JP5100689B2 - Video encoding apparatus and video decoding apparatus - Google Patents

Description

  The present invention relates to a moving image encoding device that encodes an interlaced moving image signal and a moving image decoding device that decodes encoded data in which the interlaced moving image signal is encoded. .

There are two types of moving image signals, progressive signals and interlace signals.
The progressive signal is called “sequential scanning”, and scans a scanning line used for a PC monitor or the like line by line from top to bottom.
On the other hand, the interlaced signal is called “interlaced scanning”, in which one frame is divided into two fields, and scanning is performed every other line where the fields do not overlap each other.

Since an interlace signal can form one field with half the number of scanning lines compared to a progressive signal, even if the signal amount is the same, an apparently double motion (frame rate) can be expressed. it can.
As an international standard system for efficiently encoding the interlaced signal, there are “MPEG-2” disclosed in the following Non-Patent Document 1, “AVC” disclosed in Non-Patent Document 2, and the like.

FIG. 6 is a block diagram showing a general moving image encoding apparatus that performs encoding processing in MPEG-2.
In MPEG-2, encoding processing is performed using a macroblock having a luminance signal of 16 pixels × 16 lines as a unit.
When the moving image signal is an interlaced signal, a signal in a field not shaded in FIG. 7 (hereinafter referred to as “top field”) 16 pixels × 8 lines, and a signal in a field shaded ( A macroblock is composed of a signal of 16 pixels × 16 lines combined with 16 pixels × 8 lines (hereinafter referred to as “bottom field”).

In MPEG-2, when DCT is used to encode an interlaced signal in units of 8 × 8 size blocks, the blocking unit should have a 1-bit flag “dct_type” of “0”. For example, as shown in FIG. 8, a frame block is configured by alternately arranging a top field and a bottom field of an interlace signal.
On the other hand, if “dct_type” is “1”, as shown in FIG. 9, the field block is configured by separately arranging the top field and the bottom field of the interlace signal.

In this way, the blocking unit constitutes either a frame block or a field block. Usually, when the moving image signal is stationary (the moving amount of the moving image signal is small and the amount of movement satisfies the reference value). If not, a frame block is formed. If the moving image signal is in a moving state (when the moving image signal has a large movement and the amount of movement is equal to or greater than a reference value), a field block is formed.
In other words, when the video signal is stationary, there is no temporal change in the signal between the top field and the bottom field, so it is better to use a signal with a close spatial distance without dividing it into fields. Is obtained. Therefore, a frame block is formed when the moving image signal is in a still state.
Conversely, when the moving image signal is in a state of motion, a temporal change occurs between the two fields. Therefore, a signal with a short time distance is used for each field rather than a close spatial distance. The higher encoding efficiency is obtained. Therefore, when the moving image signal is in a state of motion, a field block is configured.

In the DCT unit, when the blocking unit forms a frame block or a field block, DCT is performed on the frame block or field block to calculate a DCT coefficient.
The quantization unit quantizes the DCT coefficient calculated by the DCT unit using a prepared quantization matrix to calculate a quantization index.
The variable length encoding unit generates encoded data by variable length encoding the quantization index calculated by the quantization unit.

ISO / IEC 13818-2 “Information technology-Generic coding of moving pictures and associated audio information-Part 2: Video” ISO / IEC 14496-10 “Information technology-Coding of audio-visual objects-Part 10: Advanced video coding”

  Since the conventional moving image encoding apparatus is configured as described above, the quantization index is calculated by quantizing the DCT coefficient calculated by the DCT unit using a quantization matrix prepared in advance. However, since the same quantization matrix is always used regardless of whether the block type configured by the blocking unit is a frame block or a field block, it is not necessarily used for a frame block or a field block. There is a problem that a suitable quantization process cannot be performed and an efficient encoding process cannot be realized.

Here, it will be specifically described that an efficient encoding process cannot be realized.
Consider the case of encoding an interlaced signal in which four circles are arranged in a macroblock as shown in FIG.
When a frame block is composed of such interlaced signals, as shown in FIG. 10B, since one circle is included in one block, the horizontal signal state and the vertical signal state Are equal without any difference. For this reason, the DCT coefficient also has the same value in the horizontal direction and the vertical direction.
On the other hand, when configuring a field block, since the block is configured using signals every other line, the area occupied by one block on the screen is doubled in the vertical direction as compared with the frame block. Therefore, as shown in FIG. 10C, two circles are included in one block, so that the signal in the vertical direction is twice as dense as the signal in the horizontal direction. Therefore, the DCT coefficient is twice as large in the vertical direction as compared with the horizontal value.
As described above, although the distribution of DCT coefficients (particularly the distribution in the vertical direction) tends to be different between the frame block and the field block, only one type of quantization matrix is prepared in the conventional moving picture coding apparatus. Therefore, an efficient encoding process cannot be realized.

The present invention has been made to solve the above-described problems, and provides a moving picture coding apparatus capable of realizing efficient coding processing regardless of whether a frame block or a field block is configured. For the purpose.
The present invention also provides a moving picture decoding apparatus capable of efficiently decoding encoded data obtained by encoding an interlaced moving picture signal by the moving picture encoding apparatus and outputting a decoded picture. The purpose is to obtain.

The moving picture coding apparatus according to the present invention refers to an index signal indicating an aspect ratio of pixels constituting a moving picture signal when selecting a quantization matrix, and the type of block constituted by the block constituting means is a frame. If it is a block, the quantization matrix corresponding to the frame block is used to calculate the quantization index by quantizing the transform result of the discrete cosine transform unit, and the type of block configured by the block configuration unit is the field. If the block is a block, the quantization matrix corresponding to the field block is used to quantize the conversion result of the discrete cosine transform unit, thereby calculating a quantization index, and the pixels constituting the moving image signal. An index signal indicating the aspect ratio and a quantization index calculated by the quantization means The block type signal indicating the type of configuration block by fine said block means and variable length coding is obtained by a so that a variable-length coding means for outputting coded data.

According to the present invention, when the quantization matrix is selected, the index signal indicating the aspect ratio of the pixels constituting the moving image signal is referred to, and if the type of the block configured by the block configuration unit is a frame block, Using the quantization matrix corresponding to the frame block, the quantization index is calculated by quantizing the conversion result of the discrete cosine transform unit, and if the type of block configured by the block configuration unit is a field block, Quantization means for calculating a quantization index by quantizing the conversion result of the discrete cosine transform means using a quantization matrix corresponding to the field block, and an index indicating the aspect ratio of the pixels constituting the moving image signal Signal, quantization index calculated by the quantization means, and the block Since it is configured to so that a variable-length coding means for variable length coding and outputs the encoded data block type signal indicating the type of configuration block by forming means, constituting any frame block or field block Even if it does, there exists an effect which can implement | achieve an efficient encoding process.

It is a block diagram which shows the moving image encoder by Embodiment 1 of this invention. It is explanatory drawing which shows an example of the quantization matrix corresponding to a frame block, and the quantization matrix corresponding to a field block. It is a block diagram which shows the moving image encoder by Embodiment 2 of this invention. It is a block diagram which shows the moving image decoding apparatus by Embodiment 3 of this invention. It is a block diagram which shows the moving image decoding apparatus by Embodiment 4 of this invention. It is a block diagram which shows the general moving image encoding apparatus which implements an encoding process by MPEG-2. It is explanatory drawing which shows the macroblock of an interlace signal. It is explanatory drawing which shows a frame block. It is explanatory drawing which shows a field block. It is explanatory drawing which shows an example of the frame block and field block which are comprised from the same interlace signal.

Embodiment 1 FIG.
1 is a block diagram showing a moving picture coding apparatus according to Embodiment 1 of the present invention.
In FIG. 1, when a screen rearrangement unit 1 inputs an interlace signal which is an interlaced moving image signal (I picture, P picture, B picture), the screen rearrangement unit 1 rearranges the pictures in the encoding order and A process of dividing the signal into macroblock units is performed.
The subtracter 2 obtains a difference between the macroblock unit interlace signal output from the screen rearrangement unit 1 and the image signal of the predicted image generated by the motion compensation unit 14, and outputs a difference signal indicating the difference. To implement.

  The mode determination unit 3 inputs the macroblock unit interlaced signal output from the screen rearrangement unit 1 and the difference signal output from the subtracter 2 and adopts intraframe encoding processing as the encoding processing method. In this case, when the interlaced signal is output to the blocking unit 4 as the encoding target signal and the interframe encoding process is adopted as the encoding processing method, the difference signal is output as the encoding target signal to the blocking unit 4. Execute the process to output to.

If the encoding target signal output from the mode determination unit 3 is a signal with no motion (a signal with small motion), the blocking unit 4 alternately arranges the top field and the bottom field of the encoding target signal. A frame block (see FIG. 8) is configured, and if the signal has a motion (a signal with a large motion), the field block (see FIG. 9) is separated by arranging the top field and the bottom field separately. Implement the process to configure.
Further, the blocking unit 4 performs a process of outputting dct_type (block type signal) indicating whether the type of the configured block is a frame block or a field block.
The blocking unit 4 constitutes a block constituting unit.
In the first embodiment, when the motion of the signal to be encoded is small and the amount of motion is less than the reference value, the signal to be encoded is referred to as a “no-motion signal” and the motion of the signal to be encoded is large. When the amount of motion is greater than or equal to the reference value, the signal to be encoded is referred to as a “motion signal”.

The DCT unit 5 performs DCT (Discrete Cosine Transform) on the frame block or field block configured by the blocking unit 4 and outputs DCT coefficients. The DCT unit 5 constitutes a discrete cosine transform unit.
When the quantization unit 6 prepares a quantization matrix corresponding to the frame block and a quantization matrix corresponding to the field block in advance, and the dct_type output from the blocking unit 4 indicates that it is a “frame block”, When a quantization index is calculated by quantizing the DCT coefficient output from the DCT unit 5 using the quantization matrix corresponding to the frame block, and the dct_type indicates “field block”, the field block A process for calculating a quantization index is performed by quantizing the DCT coefficient output from the DCT unit 5 using a quantization matrix corresponding to. The quantization unit 6 constitutes a quantization means.

The inverse quantization unit 7 prepares in advance a quantization matrix corresponding to a frame block and a quantization matrix corresponding to a field block, and the dct_type output from the blocking unit 4 indicates that it is a “frame block” If the quantization index calculated by the quantization unit 6 is dequantized using the quantization matrix corresponding to the frame block and a DCT coefficient is output, and the dct_type indicates “field block”, Using the quantization matrix corresponding to the block, a process of dequantizing the quantization index calculated by the quantization unit 6 and outputting a DCT coefficient is performed.
The IDCT unit 8 performs IDCT (Inverse Discrete Cosine Transform) on the DCT coefficient output from the inverse quantization unit 7 and outputs a frame block or field block that is a local decoded signal.

The switch 9 outputs a zero value to the adder 10 as the image signal of the predicted image when the mode determination unit 3 adopts the intra-frame encoding process as the encoding process method, and the inter-frame encoding process as the encoding process method. Is employed, a process of outputting the image signal of the predicted image generated by the motion compensation unit 14 to the adder 10 is performed.
The adder 10 performs a process of adding the image signal of the prediction image output from the switch 9 and the local decoded signal output from the IDCT unit 8 to output the image signal of the reference image for prediction.

The switch 11 performs a process of switching the frame memory 12 that stores the image signal of the prediction reference image output from the adder 10 in accordance with the picture to be encoded.
The frame memory 12 is a memory for storing the image signal of the prediction reference image output from the switch 11.
The motion detection unit 13 performs a process of detecting a motion vector between the interlace signal in units of macroblocks output from the screen rearrangement unit 1 and the image signal of the reference image for prediction stored in the frame memory 12. .

The motion compensation unit 14 performs motion compensation of the image signal of the prediction reference image stored in the frame memory 12 using the motion vector detected by the motion detection unit 13, and generates an image signal of the prediction image. Perform the process.
The variable length encoding unit 15 performs variable length encoding on the quantization index calculated by the quantization unit 6, the dct_type output from the blocking unit 4, the motion vector detected by the motion detection unit 13, and the like, thereby encoding the encoded data. Perform the output process. The variable length encoding unit 15 constitutes variable length encoding means.

Next, the operation will be described.
When the screen rearrangement unit 1 receives an interlace signal that is an interlaced moving image signal, the screen rearrangement unit 1 rearranges the pictures in the encoding order and divides the interlace signal into units of macroblocks.
When the subtractor 2 receives the interlace signal in units of macroblocks from the screen rearrangement unit 1, it obtains the difference between the interlace signal and the image signal of the predicted image generated by the motion compensation unit 14, and indicates the difference. The difference signal is output to the mode determination unit 3.

The mode determination unit 3 inputs the macroblock-unit interlace signal output from the screen rearrangement unit 1 and the difference signal output from the subtracter 2 and adopts intraframe encoding processing as the encoding processing method. In this case, the interlaced signal is output to the blocking unit 4 as the encoding target signal, and when the interframe encoding processing is adopted as the encoding processing method, the difference signal is used as the encoding target signal. 4 is output.
Note that when determining the encoding processing method to be employed, the mode determination unit 3 performs, for example, encoding efficiency when encoding is performed using intra-frame encoding processing, and encoding when encoding is performed using inter-frame encoding processing. Compare the efficiency and select the encoding processing method with the higher encoding efficiency.

When receiving the encoding target signal from the mode determination unit 3, the blocking unit 4 determines whether or not the encoding target signal is moving (the magnitude of the motion). Since a known technique may be used as a method for determining the presence or absence of movement, detailed description thereof is omitted here.
The blocking unit 4 determines the presence / absence of motion of the encoding target signal. If the signal is not moving, the blocking unit 4 alternately arranges the top field and the bottom field of the encoding target signal as shown in FIG. Thus, a frame block is configured, and the frame block is output to the DCT unit 5.
On the other hand, if the signal has motion, as shown in FIG. 9, the top field and the bottom field are separately arranged to form a field block, and the field block is output to the DCT unit 5.
When configuring the frame block or the field block, the blocking unit 4 outputs dct_type indicating the type of the configured block to the quantization unit 6, the inverse quantization unit 7 and the variable length coding unit 15.

  When the DCT unit 5 receives the frame block or field block from the blocking unit 4, the DCT unit 5 performs DCT on the frame block or field block and outputs the DCT coefficient to the quantization unit 6.

The quantization unit 6 prepares in advance a quantization matrix corresponding to a frame block and a quantization matrix corresponding to a field block.
Here, FIG. 2 is an explanatory diagram showing an example of a quantization matrix corresponding to a frame block and a quantization matrix corresponding to a field block.
As shown in FIG. 10B, the signal of the frame block has the same signal state in the horizontal direction and the vertical direction, so that the quantization process can be performed equally regardless of the direction. As shown in FIG. 2 (A), the values of the conversion matrix are prepared such that the horizontal direction and the vertical direction have the same tendency.
On the other hand, as shown in FIG. 10C, the signal in the field block has twice the density of the signal in the vertical direction as compared with the horizontal direction. The matrix value is set to be double in the horizontal direction with respect to the vertical direction.
As a result, the quantization error of the DCT coefficient in the vertical direction can be reduced by a factor of two compared to the quantization error of the DCT coefficient in the horizontal direction, so that the state of the signal in the vertical direction can be encoded with higher performance. It becomes possible.

When the dct_type output from the blocking unit 4 indicates “frame block”, the quantizing unit 6 outputs from the DCT unit 5 using the quantization matrix of FIG. 2A corresponding to the frame block. The quantization index is calculated by quantizing the DCT coefficient.
On the other hand, when the dct_type indicates “field block”, the quantization matrix of FIG. 2B corresponding to the field block is used to quantize the DCT coefficient output from the DCT unit 5 to quantize. Calculate the index.

Similar to the quantization unit 6, the inverse quantization unit 7 prepares two types of quantization matrices (quantization matrices in FIGS. 2A and 2B), and the dct_type output from the blocking unit 4. 2 indicates that it is a “frame block”, the quantization index calculated by the quantization unit 6 is inversely quantized using the quantization matrix of FIG. 2A corresponding to the frame block, and the DCT coefficient is IDCT. Output to unit 8.
On the other hand, when the dct_type indicates “field block”, the quantization index calculated by the quantization unit 6 is inversely quantized using the quantization matrix of FIG. 2B corresponding to the field block. The DCT coefficient is output to the IDCT unit 8.
When receiving the DCT coefficient from the inverse quantization unit 7, the IDCT unit 8 performs IDCT on the DCT coefficient and outputs a frame block or a field block that is a local decoded signal to the adder 10.

The switch 9 outputs a zero value as an image signal of the predicted image to the adder 10 when the mode determination unit 3 adopts the intra-frame encoding process as the encoding process method, and the inter-frame encoding as the encoding process method. When the process is employed, the image signal of the predicted image output from the motion compensation unit 14 is output to the adder 10.
The adder 10 adds the image signal of the predicted image output from the switch 9 and the local decoded signal output from the IDCT unit 8, and outputs the image signal of the reference image for prediction to the switch 11.
The switch 11 stores the image signal of the prediction reference image output from the adder 10 in the frame memory 12.

When the motion detection unit 13 receives an interlace signal in units of macroblocks from the screen rearrangement unit 1, the motion detection unit 13 calculates a motion vector between the interlace signal and the image signal of the prediction reference image stored in the frame memory 12. To detect.
When the motion detection unit 13 detects a motion vector, the motion compensation unit 14 performs motion compensation of the image signal of the prediction reference image stored in the frame memory 12 using the motion vector, and An image signal is generated.
The variable length encoding unit 15 includes a quantization index calculated by the quantization unit 6, a dct_type output from the blocking unit 4, a motion vector detected by the motion detection unit 13, and encoding scheme information (encoding process). Information indicating whether intra-frame coding processing or inter-frame coding processing is adopted as a method) is variable-length coded, and encoded data is output.

As is apparent from the above, according to the first embodiment, when the dct_type output from the blocking unit 4 indicates “frame block”, the quantization of FIG. 2A corresponding to the frame block is performed. When the quantization index is calculated by quantizing the DCT coefficient output from the DCT unit 5 using the matrix and the dct_type output from the blocking unit 4 indicates “field block”, the field block 2B, the quantization unit 6 that calculates the quantization index by quantizing the DCT coefficient output from the DCT unit 5 is provided. Regardless of whether a block or a field block is configured, an effective encoding process can be realized. .
In the first embodiment, the block that the block forming unit 4 switches with / without motion is shown. However, both blocks are configured regardless of whether there is motion or not. And then select the block with the smaller code amount, the block with the smaller coding noise generated by the coding process, or the block with the better balance between the code amount and the coding noise. You may make it do.
With such a configuration, an effect of improving the encoding efficiency can be obtained.

Embodiment 2. FIG.
FIG. 3 is a block diagram showing a moving picture coding apparatus according to Embodiment 2 of the present invention. In the figure, the same reference numerals as those in FIG.
The mode determination unit 21 performs the process of outputting the encoding target signal to the blocking unit 4 in the same manner as the mode determination unit 3 of FIG. 1 and also uses an encoding processing method as an index signal used to select a quantization matrix. A process of outputting a flag “mtx_flag” indicating whether or not is an intra-frame encoding process or an inter-frame encoding process to the quantization unit 22 is performed.

  The quantization unit 22 prepares in advance a plurality of quantization matrices corresponding to frame blocks and a plurality of quantization matrices corresponding to field blocks, respectively, and indicates that the dct_type output from the blocking unit 4 is a “frame block”. In the case of showing, referring to the flag “mtx_flag” output from the mode determination unit 21, a quantization matrix used for quantization is selected from a plurality of quantization matrices corresponding to the frame block, and the quantization matrix is used. Then, the quantization index is calculated by quantizing the DCT coefficient output from the DCT unit 5. On the other hand, when the dct_type indicates that it is a “field block”, the flag “mtx_flag” output from the mode determination unit 21 is used for quantization among a plurality of quantization matrices corresponding to the field block. A process of calculating a quantization index by selecting a quantization matrix and quantizing the DCT coefficient output from the DCT unit 5 using the quantization matrix is performed. The quantization unit 22 constitutes a quantization means.

  The inverse quantization unit 23 prepares in advance a plurality of quantization matrices corresponding to the frame blocks and a plurality of quantization matrices corresponding to the field blocks, respectively, and the fact that the dct_type output from the blocking unit 4 is “frame block”. , Referring to the flag “mtx_flag” output from the mode determination unit 21, a quantization matrix used for quantization is selected from a plurality of quantization matrices corresponding to the frame block, and the quantization matrix is selected. The quantization index calculated by the quantization unit 22 is inversely quantized and a DCT coefficient is output. On the other hand, when the dct_type indicates that it is a “field block”, the flag “mtx_flag” output from the mode determination unit 21 is used for quantization among a plurality of quantization matrices corresponding to the field block. A process of selecting a quantization matrix, dequantizing the quantization index calculated by the quantization unit 22 using the quantization matrix, and outputting a DCT coefficient is performed.

  The variable length coding unit 24 detects the quantization index calculated by the quantization unit 22, the dct_type output from the blocking unit 4, the flag “mtx_flag” output from the mode determination unit 21, and the motion detection unit 13. A process of variable-length coding a motion vector or the like and outputting encoded data is performed. The variable length coding unit 24 constitutes variable length coding means.

Next, the operation will be described.
When the mode determination unit 21 receives the macroblock-unit interlace signal output from the screen rearrangement unit 1 and the difference signal output from the subtractor 2, the mode determination unit 21 encodes the code as in the mode determination unit 3 of FIG. The interlaced signal or the difference signal is output to the blocking unit 4 as the signal to be converted.
Further, the mode determination unit 21 quantizes the flag “mtx_flag” indicating whether the encoding processing method is intra-frame encoding processing or inter-frame encoding processing as an index signal used for selecting a quantization matrix. Output to the unit 22, the inverse quantization unit 23, and the variable length coding unit 24.

The quantization unit 22 prepares in advance two or more types of quantization matrices corresponding to frame blocks, and prepares two types of quantization matrices corresponding to field blocks.
When the quantization unit 22 indicates that the dct_type output from the blocking unit 4 is a “frame block”, the quantization unit 22 refers to the flag “mtx_flag” output from the mode determination unit 21 and corresponds to the plurality of frame blocks. A quantization matrix used for quantization is selected from the quantization matrices.
For example, when the flag “mtx_flag” indicates that the intra-frame encoding process is to be adopted, a quantization matrix suitable for the intra-frame encoding process is selected from the plurality of quantization matrices corresponding to the frame block. (E.g., a quantization matrix capable of finely quantizing AC coefficients), if the flag "mtx_flag" indicates that interframe coding processing is adopted, quantization suitable for interframe coding processing A matrix is selected (for example, a quantization matrix capable of coarsely quantizing AC coefficients).

On the other hand, when the dct_type output from the blocking unit 4 indicates “field block”, a plurality of quantization matrices corresponding to the field block are referred to with reference to the flag “mtx_flag” output from the mode determination unit 21. A quantization matrix to be used for quantization is selected from the list.
For example, when the flag “mtx_flag” indicates that the intra-frame encoding process is to be adopted, a quantization matrix suitable for the intra-frame encoding process is selected from a plurality of quantization matrices corresponding to the field block. (E.g., a quantization matrix capable of finely quantizing AC coefficients), if the flag "mtx_flag" indicates that interframe coding processing is adopted, quantization suitable for interframe coding processing A matrix is selected (for example, a quantization matrix capable of coarsely quantizing AC coefficients).
When the quantization unit 22 selects a quantization matrix used for quantization, the quantization unit 22 calculates a quantization index by quantizing the DCT coefficient output from the DCT unit 5 using the quantization matrix.

Similar to the quantization unit 22, the inverse quantization unit 23 prepares two or more types of quantization matrices corresponding to frame blocks, and prepares two types of quantization matrices corresponding to field blocks.
The inverse quantization unit 23 selects a quantization matrix used for inverse quantization in the same manner as the quantization unit 22, and uses the quantization matrix to reverse the quantization index calculated by the quantization unit 22. The quantization is performed and the DCT coefficient is output to the IDCT unit 8.

The variable length encoding unit 24 is detected by the quantization index calculated by the quantization unit 22, dct_type output from the blocking unit 4, the flag “mtx_flag” output from the mode determination unit 21, and the motion detection unit 13. The motion vector, the encoding method information, and the like are subjected to variable length encoding to output encoded data.
However, since the variable length encoding unit 24 performs variable length encoding on the encoding scheme information, the flag “mtx_flag” output from the mode determination unit 21 does not necessarily need to be variable length encoded.

  As apparent from the above, according to the second embodiment, when the quantizing unit 22 indicates that the dct_type output from the blocking unit 4 is a “frame block”, it is output from the mode determining unit 21. If the quantization matrix used for quantization is selected from a plurality of quantization matrices corresponding to the frame block with reference to the flag “mtx_flag”, and the dct_type indicates “field block”, the mode is determined. With reference to the flag “mtx_flag” output from the unit 21, the quantization matrix used for quantization is selected from the plurality of quantization matrices corresponding to the field block. In addition, the quantization matrix used for quantization is optimized and more efficient encoding processing is implemented. An effect that can be.

In the second embodiment, the mode determination unit 21 indicates whether the encoding processing method is an intra-frame encoding process or an inter-frame encoding process as an index signal used for selecting a quantization matrix. Although the flag “mtx_flag” is output to the quantization unit 22, the inverse quantization unit 23, and the variable length coding unit 24, as an index signal used for selecting a quantization matrix, the pixel of the interlace signal is displayed. The aspect ratio (ratio between the vertical direction and the horizontal direction in which one pixel occupies the screen) may be given to the quantization unit 22, the inverse quantization unit 23, and the variable length coding unit 24.
In this case, the quantization unit 22 and the inverse quantization unit 23 increase the signal density in the vertical direction as shown in FIG. 10C when the aspect ratio of the pixel is vertically long. A quantization matrix having a tendency like the quantization matrix for the field block is selected. On the other hand, when the aspect ratio of the pixel is the same in the horizontal direction and the vertical direction, the signal state as shown in FIG. 10 (B) is obtained, so that the tendency like the quantization matrix for the frame block in FIG. Select the quantization matrix to be shown.

Further, as an index signal used for selecting a quantization matrix, the distribution of DCT coefficients output from the DCT unit 5 may be provided to the quantization unit 22, the inverse quantization unit 23, and the variable length coding unit 24. .
In this case, the quantizing unit 22 and the inverse quantizing unit 23, for example, if there are a large number of DCT coefficients having a large value in the vertical direction, the quantum indicating a tendency like the quantization matrix for the field block in FIG. Select the optimization matrix. On the other hand, if there is no particular bias between the DCT coefficients in the vertical direction and the DCT coefficients in the horizontal direction, a quantization matrix showing a tendency like the quantization matrix for the frame block in FIG. 2A is selected.

Embodiment 3 FIG.
4 is a block diagram showing a moving picture decoding apparatus according to Embodiment 3 of the present invention.
In FIG. 4, a variable length decoding unit 31 performs variable length decoding on the encoded data output from the moving picture encoding apparatus of FIG. 1, outputs a quantization index to the inverse quantization unit 32, and outputs dct_type (block type signal). ) Is output to the inverse quantization unit 32 and the macroblocking unit 34, and the process of outputting the motion vector to the motion compensation unit 38 is performed. The variable length decoding unit 31 constitutes a variable length decoding unit.

  The inverse quantization unit 32 prepares a quantization matrix corresponding to the frame block and a quantization matrix corresponding to the field block in advance, and indicates that the dct_type output from the variable length decoding unit 31 is a “frame block”. In this case, the quantization index output from the variable length decoding unit 31 is inversely quantized using a quantization matrix corresponding to the frame block, and a DCT coefficient is output, indicating that the dct_type is “field block”. Using the quantization matrix corresponding to the field block, a process of dequantizing the quantization index output from the variable length decoding unit 31 and outputting a DCT coefficient is performed. Note that the inverse quantization unit 32 constitutes an inverse quantization means.

The IDCT unit 33 performs IDCT on the DCT coefficient output from the inverse quantization unit 32 and outputs a frame block or a field block. The IDCT unit 33 constitutes an inverse discrete cosine transform unit.
When the dct_type output from the variable length decoding unit 31 indicates “frame block”, the macroblock forming unit 34 rearranges the fields in the frame block output from the IDCT unit 33 to form the original macroblock. If the dct_type output from the variable length decoding unit 31 indicates “field block”, the process of rearranging the fields in the field block output from the IDCT unit 33 to configure the original macroblock is performed. To do. The macroblock forming unit 34 constitutes a macroblock forming unit.

The adder 35 corresponds to a macroblock signal output from the macroblocking unit 34 (an encoding target signal (interlaced signal or difference signal in units of macroblocks) output from the mode determination unit 3 in FIG. 1). Signal) and the image signal of the predicted image output from the motion compensation unit 38 are added to generate a decoded image signal.
The switch 36 performs a process of switching the frame memory 37 that stores the image signal of the decoded image output from the adder 35 in accordance with the picture of the decoded image.
The frame memory 37 is a memory for storing the image signal of the decoded image.
The motion compensation unit 38 performs motion compensation of the image signal of the decoded image stored in the frame memory 37 using the motion vector output from the variable length decoding unit 31, and generates an image signal of the predicted image To implement.

Next, the operation will be described.
When the variable length decoding unit 31 receives the encoded data output from the moving image encoding apparatus in FIG. 1, the variable length decoding unit 31 performs variable length decoding on the encoded data and outputs the quantization index to the inverse quantization unit 32. dct_type is output to the inverse quantization unit 32 and the macroblocking unit 34. The motion vector is output to the motion compensation unit 38.

The inverse quantization unit 32 prepares the same two types of quantization matrices as the quantization unit 6 of FIG. 1 (quantization matrices of FIGS. 2A and 2B), and is output from the variable length decoding unit 31. When the dct_type indicates “frame block”, the quantization index output from the variable-length decoding unit 31 is inversely quantized using the quantization matrix of FIG. The coefficient is output to the IDCT unit 33.
On the other hand, when the dct_type indicates that it is a “field block”, the quantization index output from the variable length decoding unit 31 is dequantized using the quantization matrix of FIG. 2B corresponding to the field block. The DCT coefficient is output to the IDCT unit 33.
Upon receiving the DCT coefficient from the inverse quantization unit 32, the IDCT unit 33 performs IDCT on the DCT coefficient and outputs a frame block or a field block to the macroblock conversion unit 34.

When the dct_type output from the variable length decoding unit 31 indicates “frame block”, the macroblocking unit 34 rearranges the fields in the frame block output from the IDCT unit 33 (see FIG. 1). The original macroblock is configured and the macroblock is output to the adder 35.
On the other hand, when the dct_type indicates that it is a “field block”, the original macro is obtained by rearranging the fields in the field block output from the IDCT unit 33 (the reverse process of the blocking unit 4 in FIG. 1). A block is configured, and the macroblock is output to the adder 35.
Here, the original arrangement of macroblocks is not mentioned because it does not affect the present invention. For example, the state of the original macroblock may be a frame state (a state shown in FIG. 7) in which two fields are combined, a state where each field is separated, or other arrangements. There may be.

When the adder 35 receives the macroblock signal from the macroblocking unit 34, the adder 35 adds the macroblock signal and the image signal of the predicted image output from the motion compensation unit 38 to generate an image signal of the decoded image. To do.
The switch 36 stores the image signal of the decoded image generated by the adder 35 in the frame memory 37.
The motion compensation unit 38 performs motion compensation of the image signal of the decoded image stored in the frame memory 37 using the motion vector output from the variable length decoding unit 31, and generates an image signal of the predicted image. The image signal of the predicted image is output to the adder 35.

  As apparent from the above, according to the third embodiment, when the inverse quantization unit 32 indicates that the dct_type output from the variable length decoding unit 31 is a “frame block”, it corresponds to the frame block. When the quantization index output from the variable length decoding unit 31 is inversely quantized using the quantization matrix and a DCT coefficient is output, and the dct_type indicates “field block”, the quantum corresponding to the field block Since the quantization index output from the variable-length decoding unit 31 is inversely quantized using the quantization matrix and the DCT coefficient is output, the interlace signal is encoded by the moving image encoding apparatus of FIG. It is possible to efficiently decode the encoded data and output a decoded image.

Embodiment 4 FIG.
5 is a block diagram showing a moving picture decoding apparatus according to Embodiment 4 of the present invention. In the figure, the same reference numerals as those in FIG.
The variable length decoding unit 41 performs variable length decoding on the encoded data output from the moving image encoding apparatus in FIG. 3, outputs the quantization index and the flag “mtx_flag” to the inverse quantization unit 42, and sets dct_type (block type). Signal) is output to the inverse quantization unit 42 and the macroblocking unit 34, and the motion vector is output to the motion compensation unit 38. The variable length decoding unit 41 constitutes variable length decoding means.

  The inverse quantization unit 42 prepares in advance a plurality of quantization matrices corresponding to frame blocks and a quantization matrix corresponding to field blocks, and the dct_type output from the variable length decoding unit 41 is a “frame block”. When indicating that, the quantization matrix used for the inverse quantization is selected from the plurality of quantization matrices corresponding to the frame block with reference to the flag “mtx_flag” output from the variable length decoding unit 41, and the quantum Using the quantization matrix, the quantization index output from the variable length decoding unit 41 is inversely quantized to output a DCT coefficient. On the other hand, when the dct_type indicates “field block”, the dequantization is performed from among a plurality of quantization matrices corresponding to the field block with reference to the flag “mtx_flag” output from the variable length decoding unit 41. A quantization matrix to be used is selected, and a quantization index output from the variable length decoding unit 41 is inversely quantized using the quantization matrix to output a DCT coefficient. The inverse quantization unit 42 constitutes inverse quantization means.

Next, the operation will be described.
Except for the variable length decoding unit 41 and the inverse quantization unit 42, the present embodiment is the same as the third embodiment.
When the variable length decoding unit 41 receives the encoded data output from the moving image encoding apparatus in FIG. 3, the variable length decoding unit 41 performs variable length decoding on the encoded data, and dequantizes the quantization index and the flag “mtx_flag”. 42, and dct_type is output to the inverse quantization unit 42 and the macroblocking unit 34. The motion vector is output to the motion compensation unit 38.

Similar to the quantization unit 22 of FIG. 3, the inverse quantization unit 42 prepares a plurality of quantization matrices corresponding to frame blocks and a plurality of quantization matrices corresponding to field blocks.
When the inverse quantization unit 42 indicates that the dct_type output from the variable length decoding unit 41 is a “frame block”, the inverse quantization unit 42 refers to the flag “mtx_flag” output from the variable length decoding unit 41 to convert the frame block into a frame block. A quantization matrix to be used for inverse quantization is selected from a plurality of corresponding quantization matrices, and the quantization index output from the variable length decoding unit 41 is inversely quantized using the quantization matrix to obtain a DCT coefficient. Output.
On the other hand, when the dct_type indicates “field block”, the dequantization is performed from among a plurality of quantization matrices corresponding to the field block with reference to the flag “mtx_flag” output from the variable length decoding unit 41. A quantization matrix to be used is selected, and the quantization index output from the variable length decoding unit 41 is inversely quantized using the quantization matrix to output a DCT coefficient.

  As is apparent from the above, according to the fourth embodiment, when the inverse quantization unit 42 indicates that the dct_type output from the variable length decoding unit 41 is a “frame block”, the variable length decoding unit 41 Referring to the flag “mtx_flag” output from, a quantization matrix used for inverse quantization is selected from a plurality of quantization matrices corresponding to the frame block, and indicates that the dct_type is a “field block” In this case, with reference to the flag “mtx_flag” output from the variable length decoding unit 41, the quantization matrix used for inverse quantization is selected from a plurality of quantization matrices corresponding to the field block. In addition to the third embodiment, the quantization matrix used for inverse quantization is optimized, and more efficient decoding is achieved. An effect that can be achieved sense.

In the first to fourth embodiments, the description has been made assuming that the moving image encoding device or the moving image decoding device has two or more quantization matrices in advance. It is also possible to switch the quantization matrix each time, for example, by generating a quantization matrix and including the quantization matrix information in a part of the encoded data and outputting it to the video decoding device.
With this configuration, an optimal quantization matrix is designed according to the state of the image signal input to the video encoding device and the parameters such as the bit rate set by the video encoding device. Using this, it becomes possible to perform the encoding process.

  DESCRIPTION OF SYMBOLS 1 Screen rearrangement part, 2 Subtractor, 3 Mode determination part, 4 Blocking part (block formation means), 5 DCT part (discrete cosine transformation means), 6 Quantization part (quantization means), 7 Inverse quantization part , 8 IDCT section, 9 switch, 10 adder, 11 switch, 12 frame memory, 13 motion detection section, 14 motion compensation section, 15 variable length coding section (variable length coding means), 21 mode determination section, 22 quantum Quantization unit (quantization means), 23 inverse quantization unit, 24 variable length coding unit (variable length coding unit), 31 variable length decoding unit (variable length decoding unit), 32 inverse quantization unit (inverse quantization unit) ), 33 IDCT unit (inverse discrete cosine transform unit), 34 Macroblock unit (macroblock configuration unit), 35 Adder, 36 switch, 37 frame memory, 38 motion compensation unit, 41 Variable-length decoding section (variable length decoding means), 42 inverse quantization unit (inverse quantization means).

Claims (2)

A block configuration means for configuring a frame block in which top and bottom fields of an interlaced video signal in a macro block are alternately arranged or a field block in which the top field and bottom field of the video signal are separately arranged; ,
Discrete cosine transform means for performing discrete cosine transform on the frame block or field block configured by the block configuration means;
When the quantization matrix is selected, the index signal indicating the aspect ratio of the pixels constituting the moving image signal is referred to, and if the type of block configured by the block configuration means is a frame block, it corresponds to the frame block. The quantization index is calculated by quantizing the conversion result of the discrete cosine transform unit using the quantization matrix to be processed, and if the type of the block configured by the block configuration unit is a field block, the field block A quantization means for calculating a quantization index by quantizing the conversion result of the discrete cosine transform means using a quantization matrix corresponding to
The index signal indicating the aspect ratio of the pixels constituting the moving image signal, the quantization index calculated by the quantizing unit, and the block type signal indicating the type of block configured by the block forming unit are variable-length encoded. And a variable-length encoding means for outputting encoded data. Variable-length decoding means for variable-length decoding encoded data and outputting an index signal indicating a pixel aspect ratio, a quantization index, and a block type signal;
When an index signal indicating the aspect ratio of the pixel is selected when selecting a quantization matrix, and the block type indicated by the block type signal output from the variable-length decoding unit is a frame block, When the quantization index output from the variable length decoding unit is inversely quantized using a corresponding quantization matrix, and the block type indicated by the block type signal output from the variable length decoding unit is a field block, Dequantizing means for dequantizing the quantization index output from the variable length decoding means, using a quantization matrix corresponding to the field block;
An inverse discrete cosine transform unit that outputs a frame block or a field block by performing an inverse discrete cosine transform on the inverse quantization result of the inverse quantization unit;
A moving picture decoding apparatus comprising: a macroblock configuration unit configured to rearrange fields in a frame block or a field block output from the inverse discrete cosine transform unit to configure an original macroblock.

Images