction is detected. The detection of the movement direction is performed by using each DC component extracted by the DC component extraction unit 5 and the reference D stored in the DC component memory 6.
This is performed for each block by comparing the C component. The detection of the movement direction will be described in detail with reference to FIG.

FIGS. 6A and 6B are diagrams for explaining a method of detecting a motion direction. FIG. 6A shows a method of setting a search range to an 8-pixel range, and FIG. 6B shows a method of setting a search range to a 16-pixel range. Here, a case where two pixels are selected as candidates for the movement direction will be described as an example. First, an absolute value of a difference between a DC component for detecting a motion direction, a DC component of a pixel located at the same position among reference DC components, and DC components of eight pixels around the DC component is calculated. Two of the absolute values are selected from the smaller one, and the direction in which each pixel of the two absolute values exists is determined by no motion or upper, lower, left, right, upper right, lower right, upper left, or This is indicated by the lower left direction, and this direction is defined as the movement direction.

FIG. 6A shows that the direction of motion is detected in units of 8 pixels in a range of 8 pixels when considered in units of DCT blocks. Arrows represent motion vectors. The search range can be further expanded, and the motion direction can be detected in a 16-pixel range. It is also possible to detect the motion direction in a wider search range such as a 24-pixel range. As shown in FIG. 6B, when the search range is expanded to a range of 16 pixels, the number of motion vectors expressing the motion direction increases.

In the MPEG format, a motion vector needs to be detected in units of macroblocks of 16 × 16 pixels. In the above detection method, two motion directions are detected in units of blocks of 8 × 8 pixels. Therefore, the direction most frequently used and the direction used secondly among the motion directions of the four blocks forming the macroblock are defined as the motion direction of the macroblock. In this case, the number of motion direction candidates is two, but one or three or more motion direction candidates can be similarly detected.

The encoding section 12 includes a motion direction detecting section 11
A search is made for a motion vector around the motion direction of the macroblock detected by the above to obtain a motion vector, and encoding is performed using the image data generated by the resize IDCT unit 4 to generate MPEG data. In addition, the resize IDCT unit 4 can perform the same for other combinations of image formats and image sizes. Also, the compression format of the input compressed data 1 is intra-frame coded compressed data such as motion JPEG, and the compression format of the output compressed data 7 is H.264. 263; 261 or the like may be compressed data that has been inter-frame coded.

Embodiment 2 FIG. 7 is a block diagram showing a configuration of a video signal encoding device according to Embodiment 2 of the present invention. Hereinafter, a configuration will be described in which compressed data compressed in the DV format, which is input compressed data, is converted into output compressed data in the MPEG format. As shown in FIG. 7, the video signal encoding device according to the second embodiment is configured as shown in FIG.
A motion determining unit 71 is inserted between the inverse quantization unit 3 and the resize IDCT unit 4. The same parts as those in the first embodiment are denoted by the same reference numerals, and duplicate description will be omitted.

A first input terminal of the motion discriminating unit 71 is connected to an output terminal of the inverse quantization unit 3, and four output terminals are an input terminal of the resize IDCT unit 4 and a first input terminal of the motion direction detecting unit 74, respectively. Input end, input end of DCT coefficient conversion section 73 and DCT
It is connected to the input terminal of the coefficient memory 72. The other two input terminals of the motion determining unit 71 are connected to the output terminal of the DCT memory 72 and the output terminal of the encoding unit 75, respectively. The output end of the DCT coefficient conversion unit 73 is the encoding unit 7
5 is connected to the first input terminal. An output terminal of the DCT coefficient conversion unit 73 is connected to an input terminal of the encoding unit 75. The other two input terminals of the movement direction detecting unit 74 are connected to the output terminals of the DC component extracting unit 5 and the DC component memory 6, respectively. An output terminal of the movement direction detecting unit 74 is connected to another input terminal of the encoding unit 75. Other configurations are the same as those in FIG.

Hereinafter, the operation of the video signal encoding apparatus according to the second embodiment will be described with reference to FIG. 5B, FIG. 7 and FIG. The VLD unit 2, the inverse quantization unit 3, the resize I
DCT unit 4, DC component extraction unit 5, and DC component memory 6
Are the same as those in the first embodiment, and a description thereof will be omitted. In FIG. 7, the DCT coefficient generated by the inverse quantization unit 3 is referred to as D
It is stored in the DCT coefficient memory 72 as a CT coefficient. The motion discrimination unit 71 compares the DCT coefficient generated by the inverse quantization unit 3 with the reference DCT stored in the DCT coefficient memory 72.
A difference is calculated for each element of the coefficient to generate a difference DCT coefficient. When taking the difference, when both the DCT coefficient and the reference DCT coefficient are in the 2-4-8 DCT mode, the difference is taken between the DCT coefficients of the same field, and one is taken as 2-4-8.
In the 8DCT mode, as shown in FIG.
The difference is calculated with the higher-order data set to 0.

Next, the absolute value of each element of the differential DCT coefficient is obtained, and the sum is calculated. When the sum of the absolute values is smaller than the predetermined value, it is determined that there is no motion, and a difference DCT coefficient is output. If the sum of the absolute values is larger than the predetermined value, it is determined that there is a motion, and the DCT coefficient and the position of the moving block are output. However, since MPEG data performs motion detection on a macroblock basis, if any one of a plurality of blocks forming a macroblock is determined to have motion, the block determined to have no motion is also determined to have motion. Check again.

The DCT coefficient memory 72 holds the DCT coefficients generated by the inverse quantization unit 3 as reference DCT coefficients. The DCT coefficient conversion unit 73 receives as input either the DCT coefficient of the frame to be encoded into an I picture by the encoding unit 75 or the difference DCT coefficient generated by the motion discrimination unit 71, and converts the size and format of the DCT coefficient. hand,
A new DCT coefficient is generated without changing the luminance and color difference data by the IDCT operation. If the size and format of DCT coefficients are the same, DC
Output T coefficient or difference DCT coefficient.

Here, the image size is reduced by half in both the vertical and horizontal directions.
Will be described with reference to FIG. 8A and 8B are diagrams for explaining this image size conversion method. FIG. 8A shows an array of four 8 × 8 DCT coefficient blocks 26 before conversion, and FIG. 8B shows an 8 × 8 pseudo DCT coefficient 27 after conversion. The following shows the sequence. As shown in FIG. 8, the lower 4 × 4 DCT coefficients of the four 8 × 8 DCT coefficient blocks 26 are extracted,
This is set as an 8 × 8 pseudo DCT coefficient 27. This 8x8
4 pseudo IDCT coefficients and 8 pseudo DCT coefficients
A new DCT coefficient is generated by performing a two-dimensional operation using a coefficient newly created by combining the DCT coefficients of the points.
Assuming that the coefficient at this time is Cn = cos (2π / 16),
An example is shown in the following equation.

[0039]

(Equation 1)

The movement direction detecting section 74 basically operates in the same manner as the movement direction detecting section 11 of FIG. The difference from the operation of the motion direction detecting unit 11 is that the motion discriminating unit 7 indicates the position of the DCT block judged to have motion by the motion discriminating unit 71.
The movement direction is detected only for the DC component corresponding to the output of No. 1. The encoding unit 75 obtains a motion vector by searching for a motion vector around the motion direction of the macroblock detected by the motion direction detection unit 74, and encodes using the image data generated by the resize IDCT unit 4. , MPEG data.

When an I picture is generated, the new DCT coefficient generated by the DCT coefficient conversion unit 73 is subjected to quantization and variable length coding, and converted to MPEG data.
In the P and B pictures, for the DCT block determined to have no motion by the motion determining unit 71, quantization and variable length coding are performed on the new differential DCT coefficient generated by the DCT coefficient converting unit 73, Convert to MPEG data.

The DCT coefficient conversion unit 73 similarly converts the size and format of the DCT coefficient by generating a new coefficient and performing a two-dimensional operation on the size of the DCT coefficient other than that described above. It is possible to generate various DCT coefficients. It is also possible to construct a high-speed algorithm based on a newly generated DCT coefficient and perform an operation. The compression format of the input compressed data 1 is intra-frame coded compressed data such as motion JPEG, and the compression format of the output compressed data 7 is H.264. 263; 261 or the like may be compressed data that has been inter-frame coded.

Embodiment 3 FIG. 9 is a block diagram showing a configuration of a video signal encoding apparatus according to Embodiment 3 of the present invention. Hereinafter, a configuration for compressing the input compressed data compressed in the DV format into the MPEG format will be described. As shown in FIG. 9, the video signal encoding apparatus according to the third embodiment includes a motion determining unit 71 of the video signal encoding apparatus shown in FIG.
This is replaced with a T-mode determination unit 91. Example 2
The same parts as those described above are denoted by the same reference numerals, and redundant description will be omitted. The DCT mode determination unit 91 determines whether or not to detect the motion direction from the mode of the DCT coefficient generated by the inverse quantization unit 3.

Hereinafter, the operation of the video signal encoding apparatus according to the third embodiment will be described with reference to FIG. The operations of the VLD unit 2, the inverse quantization unit 3, the resize IDCT unit 4, the DC component extraction unit 5, and the DC component memory 6 are the same as those described in the first embodiment, and the DCT coefficient memory 72, the DCT coefficient conversion unit 73, a movement direction detection unit 74, and an encoding unit 7
The operation of No. 5 is the same as that described in the second embodiment, and a description thereof will be omitted.

In the compressed data of the DV format, the DCT coefficient mode uses the 2-4-8 DCT mode when the difference between fields is large, and uses the 8-8 DCT mode when the difference is small. . D
When the DCT coefficient input from the inverse quantization unit 3 is in the 2-4-8 DCT mode, the CT mode determination unit 91 determines that the motion is large. That is, any one of the plurality of blocks forming the macroblock is 2-4.
-8 DCT mode and DCT coefficient memory 7
If any one of the plurality of blocks at the macroblock position located at the same position as the DCT block of the reference DCT coefficient held in 2 is in the 2-4-8 DCT mode, the DCT mode determining unit 91 D
It outputs CT coefficients and block position data.

In other cases, that is, when a plurality of blocks forming a macroblock and a plurality of blocks at the same macroblock position of the reference DCT coefficient held in the DCT coefficient memory 72 all have 8-8 DCT coefficients. Indicates that the DCT mode determining unit 91 determines that there is no motion,
A difference between each DCT coefficient and each reference DCT coefficient for each element is obtained, and a difference DCT coefficient is generated and output. Example 2
Similarly, the compression format of the output compressed data 7 is set to H.264.
263; 261 or the like may be compressed data that has been inter-frame coded.

In the DCT mode determining section 91,
After determining that there is no motion and generating a difference DCT coefficient, the absolute value of each difference is summed up. If this value is large, it is determined that there is a motion again, and a macro block including the block determined to have motion is formed. It is also possible to determine again that all the blocks to be moved are present and output the DCT coefficients and block positions.

[0048]

As described above in detail in each embodiment,
The video signal encoding device of the present invention has the following effects. Since the luminance data and the chrominance data obtained by converting the image size and the image format are generated at the time of decoding, it is not necessary to perform the filtering process on the decoded video, and the processing amount can be reduced. Furthermore, since the motion direction is detected for each block and a motion vector is obtained based on the detected motion direction, the processing amount can be reduced by narrowing the search range of the motion direction. Also, for a block determined to have no motion, a new DT coefficient is returned to the image data without returning the DCT coefficient to the image data.
By generating the CT coefficients, the processing amount can be reduced.

Further, the motion direction is detected only for the block determined to have motion, and the search range for obtaining a motion vector can be narrowed using this motion direction. Since the motion vector detection is performed only for the selected block, the number of blocks for which the motion vector is to be obtained can be reduced, and thus the processing amount can be reduced. In addition, since only the data of one field is decoded and encoded, it is possible to smooth the boundary of the object even in a place where the motion is large. As a result, it is possible to obtain a video signal encoding device that converts input compressed data compressed within a frame into output compressed data compressed between frames at a high speed.

[Brief description of the drawings]

FIG. 1 is a block diagram illustrating a configuration of a video signal encoding device according to a first embodiment of the present invention.

FIGS. 2A and 2B are diagrams showing an image format to which the first embodiment of the present invention is applied, wherein FIG. 2A shows a 4: 1: 1 format, FIG. 2B shows a 4: 2: 2 format, and FIG. )
Indicates a 4: 2: 0 format.

FIG. 3 is a diagram illustrating a DCT mode in the video signal encoding device according to the first embodiment of the present invention, where (a) illustrates an 8-8 DCT;
(B) shows a 2-4-8 DCT mode.

FIG. 4 shows a video signal encoding apparatus according to a first embodiment of the present invention in which the image format is 4: 2: 0, the image size is 360 × 240 for the luminance signal Y, and 18 for the color difference signals Cr and Cb.
FIGS. 7A and 7B are diagrams illustrating a method of an IDCT operation for converting each signal to 0 × 120, wherein FIG. 9A illustrates an operation method of an 8-8 DCT coefficient of the luminance signal Y, and FIG.
(C) shows a method of calculating the 8-8 DCT coefficient of the color difference signals Cr and Cb, and (d) shows a method of calculating the 2-4-8 DCT coefficient of the color difference signals Cr and Cb.

FIG. 5 shows a video signal encoding apparatus according to a first embodiment of the present invention, in which the image format is 4: 2: 2, the image size is 720 × 480 for the luminance signal Y, and 36 for the color difference signals Cr and Cb.
0x480 to generate image data
8A and 8B are diagrams illustrating a method of calculating a DCT coefficient, in which FIG. 7A illustrates a method of calculating an 8-8 DCT coefficient of a luminance signal Y, and FIG.
Shows the method of calculating the 2-4-8 DCT coefficient of the luminance signal Y, (c) shows the method of calculating the 8-8 DCT coefficient of the color difference signals Cr and Cb, and (d) shows the method of calculating the 2-8 DCT coefficient of the color difference signals Cr and Cb.
The method of calculating the -4-8 DCT coefficient will be described.

FIG. 6 is a diagram for explaining a motion direction of block data in the video signal encoding device according to the first embodiment of the present invention;
(A) shows a movement direction in a search range of 8 pixels,
(B) shows the movement direction in the range of 16 pixels.

FIG. 7 is a block diagram illustrating a video signal encoding device according to a second embodiment of the present invention.

FIG. 8 is a diagram illustrating a method of generating an 8 × 8 pseudo DCT coefficient in the video signal encoding device according to the second embodiment of the present invention, where (a) illustrates 8 × 8 DCT coefficients of four blocks; b) shows the generated 8 × 8 pseudo DCT coefficients.

FIG. 9 is a block diagram illustrating a video signal encoding device according to a third embodiment of the present invention.

FIG. 10 is a block diagram showing a conventional video signal encoding device.

[Explanation of symbols]

 1, 101 ... input compressed data 2 ... VLD section 3 ... inverse quantization section 4 ... resize IDCT section 5 ... DC component extraction section 6 ... DC component memory 7, 106 ... · Output compressed data 11, 74 · · · Motion direction detection unit 12, 75 · · · Encoding unit 71 · · · Motion determination unit 72 · · · DCT coefficient memory 73 · · · DCT coefficient conversion unit 91 · · · DCT mode Discriminator

Claims (13)

[Claims] 1. VLD (Variable Length Decoding) means for performing variable length decoding of input compressed data compressed in a frame to generate a quantized DCT (Discrete Cosine Transform) coefficient, The quantized DCT coefficient Inverse quantizing means for inversely quantizing the DCT coefficients to generate DCT coefficients; resizing IDCT (inverse discrete cosine transform) means for converting the size of the DCT coefficients and image format to generate block data; A DC component extracting means for extracting only a DC (direct current) component, a DC component memory for holding the DC component for one frame as a reference DC component, and comparing a DC component of an adjacent frame with the reference DC component. , A motion direction detection for detecting a motion direction of the block data. Means for detecting a motion vector from the motion direction, encoding the block data for one frame based on the motion vector, and outputting compressed data compressed between frames. Signal encoding device. 2. VLD means for performing variable length decoding of input compressed data compressed in a frame to generate a quantized DCT coefficient, and inverse quantization for dequantizing the quantized DCT coefficient to generate a first DCT coefficient. Conversion means, a DCT coefficient memory for holding the first DCT coefficient as a reference DCT coefficient, a first DCT coefficient of an adjacent frame and the reference DCT
It is determined whether or not to detect the motion direction from the first difference DCT coefficient obtained by calculating the difference between the coefficient and the first DCT coefficient. Is not performed, the first difference D
A motion discriminating means for outputting a CT coefficient; converting the size and the image format of the first differential DCT coefficient or the first DCT coefficient for performing intra-frame compression to obtain a second differential DCT coefficient or a second DCT coefficient; DCT coefficient converting means for generating, resizing IDCT means for converting the size and image format of the first DCT coefficient for performing motion detection to generate block data, extracting only a DC component from the first DCT coefficient DC component extracting means, a DC component memory for holding the DC component for one frame as a reference DC component, a motion for comparing a DC component of an adjacent frame with the reference DC component, and detecting a motion direction of the block data Direction detecting means, and a motion vector is detected from the motion direction, and a motion vector for one frame is detected based on the motion vector. A video signal code comprising encoding means for encoding the block data, the second differential DCT coefficient, or the second DCT coefficient as a DCT coefficient and outputting inter-frame compressed data. Device. 3. VLD means for variable-length decoding input compressed data compressed in a DV (Digital Video) format to generate a quantized DCT coefficient, and inversely quantizes the quantized DCT coefficient to obtain a first DCT coefficient. A DCT coefficient memory for holding the first DCT coefficient as a reference DCT coefficient; determining a mode of the first DCT coefficient; determining whether to detect a motion direction; When the direction detection or intra-frame compression is performed, the first DCT coefficient is used. When the motion direction detection is not performed, the first DCT coefficient of the adjacent frame is used.
And the first DCT coefficient obtained by calculating the difference between the DCT coefficient of
DCT mode discriminating means for outputting a differential DCT coefficient of the first DCT coefficient or a second DCT coefficient of the second DCT coefficient or a second DCT coefficient for converting the size and image format of the first DCT coefficient or the first DCT coefficient for performing intra-frame compression DCT coefficient conversion means for generating DCT coefficients; resize ID for converting the size and image format of the first DCT coefficient determined to detect the motion direction by the DCT mode determination means to generate block data
CT means; DC component extracting means for extracting only a DC component among the first DCT coefficients; DC component memory for holding the DC component for one frame as a reference DC component; DC components of an adjacent frame and the reference A motion direction detecting means for comparing a DC component and detecting a motion direction of the block data; and detecting a motion vector from the motion direction, the block data, the second differential DCT coefficient,
Alternatively, the video signal encoding apparatus includes encoding means for encoding the second DCT coefficient as a DCT coefficient and outputting compressed data compressed between frames. 4. The video signal encoding apparatus according to claim 1, wherein said input compressed data is in a format compressed by intra-frame encoding of JPEG format or DV format. 5. The output compressed data is MPEG data.
Format, H. H.263 format, or H.263 format. 2
The video signal encoding device according to claim 1, wherein the video signal encoding device is a format compressed by hybrid encoding combining intra-frame encoding and inter-frame predictive encoding in a 61 format. 6. The resize IDCT means includes an IDCT unit.
4. The video signal encoding apparatus according to claim 1, wherein interpolation or thinning is performed in the frequency domain by changing the order of the operation, and block data is generated by converting an image size and an image format. 7. The video signal encoding apparatus according to claim 1, wherein said resize IDCT means generates only data of one field by an IDCT operation. 8. The DCT coefficient conversion means performs data conversion without restoring one or a plurality of the first DCT coefficients or one or a plurality of the first differential DCT coefficients to a signal component. The video signal encoding apparatus according to claim 2, wherein one of the second DCT coefficients or one of the second differential DCT coefficients is generated. 9. The motion judging means judges whether or not to detect a motion direction for each DCT block, and judges whether to detect a motion direction of a macroblock based on a result of the judgment. 2. The video signal encoding device according to 1. 10. The video signal encoding apparatus according to claim 3, wherein said DCT mode determining means determines whether to detect a motion direction according to a mode of said first DCT coefficient. . 11. The video signal encoding apparatus according to claim 1, wherein said motion direction detecting means indicates a motion direction of said block data by one or a plurality of directions. 12. The apparatus according to claim 1, wherein said motion direction detecting means obtains a motion direction of a macroblock from a plurality of motion directions of said block data, and indicates the direction by one or more directions. 12. The video signal encoding device according to 3 or 11. 13. The apparatus according to claim 1, wherein said encoding means searches for a motion vector from a center of one or a plurality of movement directions detected by said movement direction detection means. Video signal encoding device.