JP3363036B2 - Video encoding bit stream converter - Google Patents

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a moving picture coded bitstream conversion apparatus, and more particularly, to a moving picture encoding apparatus using various kinds of information including moving picture information included in a compressed bitstream. The present invention relates to a moving image coded bitstream conversion device capable of easily converting a coded bitstream in a television system with high image quality.

[0002]

2. Description of the Related Art Conventionally, when performing television transmission between regions having different television systems, TV system conversion is performed. Further, conventionally, in order to maintain a high image quality of a TV image at the time of this TV system conversion, a motion correction TV system conversion utilizing a motion vector which is a result of estimating a motion of a moving image is adopted. Regarding this motion compensation TV system conversion method, for example, Japanese Patent Publication No.
No. 25119, "Motion picture frame rate conversion method using motion amount" and the like are described in detail.

Here, the outline will be described. A general moving image including a television image is not given as a temporally continuous signal on the basis of displaying still images one after another at a constant frame rate per second. For example, looking at the broadcast television system, a frame rate of 30 frames / second is adopted in Japan, while a frame rate of 25 frames / second is adopted in European countries. Therefore, when exchanging broadcast programs between countries that adopt television systems having different frame rates, a frame rate conversion technique is indispensable.

FIG. 8 shows a principle diagram of the frame rate conversion. In the figure, the frame rate is 3 from the image (a25, b25, c25, ...) with a frame rate of 25 frames / sec.
It shows conversion to images (a30, b30, c30, ...) Of 0 frames / sec. Now, assuming that a25 and a30 frames are synchronized, the frame rate ratio is 25:
Since 30 = 5: 6, both frames are synchronized every 6 frames.

As the frame to be synchronized, the frame of the original image can be used as it is, but the frame not to be synchronized is created by interpolating two adjacent frames of the original image.
For example, a b30 frame is created by interpolating adjacent a25 and b25 frames, and a c30 frame is created by interpolating adjacent b25 and c25 frames.
When generating this interpolated frame or interpolated field, it is necessary to correct the motion of the original image.

In various motion estimations when performing field interpolation by performing this motion correction, conventionally, assuming the uniformity of motion in an image, one field image is divided into a number of blocks, The motion estimation is performed every time. Specifically, a motion is detected for each block between two adjacent fields, and an interpolation field is created based on the vector by referring to the two fields.

Next, although it is a technique completely different from the TV system conversion, in recent years, in order to transmit a moving image, MPE is used.
Digital compression encoding represented by G-2 and the like has come to be used. Regarding this MPEG-2, ITU-T Recommendation H.264 is used. 262 and the like. Many of these compression encoding devices are MPEG-2.
The motion vector, which is the result of estimating the motion of the moving image, is used to improve the compression efficiency.

Conventionally, the motion-compensated TV system conversion and compression encoding are performed independently, and no unified processing is performed. An example of a conventional device that realizes both of the above functions will be described with reference to FIG.

When a coded input bitstream a of a TV image is input, the decoding device 31 decodes it and restores it to the original TV system image (hereinafter referred to as input system video) b. . The motion compensation system conversion device 32 is composed of a motion estimation unit 32a and a line number / field number conversion unit 32b. The motion estimation unit 32a obtains a motion vector from the input system image b and sets the motion vector c to the line number / line.
It is supplied to the field number conversion unit 32b. The line number / field number conversion unit 32b uses the motion vector c to perform conversion to another system using the technique disclosed in Japanese Patent Publication No. 3-25119. The converted TV system image (hereinafter referred to as an output system video) d is output from the line number / field number conversion unit 32b.

Next, the output method image d is input to a compression encoding device such as MPEG-2, that is, a motion compensation encoding device 33. The motion compensation coding device 33 includes a motion estimation unit 3
3a and an encoding unit 33b, and the motion estimating unit 33a generates a motion vector from the output method image d. The encoding unit 33b encodes the output system video d using the motion vector e and outputs it as a new output bit stream f.

As described above, conventionally, the motion compensation system conversion device 32 and the motion compensation coding device 33 operate completely independently.

[0012]

[Problems to be Solved by the Invention] Motion compensation TV set forth above
The processing for obtaining the motion vector occupies a large part of the hardware in both the system conversion device 32 and the motion compensation encoding device 33, but this cannot be omitted in either device. That is, in the system conversion device 32, the quality of the TV image after the TV system conversion cannot be kept good unless the motion estimation of the moving image is accurately performed. In other words, accurate motion estimation of moving images is
It can be said that this is the first important factor for keeping the converted image quality high. On the other hand, also in the motion compensation coding apparatus 33, the motion estimation is an important factor for significantly increasing the compression efficiency.

As described above, since the motion compensation TV system conversion and the motion compensation coding are conventionally performed independently, it is necessary to duplicate the motion estimation processing, and a device realizing both functions will be considered. In that case, there is a problem in that the hardware of the entire apparatus is increased. In other words, the motion estimators 32a and 33a are respectively the motion compensation TV system converter 32 and the motion compensation encoder 33.
Occupies a large part of the hardware, and there was a problem that the overall hardware scale would be enormous.

Further, in the prior art, as is apparent from FIG.
Since the motion compensation TV system conversion device 32 generates a motion vector using the image once decoded by the decoding device 31, if the decoded image is deteriorated, the motion estimation unit 32a causes the deteriorated image to be generated. Motion estimation based on
There is a problem in that the motion estimation may cause an erroneous estimation.

An object of the present invention is to eliminate the above-mentioned problems of the prior art and to effectively use various coding control information such as motion vectors existing in these compressed bitstreams, while converting motion compensation methods. Another object of the present invention is to provide a moving image coded bitstream conversion device capable of performing high-quality and simple moving image coded bitstream conversion by performing re-encoding thereafter.

[0016]

In order to achieve the above-mentioned object, the present invention provides a moving picture coded bit stream conversion for converting a bit stream obtained by coding a TV signal of a certain system into a bit stream of another system. In the apparatus, encoding control information extracting means for extracting encoding control information including a motion vector for each macroblock, an encoding mode, and a quantization step size from an input method bitstream, and using the encoding control information, Decoding means for decoding a bitstream, motion vector calculating means for converting the extracted motion vector into a motion vector for motion compensation method conversion using the encoding mode information, and the decoding using the motion vector System conversion means for converting the image obtained by the conversion into an image of another system, the motion vector, and the interpolation field position information It generates a motion vector using the fine the quantization step size, is characterized in that comprises a coding means for coding the image after the system conversion using animal vector.

According to the present invention, by using the encoding control information of the bit stream extracted by the encoding control information extracting means, the motion vector required for the TV system conversion and the motion vector used for the re-encoding are used. As a result, it is possible to generate the image, and it is possible to improve the image quality of the image converted by the TV system, and it is possible to significantly reduce the hardware scale of the device.

[0018]

BEST MODE FOR CARRYING OUT THE INVENTION An embodiment of the present invention will be described in detail below with reference to the drawings. FIG. 1 is a block diagram showing the configuration of an embodiment of the present invention. In addition, in FIG.
Signals that are the same as or equivalent to those in FIG. 9 are assigned the same reference numerals.

In FIG. 1, an input bit stream a, which is an image of a certain TV system and is encoded, is input to the decoding device 1. The decoding device 1 includes a coding control information extraction unit 1a that extracts coding control information such as motion vector information, coding control information (coding mode information), and a quantization step size from the input bitstream a, The decoding unit 1b is configured to decode the input bitstream a using the coding control information m.

Next, the motion compensation TV system conversion device 2 uses the motion vector and the interpolated field position information p in the encoding control information m to calculate a motion vector n suitable for TV system conversion.
And a line number / field number conversion unit 2b. Further, the motion compensation coding apparatus 3 includes a coding control unit 3a for generating a motion vector q required for re-coding from the coding control information m, the motion vector n and the interpolated field position information p, and the motion control coding unit 3a. It is composed of an encoding unit 3b that performs re-encoding using the vector q.

Next, the operation of the motion vector calculator 2a will be described in detail. The motion vector calculation unit 2a calculates a motion vector of an interpolated field (for example, a30, b30, c30, ... Of FIG. 8) generated by the conversion of the TV system.
For example, as shown in FIG. 2, it functions to calculate the motion vector of the interpolation field 6 from the motion vectors of the front field 5 and the rear field 7.

First, it is assumed that the motion vector MV extracted from the MPEG-2 stream is already registered in the macroblock of each field, for example, the macroblock A of the preceding field 5 and the macroblock C of the following field 7. , The motion vector of the interpolation field 6 is obtained as follows.

(1) As shown in FIG. 3 (a), when the backward vector v1 exists in the macroblock A and the forward vector v2 exists in the macroblock C, both the A and C interpolation fields 6 There is a prediction vector from the direction. Therefore, since both of the vectors have relatively high reliability, the vector v2 which is closer to the macroblock B in terms of distance is adopted as the motion vector of the macroblock B of the interpolation field 6. The position information p of the interpolation field 6 may be obtained from the line number / field number conversion unit 2b, or may be held in advance by the motion vector calculation unit 2a as data.

(2) As shown in FIG. 3 (b), when the backward vector v1 exists in A and the forward vector does not exist in C, the vector from the direction of the interpolating field 6 exists in C. Does not exist. Therefore, there is a possibility that the continuity is not maintained in the rear field 7 and the fields before it. Therefore, since the vector V1 of A is more reliable, the vector V1 of A is adopted as the motion vector of the macroblock B.

(3) As shown in FIG. 3 (c), when the forward vector v2 exists in C and the backward vector does not exist in A, for the same reason as in (2), The vector v2 is adopted as the motion vector of the macroblock B.

(4) As shown in FIG. 3 (d), when there is no backward vector in A and no forward vector in C, both A and C are from the interpolation field 6 direction. The prediction vector of does not exist. In such a case, it may be possible that the motion of the image is complicatedly changed. Therefore, it is safer to allocate the 0 vector rather than forcibly estimating the motion of B. Therefore, the 0 vector is adopted as the motion vector of the macroblock B.
The case where A and C are intra-coded is also included in this case.

Next, the motion vectors z1 to nz of the macro blocks B1, B2, ..., Bn on the interpolation field 6 obtained in the above (1) to (4) are as shown in FIG. 4, for example. It is thought that they are different from each other. When the macroblock vectors z1 to nz are used as they are for the conversion of the line number / field number conversion unit 2b, block distortion becomes conspicuous. Therefore, an order statistical filter is applied to the vector on the interpolation field 6 to remove this variation. Regarding the configuration example and operation example of this order statistical filter, Japanese Patent Application No. 6-192253 “Motion vector processing device”, which is a patent application by the present inventor, is given.
In detail. The outline of the operation of the order statistical filter is to obtain the size of each vector of a macroblock of interest and macroblocks around it (total 9 macroblocks), and arrange the sizes in ascending or descending order, A vector having a size located in the middle is used as the size of the macroblock of interest.

Next, since one macroblock B has 16 pixels × 16 lines, it is too large as a block for motion compensation system conversion that requires true motion, so that it can cope with the motion of a smaller object. I have to
Therefore, the block on the interpolation field 6 is subdivided into blocks of, for example, 4 pixels × 2 lines. The vector of the subdivided block is assigned to each subdivided block by well-known linear interpolation using the macroblock vector obtained by the above order statistical filter as a representative vector.

Next, one of the deteriorations of the motion compensation method conversion is that a stationary object such as a ball or a caption in the screen is erroneously given a motion. To prevent this, the vector of each subdivided block is compared with the 0 vector. Specifically, as shown in FIG. 5, the subdivided block 10 on the interpolated field 6 is replaced by the preceding and following fields 5.
And 7 to obtain the difference absolute value DFD, the front and rear fields 5 of 0 vector and the 7 difference absolute value DFD, and the smaller difference absolute value DFD is adopted.

Finally, in order to further remove the variation of the vector, the order statistical filter is again applied to the vector obtained by the difference absolute value DFD. The motion vector n (see FIG. 1) thus obtained is used to perform the motion correction method conversion in the line / field number converter 2b.

Next, the line / field number converter 2
The operation of the motion compensation coding apparatus 3 that re-encodes the output format video d format-converted by b will be described. In the output method image d, as shown in (a) 30 and (g) 30 of FIG. 8, fields created by conversion only by line interpolation are periodically present. These are used as I or P pictures requiring high image quality in MPEG-2 because they are not deteriorated by motion vector errors. The picture structure is a field picture in view of the fact that TV system conversion is performed in field units and motion vectors are also allocated in each field.

The coding control unit 3a of the motion compensation coding apparatus 3
Is supplied to the coding unit 3b using the coding control information m extracted from the coding control information extraction unit 1a, the motion vector n calculated by the motion vector calculation unit 2a, and the interpolated field position information p. Generate a motion vector q. That is, the macroblock type is determined from the distribution method of the motion vectors assigned to the subdivided blocks in the macroblock, and the motion vector of the macroblock is also calculated from the subdivided block motion vector. Although various calculation methods can be considered, one method will be described below.

For simplification, B picture is not used. Calculate the variance σ _v ² of the subdivided motion vector in the macroblock. Certain thresholds are Th1 and Th2. Th2> σ _v ² ≧ Th1
, The variation of the vector is moderate, and it is considered that dividing the macroblock is effective.
Forecast. When σ _v ² <Th1, the vectors are relatively aligned, so field prediction is performed. In the case of σ _v ² > Th2, the vectors are extremely scattered, and it is considered that the inter coding efficiency is low, and thus intra coding is performed.

The following processing is performed for each 16 × 8 block in the case of 16 × 8 prediction, and for each 16 × 16 block in the case of field prediction.

(A) A block in which most of the 0 vectors are used is directly predicted from the same parity field (see FIG. 6).

(B) Variance σ of vector in block
_{If v2} ² (σ _v2 ² = σ _v ^{2 in} the case of field prediction) is set to a threshold value Th 3, and σ _v2 ² ≦ Th3, the motion can be regarded as uniform, and the reliability of the vector is also high. In this case, we take the average of the vectors and give it as shown in FIG.
It projects on the same parity and different parity fields, and predicts from the field where the designated position is closer to the half pixel position and the pixel position. In the case of FIG. 7, since the projected position of the vector of (A) is closer to the half pixel position of the same parity field, the prediction is performed from the same parity field.
The vector of (B) is the opposite.

(C) When σ _v2 ² > Th3, the variation of the vector is relatively large. In such a case, it is considered that the image changes rapidly, and prediction is performed from different parity fields that are close in distance.

Next, regarding the quantization parameter, the quantization scale code of each macroblock is controlled so as not to be as coarse as possible as compared with the quantization scale codes of the macroblocks of the fields before and after it is interpolated. To do.

As described above, by generating the motion vector q used for re-encoding in the encoding control unit 3a, it is not necessary for the encoding apparatus 3 to have an independent motion estimation unit as in the conventional case. , The hardware scale of the device can be significantly reduced.

[0040]

As described above, according to the present invention,
The coding control information including the motion vector information included in the input coded bit stream can be effectively used for the subsequent motion compensation method conversion and re-coding. Therefore, there is an effect that it is possible to realize a high image quality in the conversion of the moving image coded bit stream in the television system and a great simplification of the hardware of the entire apparatus.

[Brief description of drawings]

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

FIG. 2 is a conceptual diagram of an example of an interpolation field and front and rear fields.

FIG. 3 is an explanatory diagram of a method of obtaining a motion vector of an interpolation field.

4 is a diagram showing an example of a motion vector of a macroblock obtained by the method of FIG.

FIG. 5 is an explanatory diagram showing a method of correcting a vector of a subdivided block.

FIG. 6 is an explanatory diagram of an example of vector prediction in the case of 16 × 8 prediction and field prediction.

FIG. 7 is an explanatory diagram of another example of vector prediction in the case of 16 × 8 prediction and field prediction.

FIG. 8 is a diagram showing a conventional principle of frame rate conversion.

FIG. 9 is a block diagram of a conventional moving image coding bitstream conversion device.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 1 ... Decoding device, 1a ... Encoding control information extraction unit, 1b ... Decoding unit 1b, 2 ... Motion correction TV system conversion device, 2a ... Motion vector calculation unit, 2b ... Line number / field number conversion unit, 3 ... Motion Compensation coding device, 3a ... Coding control unit, 3
b ... Encoding unit.

─────────────────────────────────────────────────── ─── Continuation of the front page (56) Reference JP-A-7-123291 (JP, A) JP-A-8-70440 (JP, A) (58) Fields investigated (Int.Cl. ⁷ , DB name) H04N 7/01

Claims (6)

(57) [Claims] 1. A bit obtained by encoding a TV signal of a certain system.
A method for converting a stream into another format bitstream.
In the image coding bitstream conversion device, the motion of each macroblock is changed from the input method bitstream.
Vector, coding mode and quantization step size
Encoding control information extracting means for extracting encoding control information including
And recover the bitstream using the coding control information.
Decoding means for encoding, motion based on extracted motion vector and interpolation field position information
Compute the motion vector of the interpolation field of the correction method conversion
The motion vector calculation means and the motion vector are used to obtain the image obtained by the decoding.
Method conversion means for converting an image into an image of another method, a motion vector of the interpolation field and an interpolation field
A motion vector is generated using the position information, and the motion vector is generated.
Means for encoding the video after the system conversion using
Comprising the door, the motion vector calculating means, the extracted forward, among backward vector, and characterized in that in favor of the direction of the vector in the presence of the interpolation field employing the motion vector of the motion compensation type conversion Video coding bit stream conversion device. 2. A bit obtained by encoding a TV signal of a certain system.
A method for converting a stream into another format bitstream.
In the image coding bitstream conversion device, the motion of each macroblock is changed from the input method bitstream.
Vector, coding mode and quantization step size
Encoding control information extracting means for extracting encoding control information including
And recover the bitstream using the coding control information.
Decoding means for encoding, motion based on extracted motion vector and interpolation field position information
Compute the motion vector of the interpolation field of the correction method conversion
The motion vector calculation means and the motion vector are used to obtain the image obtained by the decoding.
Method conversion means for converting an image into an image of another method, a motion vector of the interpolation field and an interpolation field
A motion vector is generated using the position information, and the motion vector is generated.
Means for encoding the video after the system conversion using
And the motion vector calculation means adopts the 0 vector as a motion vector for motion compensation method conversion when there is no vector in the direction in which the interpolation field exists in the extracted forward and backward vectors. A moving image coding bitstream conversion device characterized by the above. 3. The video encoding bit stream converting apparatus according to claim 1 or 2, wherein the motion vector calculation means, in order to remove variations of the vector of the macro block, the vector of the macroblock A moving picture coded bit stream conversion device characterized by applying an order statistical filter. 4. The moving picture coded bit stream conversion device according to claim 3 , wherein the motion vector calculation means obtains a vector of a subdivided macro block by linear interpolation using the macro vector as a representative vector. A moving image coding bitstream conversion device characterized by the above. 5. Bits obtained by encoding a TV signal of a certain system
A method for converting a stream into another format bitstream.
In the image coding bitstream conversion device, the motion of each macroblock is changed from the input method bitstream.
Vector, coding mode and quantization step size
Encoding control information extracting means for extracting encoding control information including
And recover the bitstream using the coding control information.
Decoding means for encoding, motion based on extracted motion vector and interpolation field position information
Compute the motion vector of the interpolation field of the correction method conversion
The motion vector calculation means and the motion vector are used to obtain the image obtained by the decoding.
Method conversion means for converting an image into an image of another method, a motion vector of the interpolation field and an interpolation field
A motion vector is generated using the position information, and the motion vector is generated.
Means for encoding the video after the system conversion using
The encoding means uses a field created by conversion by line interpolation only as an I or P picture requiring high image quality in compression encoding such as MPEG-2. Video coding bitstream converter. 6. A bit obtained by encoding a TV signal of a certain system.
A method for converting a stream into another format bitstream.
In the image coding bitstream conversion device, the motion of each macroblock is changed from the input method bitstream.
Vector, coding mode and quantization step size
Encoding control information extracting means for extracting encoding control information including
And recover the bitstream using the coding control information.
Decoding means for encoding, motion based on extracted motion vector and interpolation field position information
Compute the motion vector of the interpolation field of the correction method conversion
The motion vector calculation means and the motion vector are used to obtain the image obtained by the decoding.
Method conversion means for converting an image into an image of another method, a motion vector of the interpolation field and an interpolation field
A motion vector is generated using the position information, and the motion vector is generated.
Means for encoding the video after the system conversion using
The coding means determines a macroblock type from a distribution method of motion vectors assigned to subdivided blocks in the macroblock, and calculates a macroblock motion vector from the subdivided block motion vector. A moving image coding bitstream conversion device characterized by: