JP2002125232A - Video data encoding apparatus and method - Google Patents

Abstract

(57) [Summary] Even if encoded moving image data is decoded, converted, and then re-encoded,
Provided is an encoding device that encodes moving image data in such a manner that image quality does not deteriorate due to re-encoding. SOLUTION: Moving picture data 401 of 60 frames / sec.
Out of the B pictures, the picture determination unit 113 determines a B picture B1 to be converted into an I picture at the time of re-encoding,
For the B picture B1, the control unit 11 obtains a larger bit allocation value using a dedicated calculation formula. Then, the variable-length encoding processing unit 15 adjusts the quantization step of the B picture according to the bit allocation value so that the B picture is encoded at a low compression rate.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an encoding apparatus and method for encoding moving picture data to be decoded and re-encoded later using motion compensation prediction.

[0002]

2. Description of the Related Art Conventionally, as a television broadcasting system in Japan, a scanning system is an interlace system, and a field frequency is 60 fields / second (when converted to a frame frequency, 30 fields / second).
(Frame / second).
On the other hand, in the future television broadcasting, image data produced in a progressive scanning system with a high frame rate of 60 frames / second is encoded by MPEG2 and once recorded on a medium, and the encoded data recorded on the medium is encoded. It is expected that digital broadcasting that broadcasts will become mainstream.
When a program is produced digitally, a scanning method and a frame frequency may differ between production and broadcasting.

[0003] For example, even if a program is produced by a progressive scanning method of 60 frames / second with high image quality, the program can be used as it is by MPEG.
If there is a problem such as restrictions on the broadcast band, such as that the available frequencies are not limited to encoding and broadcasting, if the encoded data once recorded on the medium is decoded, the interlaced 60 fields are decoded. / S image data and then re-encoded by MPEG2 before broadcasting. Since the re-encoding is performed with a frame structure of 30 frames / sec, the moving picture data which was 60 frames / sec in the process of “decoding → conversion → re-encoding” is increased to 30 frames / sec, the number of frames per time. Is halved.

A case will be described below in which data obtained by encoding a progressively scanned image by MPEG2 is converted into an interlaced scanning image and re-encoded by a frame structure. FIG. 4 is a diagram showing an approximate flow of the conversion process. First, the progressive scan image 401 is encoded by MPEG2 to obtain encoded data 402. Next, an image obtained by sequentially scanning the encoded data 402 is decoded to generate a decoded sequentially scanned image 403. Here, conversion of the scanning method is performed, and 6
The decoded progressively scanned image 403 of 0 frames / sec is converted into an interlaced scanned image 404 of 60 fields / sec, and then re-encoded in a frame format by MPEG2. Since re-encoding is performed in a frame structure, the 60-field / second interlaced scan image 404 is converted into 30-frame / second frame structure data 405 for re-encoding. Then, re-encoded data 406 obtained by encoding the frame structure data 405 is broadcast as broadcast data.

A characteristic feature of such conversion processing is that, as shown in FIG.
Decode the encoded one, convert it to interlaced,
When re-encoding an interlaced image, two frames in the decoded progressively scanned image 403 become one frame of the frame structure data 405 to be re-encoded (two fields in the interlaced scanned image 404). Then, the decoded progressively scanned image 403 which is the base of the frame of the I picture attribute in the frame structure data 405
, At least one of the two frames is a B picture.

[0006]

However, the above-mentioned processing, that is, the progressive scanning video data of 60 frames / second is once compressed by the MPEG2 system, decoded, and then converted to the interlaced system. Performing the process of re-encoding with a frame structure of 30 frames / sec causes a problem that the image quality of the interlaced moving image after re-encoding deteriorates. Hereinafter, a process in which the deterioration occurs will be described.

[0007] Generally, in the motion compensation prediction of the MPEG2 coding, all the frames constituting the moving image data are:
It is divided into three types of pictures, I, P and B. The I picture is directly or indirectly referred to in encoding / decoding of all pictures of other types (P, B). The image quality of the entire moving image data is degraded. Therefore, at the time of encoding, the compression ratio of the I picture is reduced (the allocation of the encoded data amount, that is, “bit allocation” is increased) so that the image quality of the I picture does not deteriorate. Conversely, this means that bit allocation to B and P pictures is relatively small. In particular, since a B picture is not referred to in encoding / decoding other pictures (since other pictures are not affected even if the picture quality of the B picture is reduced), the bit allocation per picture is as follows.
The smallest among the three types of pictures. like this,
The control algorithm for the MPEG2 bit allocation is "Test Model 5," ISO / IEC JTC / SC29 / WG11 / N040
0, 1993).

Therefore, when such a rule of bit allocation is applied to the case of “transformation, re-encoding with a frame structure” shown in FIG. 4, first, the I picture in the decoding order scan image 402 at the time of decoding The image quality of I1 'is almost inferior to the I picture I1 in the progressive scan image 401 before encoding, whereas the B picture B1' in the decoding order scan image 403 is the same as that of the progressive scan image 401 before encoding. It is inevitable that the image quality deteriorates compared to the B picture B1. However, the decoding order scanned image 403 is converted into an interlaced scanned image 404.
In the process of generating frame structure data 405 from
The I picture Ii in the frame structure data 405 is
It is generated from the I picture I1 'and the B picture B1' of the decoding order scanned image 403. In other words, the image quality of the Ii picture in the frame structure data 405 to be re-encoded is reduced due to the image quality degradation that has occurred in the B picture B1 ′. Further, the frame structure data 4
When re-encoding 05, the Ii picture becomes a reference picture in motion compensation prediction, so that the P and B pictures that use this as the reference picture carry over the influence of the image quality degradation that has occurred in the Ii picture. . Frame structure data 405
This means that the image quality of the entire (interlaced scan image 404) after re-encoding (at the time of re-decoding) is reduced.

The present invention has been made in view of the above problems, and is an encoding apparatus for compressing and encoding moving image data, in which the encoded moving image data is subjected to a process of “decoding → conversion → re-encoding”. It is an object of the present invention to provide an encoding device capable of performing re-encoding in such a manner that image quality hardly deteriorates.

[0010]

In order to achieve the above-mentioned object, an encoding apparatus according to the present invention performs compression encoding on video data to be decoded and re-encoded later using motion compensation prediction. An encoding device for performing processing, comprising: identification means for identifying a frame to be referred to as an I-picture attribute frame in motion compensation prediction during re-encoding processing, among frames constituting the moving image data The quantization step determining means for reducing the quantization step in the quantization process for the frame identified by the identification means, compared to other frames having the same picture attribute, and the quantization step determining means determines Encoding means for encoding the frame identified by the identification means in accordance with the quantization step. For this reason, when encoding a progressively scanned moving image using motion compensation prediction, a frame of a progressively scanned image that will be converted to an I-picture when conversion to an interlaced system is performed in the future is performed. The encoding is performed so that the image quality at the time of performing is higher. Therefore, even for the interlaced moving image after the conversion, the image quality hardly deteriorates due to the encoding.

In order to achieve the above object, a coding method according to the present invention provides a coding method for performing a compression coding process using motion compensation prediction on moving picture data to be decoded and recoded later. An identification method for identifying a frame to be referred to as an I-picture attribute frame in motion-compensated prediction during re-encoding processing, among the frames constituting the video data, A quantization step determining step of reducing the quantization step in the quantization process for the frame identified in the step as compared with other frames of the same picture attribute; and a quantization step determined in the quantization step determining step. An encoding step of encoding the frame identified in the identification step according to the step. The features.

[0012]

BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a block diagram showing an embodiment of an encoding apparatus according to the present invention. (Structure) FIG. 1 shows the structure of an encoding device 1 according to the present embodiment in which the moving image data encoded by the encoding device 1 is decoded by “decoding → conversion to interlacing → recoding in a frame structure”. 2 is a block diagram shown together with a conversion / re-encoding device 2 that performs a process of “encoding” and outputs the result.

(Outline) First, an outline of the processing for each of these devices will be described. First, the encoding device 1
Is a device for compressing and encoding moving image data (60 frames / second) for a broadcast program produced by a high-quality progressive scanning method for broadcasting. The encoding device 1 performs compression encoding according to the MPEG2 method. However, the broadcast method after compression of the broadcast program data (forward scan method, broadcast at 60 frames / second, or after converting to the interlace method) The compression / encoding processing method is switched according to (broadcasting at 30 frames / sec). The compression-encoded data is stored in a medium, and is used for broadcasting as it is when the broadcasting method is the progressive scanning method, but when the broadcasting method is the interlacing method,
The compression-encoded data is converted into interlaced data by the conversion / re-encoding device 2 and re-encoded in a frame structure (30 frames / sec) before being broadcast.

The conversion / re-encoding device 2 decodes the compression-coded progressive scanning data into an interlaced data (60 fields / second), and re-compresses the compressed data for broadcasting. Become Compression encoding is MPEG2
The data is re-encoded in a frame structure. The conversion / re-encoding device 2 includes a decoding unit 2 that decodes progressively scanned image data compressed and encoded by the encoding device 1.
1. It comprises a conversion unit 22 for converting decoded progressively scanned video data into an interlace format, and a re-encoding unit 23 for compression-coding the converted interlaced video data in a frame structure.

The process of encoding, converting and re-encoding moving picture data is the same as in the prior art (see FIG. 4). The processing contents of both apparatuses are summarized with reference to FIG. 4. First, the encoding processing from the progressive scan image 401 to the encoded data 402 is performed by the encoding apparatus 1, and the decoding sequential scanning from the encoded data 402 is performed. The decoding process for the image 403 is performed by the decoding unit 21, and from the decoded progressively scanned image 403 to the interlaced scanned image 404.
The conversion into the conversion unit 22, the frame structure data 405 from the interlaced scan image 404, and the re-encoded data 40
6 is performed by the re-encoding unit 23, respectively. (Conversion / re-encoding device 2) First, the conversion / re-encoding device 2
The processing contents of each component will be described.

The decoding section 21 decodes the input coded data and returns it to the decoded progressively scanned image 403. Converter 22
Converts the decoded progressively scanned image 403 into an interlaced scanned image 4
Convert to 04. The conversion method is a decoding progressive scanning image 4
03, in which only odd lines or only even lines are extracted from each frame as fields, and odd lines are extracted from odd-numbered frames, and even lines are extracted from even-numbered frames. And

The re-encoding unit 23 converts the converted interlaced scanned image 404 output from the converting unit 22 into frame structure data 405 (30 frame seconds), that is, two fields of the interlaced scanned image 404. Is converted into one frame, and then compression-coded according to the MPEG2 system. At this time, the GOP structure of the frame structure data 405 is determined so that the frame derived from the I picture frame and the subsequent B picture frame in the decoded progressive scan image 403 become the I picture.

(Configuration of Encoding Apparatus 1) Next, referring to FIG. 1 again, the configuration of the encoding apparatus 1 and the processing contents of each component will be described below. The encoding device 1 includes a compression processing unit 1
This is a configuration having a pre-processing unit 20, an output unit 30, etc., centered on 0.

The pre-processing unit 20 temporarily stores the input moving image data in a built-in memory, rearranges the pictures in the order in which compression processing is to be performed, and divides each picture into macro blocks, which are processing units. , And sequentially output to the compression processing unit 10. Note that the macroblock output from the preprocessing unit 10 has accompanying information indicating which picture belongs to which position, and which picture belongs to the P or B picture as a reference picture. Has been added,
Encoding is performed according to this information.

The GOP structure here is M = 3, N = 30
(P pictures appear every three pictures and the number of frames (pictures) included in one GOP is 30). 1 GO
If the pictures for P are displayed in the display order, "BBI BBP BB
P BBP BBP BBP BBP BBP BBP BB
P ". FIG. 2 is a schematic diagram showing an overview of the rearrangement of pictures by the preprocessing unit 50. FIG. 3A shows the order of the original pictures in the moving image data. "P",
“B” indicates the type of picture, and the number in parentheses indicates the order in the original order. On the other hand, FIG. 2B shows the order in which these pictures are compression-encoded.

The compression processing unit 10 performs motion compensation prediction (for P and B pictures), DCT conversion processing, quantization processing, and encoding processing on the image data (macroblock) output from the preprocessing unit. And outputs it to the output unit 30. The compression processing unit 10 determines a standard (bit allocation) of the data amount (encoded data amount) after the encoding process for each picture, and adjusts a quantization step in the quantization process.

The output unit 30 temporarily stores the encoded data output from the compression processing unit 20 in a built-in buffer and outputs the encoded data to a recording device (not shown) at a predetermined bit rate. (Compression processing unit 10) The compression processing unit 10 includes a motion compensation prediction unit 12, a DCT processing unit 13, a quantization processing unit 14, a variable length coding processing unit 15, an inverse quantization processing unit 16, an inverse DCT processing unit 17, The reference data memory 18 and the control unit 11 are provided.

The motion compensation prediction unit 12 detects a motion vector for an input image macroblock of a P or B picture, generates difference data from a reference image, and generates a difference data from a reference image.
Output to 3. The DCT processing unit 13 performs DCT processing on the input macroblock, converts the input macroblock into a post-DCT macroblock, and outputs the converted macroblock to the quantization processing unit 14.

The quantization processing unit 14 performs a quantization process on the post-DCT macroblock output from the DCT processing unit 13 and outputs the result to the variable-length coding processing unit 15. The amount of encoded data of the picture is determined by the size of the quantization step in the quantization process.
Changes the quantization step for each macroblock in accordance with the bit allocation value from the control unit 11 so that the amount of encoded data of each picture approaches the bit allocation value determined by the control unit 11. In the case of macroblocks of I and P pictures, the motion compensation prediction unit 12 refers to the macroblock as a reference image. The quantized data is output to the inverse quantization processing unit 16.

An inverse quantization processing section 16 and an inverse DCT processing section 17 following the inverse quantization processing section 16 perform an inverse quantization process and an inverse DCT process on the quantized data of the I and P pictures serving as reference images in the motion compensation prediction. To generate reference image data and store it in the reference data memory 18. The control unit 11 determines a quantization step of the macroblock after the compression processing for each picture and notifies the quantization processing unit 14 of the quantization step. Control unit 11
Is performed each time new picture data is transmitted from the preprocessing unit 30.

That is, the scanning method at the time of broadcasting the moving image data to be compression-encoded (whether the image data is converted to interlace and re-encoded), and the type of each picture (I,
P, B) Further, according to whether the picture is re-encoded and converted into an I-picture, bit allocation is performed for each picture, and the bit allocation value is notified to the quantization processing unit 14. This is performed not only for the purpose of controlling the bit rate of the encoded data amount after the compression encoding, but also for the purpose of increasing the image quality of the image to be re-encoded as much as possible.
The bit allocation processing by the control unit 11 is based on the test model 5 described above, and is limited to a specific B picture (B picture B1 in the example of FIG. 4) which is converted into an I picture when re-encoded. Special control is performed to increase the value of the bit allocation. As a result, the picture quality of the B picture is prevented from deteriorating in the process of compression and decoding, and the picture quality of the I picture at the time of re-encoding processing generated using the data is improved.

For this reason, the control unit 11 includes two types of bit allocation value calculation units (a first calculation unit 111 and a second calculation unit 112).
And a selection unit 113 for selecting which calculation unit to use
And the method of calculating the bit allocation value is switched between image data requiring conversion and re-encoding and image data not requiring conversion / re-encoding. The selection unit 113 receives an instruction of the broadcast format of the moving image data (whether it is a target of conversion / re-encoding) from the operator via an interface unit (not shown), and calculates a calculation unit to use based on the instruction. Select The second calculation unit 112 is selected for data that is broadcast at 60 frames / second without conversion / recoding, and the first calculation is performed for data that is converted / recoded and broadcast at 30 frames / second. Part 1
Select 11.

Next, the contents of the method of calculating the bit allocation value (calculation formula) for each of the two calculation units will be described. Note that in the following arithmetic expressions, i, p, b
The type lowercase letters indicate that they correspond to an I picture, a P picture, and a B picture, respectively. First, the second calculation unit 112 calculates a bit allocation value according to the same method of the test model 5 as in the past. Hereinafter, an outline of a method of calculating the amount of encoded data allocation by the test model 5 will be described. In determining the allocation, the amount of encoded data per GOP (or the remaining amount thereof), the complexity of each picture attribute, and the number of pictures of various Information on the number of processing frames is used. Second bit amount calculation unit 112
Uses three different formulas for each picture type,
Bit allocation is determined according to the type of the picture. The information on the picture type is transmitted from the preprocessing unit 20.

The complexity of a picture refers to the amount of encoded data (Si, Sp, S
b) and the average quantization scale (Qi, Qp, Qb) at the time of the last encoding is a value obtained by substituting into the following equation. For an I picture: Xi = Si × Qi For a P picture: Xp = Sp × Qp For a B picture: Xb = Sb × Qb Then, the second calculator 112 calculates the amount of encoded data per GOP (remaining amount): R, Complexity for each picture attribute (X
i, Xp, Xb) and the number of unprocessed frames (Np, Nb) in the GOP are substituted into any of the following three equations to determine the bit allocation (Ti, Tp, Tb) of the picture. For I picture: Ti = R / (1 + NpXp / Xi + Nb
Xb / XiKb) For P picture: Tp = R / (Np + NbXb / XpK
b) For B picture: Tb = R / (Nb + NpKbXp / X
b) Kb is a weighting coefficient for bit allocation, where 1.b.
The value is 4. The larger the weighting value is from 1.0, the smaller the bit allocation value of the B picture is for I and P pictures. Note that the value of R is updated in such a manner that the actual generated encoded data amount is subtracted each time the encoding process for one picture is completed.

On the other hand, the first calculating unit 111 obtains a bit allocation value by a method unique to the present invention, basically based on the test model 5. That is, the method of the test model 5 is partially changed so that when the same image is re-encoded by the decoding / re-encoding device 2 among the B-picture frames in the progressively scanned image, Regarding the converted image (B picture B1 in FIG. 4), the bit allocation amount is increased, and the bit allocation amount of another frame in the same GOP is reduced by the increase in the bit allocation amount of the B picture. is there. For this purpose, the first calculation unit 111 uses one type for I pictures and two types for P pictures.
There are three types of calculation formulas for the type and the B picture, and these are used properly to increase the bit allocation to a specific B picture. Broadly speaking, the second calculation unit 1
The bit allocation value is increased by excluding the weighting value applied to the B picture in 12 from being applied only to the specific B picture.

Since the processing up to the calculation of the complexity is the same as that of the second calculation unit 112, the description is omitted. The following six types of equations determine the bit allocation of a picture. For I picture: Ti = R / (1 + NpXp / Xi + (Nb-1) Xb / XiKb + Xb / Xi) <Formula I> For P picture (1): Tp1 = R / (Np + (Nb-1) Xb / XpKb + Xb / Xp) <Formula P1> For P picture (2): Tp2 = R / (Np + NbXb / XpKb) <Formula P2> For B picture (1): Tb1 = R / (Nb + NpXp / Xb + 1) <Formula B1> For B picture (2): Tb2 = R / (Nb + NpKbXp / Xb + Kb) <Formula B2> For B picture (3): Tb3 = R / (Nb + NpKbXp / Xb) <Formula B3> where Tb1 is This is a formula for calculating a bit allocation value of a B picture to be converted into an I picture in re-encoding. Tp1
Is an arithmetic expression for a P picture encoded before a B picture (I picture at the time of re-encoding) processed by the expression of Tb1, and Tp2 is a P expression encoded after the B picture. This is an operation expression for a picture. Similarly, Tb2 is an arithmetic expression for a B picture processed before the B picture, and Tb3 is an arithmetic expression for a B picture processed later. By allocating bits using such an arithmetic expression, the bit allocation becomes larger for a B picture that becomes an I picture at the time of re-encoding than for other B pictures.

When the first calculation unit 111 or the second calculation unit 112 calculates the bit allocation value of the picture based on the above equation, the control unit 11 notifies the quantization processing unit 14 of the calculation result. An instruction is given to control the quantization process using the bit allocation value as a guide. The control unit 11 receives the information on the data amount of the encoding processing result of the picture from the variable-length encoding processing unit 15 at the time when the encoding processing for the picture is completed, and holds this value. GOP
, The value of R is updated by subtracting from the remaining bit amount R, and the process proceeds to the next picture.

Further, in order for the first calculation unit 111 to use the above calculation formula properly, it is necessary to determine a specific B picture (one that is converted into an I picture at the time of re-encoding).
The control unit 11 includes a picture determination unit 1
14. This specific B picture is located immediately after the I picture in the display order in the moving image data. However, since the pictures constituting the moving picture data are rearranged by the preprocessing unit 20 (see FIG. 2), the B picture immediately following the I picture in the order of input to the encoding unit 10 is assigned to the B picture. Not necessarily a picture.

Hereinafter, a method in which the picture determination section 114 identifies the specific B picture will be described. The image data entering the compression processing unit 10 is rearranged as shown in FIG. 2, and is displayed in the order of “B (1) -B (2) -IB (3)-
B (4) -P ", the order of the encoding process is" IB (1) -B (2) -PB (3) -B (4) ". Pictures are coded fifth. The B picture (3) immediately after the I picture in the display order is a GOP
Are encoded immediately after the earliest encoded P picture among the P pictures. Thus, the picture determination unit 114 identifies “B picture coded immediately after the first P picture” using two flags.

The two flags are a P flag and a B flag, both of which have an initial value of "OFF". The P flag is switched to “ON” when the encoding of the I picture is performed, and “O” is performed when the first P picture is encoded.
FF ". On the other hand, the B flag is switched to “ON” when the first P picture (the first P picture is determined by the P flag) is encoded, and the B picture that appears next to the first P picture Is returned to "OFF" at the time when is encoded. Therefore, the P flag = “O”
The B picture appearing in the state of “FF” and B flag = “ON” is determined as the B picture to be obtained.

When this is applied to the rearranged picture in FIG. 2, the P flag and the B flag are initially set to (OFF,
OFF), but becomes (ON, OFF) when the I picture is encoded. Since the flags of the subsequent two B pictures are (ON, OFF), bit allocation is performed by equation Tb2. Then, when the P picture appears, the flag is changed to (OFF, ON), and for this P picture, bit allocation is performed by equation Tp1. Then, the desired B picture (B (3)) appears, and the expression Tb1
, And the flag becomes (OFF, OFF). After that, the value of the flag remains unchanged (OFF, OFF) until the end of the GOP.
2. Bit allocation is performed by equation Tb3. These processes are
It is repeated for each GOP. (Operation) Hereinafter, an operation when the encoding apparatus 1 processes moving image data to be decoded / re-encoded for one GOP will be described with reference to a flowchart.

FIG. 4 is a flowchart showing the operation of the encoding device 1. First, at the beginning of a GOP, a P flag and a B flag are initialized (S401). Thereafter, the image data for each picture is sequentially read from the preprocessing unit (S402). Then, the type of the picture is confirmed.
In the case of an I picture (S403: “I”), the P flag is set to “ON” (S404), and the control unit 11
Calculates the bit allocation value using the above “Equation I” (S
405). Then, the bit allocation value is notified to the variable length coding processing unit 15, and the components subsequent to the DCT processing unit 13 perform compression coding of the picture.

In the case of a P picture (S403: "P"),
The control unit 11 refers to the P flag and sets the P flag to “ON”.
If so (if it is the first P picture in the GOP), a bit allocation value is calculated using the above-mentioned "Equation P1" (S407), and the P flag is set to "OFF" and the B flag is set to "ON". It is set (S408). On the other hand, if the P flag is “OFF”, the bit allocation value is calculated using “Equation P2” (S409). And the control part 11
The obtained bit allocation value is notified to the variable-length coding processing unit 15, and the components subsequent to the DCT processing unit 13 perform compression coding of the picture.

In the case of a B picture (S403: "B"),
The control unit 11 refers to the P flag and the B flag, and calculates a bit allocation value by using an appropriate expression. That is, if the B flag is “ON” (B picture appearing next to the first P picture in the GOP) (S410: Yes), the bit allocation value is calculated using “Equation B1”, and the B flag is set to “ O
FF ”(S411), and if the B flag is“ OFF ”and the P flag is“ ON ”(S410: No, S41)
3: Yes), a bit allocation value is obtained using “Equation B2” (S414). If the B flag is “OFF” and the P flag is “OFF” (S410: No, S413: No),
A bit allocation value is obtained using “Equation B3” (S41)
5). Then, the control unit 11 notifies the variable-length encoding processing unit 15 of the obtained bit allocation value, and the components subsequent to the DCT processing unit 13 perform compression encoding of the picture.

Then, the processing of steps S402 to S415 is repeated until the last picture of the GOP is processed (S416). Note that the GOP shown in FIG.
The unit process is repeated until the end of the moving image data. (Summary) As described above, the encoding apparatus 1 according to the present embodiment
Is used to compress and encode moving image data of 60 frames / sec, which is later subjected to a process of “decoding → conversion → re-encoding”, in a re-encoding process after conversion of B pictures.
For a picture referred to as a picture, the bit allocation value at the time of encoding is increased. As a result, the converted 30
Since the deterioration of the image quality of the I picture in the moving image of frame / second can be suppressed, the deterioration of the image quality of the image after re-encoding can be suppressed.

The method of improving the image quality of a B picture that becomes an I picture during re-encoding is not limited to the method of the present embodiment. As another method, for example, after calculating the bit allocation of the picture by (step 3) of the conventional method,
The bit allocation of the B picture immediately after the I picture is multiplied by Kb. In the description of the present embodiment, the B picture located immediately after the I picture in the display order of the sequentially scanned image is assumed to be converted into an I picture at the time of re-encoding, but the immediately preceding B picture is converted into an I picture. In some cases. In that case, the bit allocation is increased for the immediately preceding B picture.

As means for increasing the coded data amount allocation of a B picture adjacent to an I picture, in the present embodiment, a configuration for performing a unique operation called a first calculation unit 111 is provided. The means for performing is not limited to this. For example, a fixed ratio of the coded data amount per GOP is excluded in advance, and the calculation of the quota for each picture is performed by a conventional calculation (second calculation unit 112). According to the calculation result, the allocation data amount of the B picture to be allocated is the same as the other B pictures,
Here, the excluded data amount is added to the data amount of the calculation result. With this method, only one calculation unit is required, so that the configuration is also simplified.

In the description of the present embodiment, the sequential scanning I
Although the image quality of the B picture immediately after the picture has been improved, it is important that the image quality of a picture referred to as an I picture during re-encoding be improved. Therefore, if an I-picture is generated from a P-picture and a B-picture at the time of re-encoding, the encoding device will perform image quality improvement processing on the P- and B-pictures.

In the present embodiment, for a B-picture whose image quality is to be higher than that of other B-pictures, the image quality is improved by adjusting the quantization step. However, other methods are also possible. As another method, for example, the attribute of the picture is changed from a B picture to an I picture or P picture.
Intra-frame encoding of the entire picture without changing to a picture or performing motion compensation prediction only for the B-picture can be considered.

In this embodiment, as a method of converting a frame of a progressively scanned image into a field of an interlaced scanned image, a method of simply thinning out lines has been described.
The conversion method is not limited to this. For example, a low-pass filter may be used. The effect of the present invention is effective as long as the I picture in the converted interlaced image is generated based at least partially on the image data of the B picture in the progressive scan image before the conversion.

Although the present embodiment has been described with respect to broadcast image data, this is merely an example, and the effect of the present invention is that once coded moving image data is decoded,
The effect is obtained without exception when re-encoding is performed with the frame frequency reduced. In the present embodiment, a method based on a flag value at the time of compression processing is adopted as a method of recognizing a B picture for which bit allocation is to be increased, but other methods are also possible. For example, before the picture is rearranged in the preprocessing unit, the target B
A picture is recognized (immediately before, immediately after an I picture, etc.), and additional information (dedicated flag) is added to the B picture and transmitted to the compression unit. For example, a signal is sent to a compression unit.

[0047]

As is apparent from the above description, the coding apparatus of the present invention performs compression coding processing using motion compensation prediction on video data to be decoded and re-coded later. An identification device for identifying a frame to be referred to as an I-picture attribute frame in motion-compensated prediction at the time of re-encoding processing, among the frames constituting the video data, For the frame identified by the means, the quantization step determining means for reducing the quantization step in the quantization process compared to other frames of the same picture attribute, and according to the quantization step determined by the quantization step determining means. Encoding means for encoding the frame identified by the identification means, whereby the MPEG2 code Even if the decoded image is subsequently decoded and re-encoded, a large amount of coded data is allocated in advance to a frame to be referred to as an I-picture at the time of re-encoding. Image quality of an I-picture at the time of encoding, that is, image quality of an image after re-encoding can be prevented.

Further, the identification means identifies a B picture attribute frame adjacent to a frame to which an I picture attribute is added in the motion compensation prediction by the encoding device in display order, and the quantization step determination means includes: If the quantization step is smaller for the frame identified by the identification means than for the other frames having the same picture attribute, the image quality of the frame referred to as the I picture in the re-encoding can be surely improved. Can be improved.

In order to obtain the above effect, specifically, the quantization step determining means increases the bit allocation for the frame identified by the identifying means so that the quantization of the frame identified by the identifying means is performed. It is assumed that the quantization step in the quantization process is made smaller than other frames having the same picture attribute.

[Brief description of the drawings]

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

FIG. 2 is a diagram showing an outline of picture rearrangement performed by the encoding device according to the embodiment.

FIG. 3 is a flowchart showing an operation of the encoding device according to the embodiment.

FIG. 4 is a diagram showing a process in which moving image data is converted through encoding and decoding processes and is re-encoded.

[Explanation of symbols]

 Reference Signs List 1 encoding device 10 compression processing unit 11 control unit 111 first calculation unit 111 second calculation unit 112 selection unit 113 picture determination unit 12 motion compensation prediction unit 13 DCT processing unit 14 quantization processing unit 15 variable-length coding processing unit 2 Conversion / re-encoding device 21 progressive scanning decoding unit 22 conversion unit 23 re-encoding unit

Claims (6)

[Claims] 1. An encoding apparatus for performing a compression encoding process using motion compensation prediction on video data to be decoded and re-encoded later, wherein: An identification unit for identifying a frame to be referred to as a frame having an I picture attribute in motion compensation prediction during re-encoding processing; and for the frame identified by the identification unit, another frame having the same picture attribute A quantization step determining means for reducing the quantization step in the quantization processing, and an encoding means for performing the encoding processing of the frame identified by the identification means according to the quantization step determined by the quantization step determination means. An encoding device comprising: 2. The identification means identifies a B picture attribute frame adjacent to a frame to which an I picture attribute is assigned in a motion compensation prediction by the encoding device in display order, and the quantization step determination means includes: 2. The encoding apparatus according to claim 1, wherein the quantization step is smaller for the frame identified by the identification means than for other frames having the same picture attribute. 3. The quantization step determining means increases the bit allocation to the frame identified by the identification means, so that the quantization step in the quantization processing of the frame identified by the identification means has the same picture attribute 2. The frame is made smaller than the frame of (1).
Or the encoding device according to 2. 4. An encoding method for performing compression encoding processing using motion compensation prediction on video data to be decoded and re-encoded later, wherein: An identification step of identifying a frame to be referred to as an I-picture attribute frame in the motion compensation prediction at the time of re-encoding processing; A quantization step determining step of making the quantization step smaller in the quantization processing than the frame; and a code for performing the coding processing of the frame identified in the identification step in accordance with the quantization step determined in the quantization step determination step. Encoding step. 5. In the identification step, a frame having a B picture attribute adjacent to a frame to which an I picture attribute is added in a motion compensation prediction in the encoding process in display order is identified. 5. The encoding method according to claim 4, wherein the quantization step is smaller for the frame identified in the identification step than for other frames having the same picture attribute. 6. In the quantization step determining step,
By increasing the bit allocation to the frame identified in the identification step, the quantization step in the quantization process of the frame identified in the identification step is made smaller than other frames of the same picture attribute,
The encoding method according to claim 4 or 5, wherein: