JPH08340536A - Image coder - Google Patents

Abstract

PURPOSE: To provide an image coder in which processing applying variable rate coding to plural image data and recording the resulting data to a medium with a prescribed capacity at a high speed sequentially. CONSTITUTION: New image data received by a moving image coder 1 are given at first go a 1st coding section 20 via an input changeover device 10, in which variable rate coding is conducted and the data generating amount of the coded data is stored in a storage section 30. The image data are given to the moving image coder again as soon as new image data are received. In this case, the image data are given to a 2nd coding section 40 via the input changeover device 10. The 2nd coding section 40 allocates a data amount to the image data based on the data generating amount stored in the storage section 30 and a limit of the data amount with respect to the entire image data and adjusts the quantization rate so that the data generating amount is less than the allocated data mount for the application of variable rate coding to the image data.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image coding apparatus for sequentially variable-rate coding a plurality of image data at high speed.

[0002]

2. Description of the Related Art Considering the method of encoding an image in terms of controlling the amount of data, it is classified into fixed rate encoding and variable rate encoding. FIG. 4 is a diagram showing the relationship between the reproduction time and the encoding bit rate when certain program data is encoded by the fixed rate encoding and the variable rate encoding, respectively. As shown in FIG. 4, in fixed rate coding, the bit rate is always constant. Further, in variable rate coding, the bit rate changes depending on the complexity of the scene. Since the period a is a relatively difficult scene period, the amount of generated data is large, and since the periods b ₁ and b ₂ are relatively simple scene periods, encoding can be performed at a low bit rate.

As described above, in variable rate coding, a large amount of data is assigned to a complicated image and a small amount of data is assigned to a simple image, so that encoding can be performed while maintaining a homogeneous image. On the other hand, in the fixed rate coding, in the period a in which the information generation amount is large, the information is insufficient as shown in the area A of FIG. 4, and the period b in which the information generation amount is small.
_{In 1} and b ₂ , excessive information is encoded as shown in area B of FIG. Therefore, when focusing on the image quality, the variable rate coding is generally considered to be a more effective coding method.

By the way, image data of a certain recording time is
When recording on a recording medium having a limited recording capacity, such as a digital video disc (DVD) device, the encoding rate must be controlled so that the image data can be recorded on the recording medium. In such a case, if fixed rate coding is performed, a bit rate R _f that satisfies Expression 1 is obtained, and the entire image data is coded at this bit rate R _f .

[0005]

[Number 1] _{R f × T ≦ S ··· (} 1) , however, R _f, coding bit rate, T is, Duration of the image data, S is a capacity of the recording medium.

When variable rate coding is performed,
In order that all the program data can be recorded, the recording area is calculated backwards to allocate a recording area to each image data, the encoding bit rate of each image is obtained based on the recording area, and the encoding is performed at that bit rate. Specifically, for example, as shown in FIG. 5, the image data is divided into scenes at each time t, and a data amount d suitable for each scene i is obtained.
(I) is set, a bit rate satisfying, for example, Expression 2 is obtained for each scene based on the ratio of the data amount d (i) in all image data, and the bit rate of that scene is calculated at that bit rate. It will be encoded.

[0007]

## EQU00002 ## _R.sub.v (i) .times.t.ltoreq. (S.times. (D (i) / D)) (2) where _R.sub.v (i) is the coding bit rate of the image data of scene i, t is the time of the scene i, S is the capacity of the recording medium, i is the scene number, d (i) is the encoded data amount of the image data of the scene i, and D is the data amount of all the image data. D = Σd (i).

[0008]

However, when image data is recorded by variable rate coding on a recording medium having such a predetermined recording capacity, the amount of data in each scene of the image data as described above is increased. In order to allocate the data, it is necessary to investigate the data amount of each scene prior to the actual encoding, and there is a problem that the encoding process as a whole takes a very long time. That is, the image data is reproduced once, encoded, the amount of generated data is checked, and the amount of data is assigned to each scene based on the result,
Then, it is necessary to perform the process of reproducing the image data again and performing the variable rate coding, which requires a processing time about twice as long as the time for reproducing the image data. This is a major obstacle to efficient work, especially when a program such as a movie is to be sequentially recorded on a recording medium such as a DVD by variable rate coding.
Improvement was desired.

Therefore, it is an object of the present invention to perform a variable rate encoding process for a plurality of image data in a shorter time, so that the supplied image data is entirely below a predetermined capacity. An object of the present invention is to provide an image coding apparatus capable of performing the above.

[0010]

In order to solve the above-mentioned problems, the above-mentioned two times of reproduction and coding necessary for the target variable rate coding are performed by pipeline processing. That is, the first encoding and the second encoding are sequentially performed in parallel, and the encoding of the image data is sequentially completed at the apparently one encoding time interval.

Therefore, the image coding apparatus of the present invention is
First coding means for variable rate coding continuous image data, data generation amount of each image data of the variable rate coded continuous image data, and coding for the entire continuous image data The data amount determining means for allocating the encoded data amount to each image data based on the limit value of the encoded data amount, and the data generation amount of each image data of continuous image data The second coding means for variable-rate coding the continuous image data so as to be equal to or smaller than the coded data amount assigned to the data, and the continuous image data to be newly coded for the first The continuous image data input to the encoding means, once encoded by the first encoding means, and the allocation of the encoded data amount by the data amount determining means is completed is converted into the first code. In simultaneous with the input of a new continuous image data in the means, and an image data input means for inputting to said second encoding means.

[0012] Preferably, the first encoding means has a simplified configuration of the second encoding means.

Specifically, the second encoding means compresses continuous image data, quantizing means for quantizing the compressed signal, and the quantized signal. Variable length coding means for variable length coding, dequantization means for dequantizing the quantized signal, decompression means for decompressing the dequantized signal, and based on the decompressed signal Motion detection means for detecting motion between images by
A motion compensating unit that performs motion compensation based on the detected motion, and a computing unit that calculates a prediction error of each image data of the continuous image data used for the compression based on the motion-compensated signal. And the first encoding means substantially includes the dequantization means of the second encoding means, the decompression means, the motion detection means, the motion compensation means, and the calculation means. It is a configuration that does not.

More specifically, the first coding means and the second coding means compress the signal based on each image data of continuous image data, and the compressed signal. Quantizing means, variable length coding means for variable length coding the quantized signal, dequantizing means for dequantizing the quantized signal, and dequantizing the dequantized signal. A decompressing means for decompressing the signal, a motion detecting means for detecting a motion vector by searching within a predetermined range for each pixel forming the image data based on the decompressed signal, and the detected motion vector An encoding unit that includes a motion compensation unit that performs motion compensation based on the motion-compensated signal; Thus, the motion detecting means of the first encoding means searches a range substantially narrower than the search range of the motion vector in the motion detecting means of the second encoding means to search the motion vector. Is a configuration for detecting.

More specifically, the first coding means and the second coding means compress the signal based on each image data of continuous image data, and the compressed signal. Quantizing means, variable length coding means for variable length coding the quantized signal, dequantizing means for dequantizing the quantized signal, and dequantizing the dequantized signal. Expansion means for expanding the signal, motion detection means for detecting motion between images based on the expanded signal, motion compensation means for performing motion compensation based on the detected motion, and motion compensation for the motion compensation. An encoding means for obtaining a prediction error of each image data of the continuous image data used for the compression based on the signal, which is the encoding means of each of the first encoding means. Any or any Multiple, or the operations performed in all, the corresponding easily performed configured with lower accuracy than the calculation of each unit of the second encoding means.

[0016]

In the image coding apparatus of the present invention, the continuous image data to be newly coded is selected by the image data input means and first inputted to the first coding means. In the first encoding means, the continuous image data is variable rate encoded, and based on the data generation amount of the encoded data and the limit value of the encoded data amount for the entire continuous image data. Then, the data amount determining means allocates the encoded data amount to each image data. Next, the continuous image data is input to the second encoding means via the image data input means. Then, in the second encoding means, the continuous image data is variable rate encoded so that the data generation amount of each image data becomes equal to or less than the assigned encoded data amount.

[0017]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS FIG. 1 shows a moving picture coding apparatus according to an embodiment of the invention.
~ It demonstrates with reference to FIG. The moving picture coding apparatus according to the present embodiment records moving picture data on a recording medium having a predetermined recording capacity, such as a digital video disc (DVD), in other words, a recording medium having a limited recording capacity. Is a moving picture coding device for moving picture coding experts group (MPEG2).
Based on the video coding method), compression, quantization,
Variable rate coding is performed to generate an image bitstream.

First, the configuration of the moving picture coding apparatus of this embodiment will be described. FIG. 1 is a block diagram showing the configuration of the moving picture coding apparatus according to the present embodiment. Video coding device 1
Is an input switch 10, a first encoding unit 20, a storage unit 30.
And a second encoding unit 40.

The input selector 10 has two input terminals a,
It is a switching means for inputting two types of continuous image data input from b in parallel to the first encoding unit 20 and the second encoding unit 40, respectively. The input terminals a and b of the two systems are respectively connected to a VTR device, and the image data reproduced by the VTR device are input from the input terminals a and b, respectively. Then, the input switch 10
First, the newly input image data first switches the transmission path of the signal so as to be input to the first encoding unit 20, is already input to the first encoding unit 20, and is then input again. The transmitted image data is switched so that the image data is input to the second encoding unit 40.

The first encoding unit 20 is means for encoding the image data input via the input switch 10 and obtaining the data generation amount. The first encoding unit 20 will be described with reference to FIG. FIG. 2 is a block diagram showing the configuration of the first encoding unit 20. First encoding unit 2
0 has a DCT unit 122, a quantization unit 123, a variable length coding unit 124, an external I / F 132, and a control unit 133.

The image data input to the first encoding unit 20 is first input to the DCT unit 122. DCT section 1
Reference numeral 22 performs a discrete cosine transform (DCT) on the input image signal to reduce spatial redundancy. The transform coefficient obtained as a result of DCT is input to the quantization unit 123. The quantizer 123
The transform coefficient obtained by the DCT unit 122 is quantized according to a predetermined quantization rate input from the control unit 133, and output to the variable length coding unit 124.

The variable length coding unit 124 variable length codes the quantized transform coefficient to generate an image bit stream, and determines the data amount of the image bit stream from the external I / F.
To 132. The external I / F 132 is an interface circuit with the storage unit 30, and the variable length coding unit 124.
The amount of input data is output to the storage unit 30 for each image. The control unit 133 is a control unit that instructs the quantization rate of the quantization unit 123 and controls each unit included in the first encoding unit 20.

The storage unit 30 of the moving picture coding apparatus 1 has a first
Is a means for recording the data generation amount of each image data output from the encoding unit 20, and is a hard disk device in this embodiment. The data recorded in the storage unit 30 is the second
It can be read by the encoding unit 40 of.

The second encoding unit 40 encodes the continuous image data input via the input switch 10 at an encoding rate determined based on the amount of data recorded in the storage unit 30. It is a means to turn into. Regarding the encoding unit 40,
This will be described with reference to FIG. FIG. 3 shows the second encoding unit 40.
FIG. 3 is a block diagram showing the configuration of FIG. Second encoding unit 40
Is an adder 101, a DCT unit 102, a quantization unit 103,
Variable-length coding unit 104, inverse quantization unit 105, inverse DCT unit 106, adder 107, frame memory 108, motion detection unit 109, motion compensation unit 110, switching circuit 111,
It has an external I / F 112 and a control unit 113.

The image data input to the encoding unit 40 is
The difference from the input from the switching circuit 111 is obtained in the adder 101, and the difference is output to the DCT unit 102. In the intra picture mode, since no significant image is input from the switching circuit 111, the adder 1
The image data input to 01 is directly applied to the DCT unit 102.
Is output to In the motion compensation prediction mode, image data based on the previous reproduced image is input from the switching circuit 111. Therefore, the difference is obtained by the adder 101,
The error is input to the DCT unit 102.

The DCT unit 102 applies a discrete cosine transform (DCT: D) to the image signal input from the adder 101.
iscrete Cosine Transform) to reduce spatial redundancy. The transform coefficient obtained as a result of the DCT is quantized by the quantizer 1.
It is input to 03. The quantization unit 103 quantizes the transform coefficient obtained by the DCT unit 102 according to the quantization rate input from the control unit 113, and outputs it to the variable length coding unit 104 and the inverse quantization unit 105. The variable length coding unit 104 performs variable length coding on the quantized transform coefficient, generates an image bit stream, and sequentially outputs the image bit stream.

The external I / F 112 is an interface circuit with the storage unit 30. The external I / F 112 reads the data generation amount of each image data when the program data recorded in the storage unit 30 is encoded by the first encoding unit 20, and outputs it to the control unit 113.

The control unit 113 calculates the quantization rate of each image based on the data generation amount of each image input from the external I / F 112, and the quantization unit 10 uses the quantization rate.
3 controls the quantizer 103 so that 3 quantizes each image. Specifically, the control unit 113 first allocates the encoded data amount to each image based on Equation 3 based on the data generation amount of each image and the recording capacity of the recording medium to be recorded. In the present embodiment, the coded data amount of the image data within the time is determined for each predetermined time t based on the equation 3.

[0029]

Dd (i) = S × (d (i) / D) (3) where dd (i) is the allocated data amount, i is the period number for each time t, and S is , Capacity of recording media, d (i)
Is the amount of data generated when the image data of the period i is encoded, D
Is a data generation amount at the time of encoding all image data.

Then, the coding bit rate is determined by the equation 4 based on the data amount allocated by the equation 3.

[0031]

## EQU00004 ## _R.sub.v (i) .ltoreq.dd (i) / t (4) where i is the period number for each time t, and _R.sub.v (i) is
The encoding bit rate of image data in period i, dd (i) is
The amount of data allocated.

The control unit 113 also records the image data in the frame memory 108, which will be described later, and the switching circuit 11.
The switching of 1 is also controlled.

The transform coefficient quantized by the quantizer 103 is also input to the inverse quantizer 105. The transform coefficient is inversely quantized by the inverse quantization unit 105, and further, inverse DC
Inverse TDC 106 performs inverse discrete cosine transform (inverse DCT). Then, in the mode in which the motion compensation prediction is performed, the adder 107 adds the image signal subjected to the inverse DCT and the image signal obtained by the motion compensation prediction to restore the original image signal, and the frame memory Record at 108. In the intra picture mode, the inverse DCT unit 10
The image signal subjected to the inverse DCT in 6 is directly recorded in the frame memory 108.

When performing motion compensation prediction, the image signal recorded in the frame memory 108 is used. That is, based on the image signal recorded in the frame memory 108, the motion detection unit 10 detects the next image source signal.
At 9, motion detection is performed to detect a motion vector.
Further, based on the motion vector, the motion compensation unit 11
At 0, motion compensated prediction is performed. Then, the image signal obtained by this motion compensation prediction is output to the adder 101 via the switching circuit 111, and the difference between the input image source signals is obtained.

In this way, the data coded by the second coding unit 40 is output from the moving picture coding apparatus 1 as coded image data.

Next, the operation of the moving picture coding apparatus 1 will be described. The moving picture coding apparatus 1 includes an external device 2 (not shown).
Two program data are reproduced by one reproducing device and are inputted from the input terminals a and b, respectively. At that time, the program data which has already been input to the moving image encoding apparatus 1 is re-input from either one of the input terminals. Video coding device 1
The two program data input to the first program are switched by the input switcher 10, and the program data first input is the first program data.
The image data that has already been input and re-input to the second encoding unit 40 is input to the second encoding unit 40.

The program data input to the first coding unit 20 is sequentially DCT-converted, quantized at a predetermined fixed rate, and variable-length coded to generate an image bit stream. Then, the data generation amount for each image data is stored in the storage unit 30.

On the other hand, the program data input to the second encoding unit 40 is the data when the program data is first input to the moving image encoding apparatus 1 and encoded by the first encoding unit 20. Since the generation amount is stored in the storage unit 30, the encoding is performed based on the data generation amount. That is, the control unit 113 determines the quantization rate when the image data is quantized by the equations 3 and 4 based on the data generation amount. Then, the DCT-converted program data is quantized at the quantization rate, variable-length coded, and output. At that time, in the second encoding unit 40, the inverse quantization unit 105 to the motion compensation unit 110 perform interframe motion compensation prediction encoding.

When the coding in the first coding unit 20 and the second coding unit 40 is completed, the first coding unit 20
The program data that has been input to is reproduced again from the beginning, and this time is input to the second encoding unit 40. In addition, new program data is simultaneously reproduced and input to the first encoding unit 20 to detect the data generation amount. After that, by repeating this processing in sequence, a plurality of pieces of program data are sequentially variable-length coded.

As described above, in the moving picture coding apparatus 1 of the present embodiment, the first coding unit performs coding once to obtain the data generation amount of each image of the image data, and the data generation amount. A data amount is assigned to each image data based on the above, and the second encoding unit performs encoding. Therefore, it is possible to appropriately record the image data on a recording medium having a limited recording capacity, such as a DVD, after performing an appropriate variable rate encoding on each image data while maintaining the image quality. Furthermore, the encoding in the first encoding unit that detects the data generation amount for new image data and the encoding in the second encoding unit that generates the finally recorded image bit stream are performed simultaneously. The program data is sequentially encoded for each reproduction time of the program data. That is, when a plurality of program data are sequentially coded and recorded, they can be processed at high speed.

The present invention is not limited to this embodiment, but various modifications can be made.

For example, the configuration of the first coding section 20 is as follows.
Although it is preferable to simplify the configuration of the second encoding unit 40 and implement it, the simplification method is as in this embodiment.
Inverse quantizer 105 to motion compensator 1 of the second encoder 40
A method other than the configuration except 10 may be used. For example, the circuit configuration may be simplified by narrowing the motion vector search range performed by the motion detection unit 109 of the second encoding unit 40, and the first encoding unit 20 may be implemented. Usually, in the motion detection unit 109,
Although the motion vectors in the 64 horizontal / vertical directions are detected, if the search is performed within the range of ± 32 or ± 16, the motion detection unit 109 can be configured to be 1/4 or 1/16. It is reduced. The circuit thus simplified may be used as the first encoding unit 20.

The DCT unit 10 of the second encoding unit 40
2, the quantization unit 103, the inverse quantization unit 105, and the inverse D
The first encoding unit 20 may have a configuration in which the accuracy of the arithmetic unit such as the CT unit 106 is lowered and the arithmetic circuit is simplified.

The transfer of data between the storage unit 30 and the first encoding unit 20 and the second encoding unit 40 is performed by the first
The second encoding unit 40 stores only the detected data generation amount in the storage unit 30, and the second encoding unit 40 stores the detected data generation amount in the storage unit 3.
Only the data generation amount stored in 0 is read. Therefore, each external I / F 112, external I / F 13
2 may have a configuration in which only writing of data and reading of data can be performed, and each can be simplified.
You may make it such a structure.

Further, in the present embodiment, the process of allocating the encoded data amount and determining the encoding bit rate based on the data generation amount of each image data detected in the first encoding unit 20 is performed. This is performed in the control unit 113 of the second encoding unit 40. However, this processing may be performed by, for example, the control unit 133 of the first encoding unit 20, or may be performed by an arithmetic means other than the first encoding unit 20 and the second encoding unit 40. It may be performed.

In the present embodiment, the image data of the source to be encoded is the image data recorded on the VTR tape. However, in addition to this, a hard magnetic disk (HD), a magneto-optical disk (MO) is used. Any media, such as a digital video disc (DVD), may be used. Further, the moving picture coding apparatus 1 may have a structure including a reproducing apparatus for such a medium.

The encoding method in the encoding unit 40 is also the same as that shown in FIG. 2 in the present embodiment.
However, the encoding method is not limited to this. For example, a method using a transform coding method other than DCT may be used.

In the case where the second encoding unit 40 is implemented by a modularized circuit such as an LSI, it is costly to implement the same in the first encoding unit 20 as well. If there is an advantage in terms of the above, such a configuration may be adopted.

[0049]

According to the moving picture coding apparatus of the present invention, continuous image data can be appropriately variable rate coded with a predetermined capacity, and further, a plurality of image data can be sequentially variable rate coded. When you do, faster than ever,
That is, the encoding can be performed in a time corresponding to one reproduction time. As a result, various image data are recorded on the DVD.
When attempting to perform compression recording on a recording medium such as, the recording can be sequentially performed at high speed, thus improving work efficiency.

[Brief description of drawings]

FIG. 1 is a block diagram showing the configuration of a moving picture coding apparatus according to an embodiment of the present invention.

2 is a block diagram showing a configuration of a first encoding unit of the moving image encoding apparatus shown in FIG.

3 is a block diagram showing a configuration of a second encoding unit of the moving image encoding apparatus shown in FIG.

FIG. 4 is a diagram showing bit rates of fixed rate coding and variable rate coding.

FIG. 5 is a diagram illustrating a method of determining a bit rate in variable rate coding.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 1 ... Moving image coding apparatus 10 ... Input switching device 20 ... 1st coding part 122 ... DCT part 123 ... Quantization part 124 ... Variable length coding part 132 ... External I / F 133 ... Control part 30 ... Storage part 40 ... 2nd coding part 101 ... Adder 102 ... DCT part 103 ... Quantization part 104 ... Variable length coding part 105 ... Dequantization part 106 ... Inverse DCT part 107 ... Adder 108 ... Frame memory 109 ... Motion Detection unit 110 ... Motion compensation unit 111 ... Switching circuit 112 ... External I / F 113 ... Control unit

Claims (5)

[Claims] 1. A first encoding means for variable rate encoding continuous image data, and data of each image data of the continuous image data variable rate encoded by the first encoding means. A data amount determining means for allocating the encoded data amount to each image data based on the generation amount and the limit value of the encoded data amount for the entire continuous image data, and each image of the continuous image data Second encoding means for variable rate encoding the continuous image data so that the data generation amount of the data is equal to or less than the encoded data amount assigned to each image data by the data amount determining means; The continuous image data to be newly encoded is input to the first encoding means, encoded by the first encoding means, and encoded by each of the image data by the data amount determining means. Continuous image data data amount allocation has been completed,
An image coding apparatus, comprising: image data input means for inputting new continuous image data to the first coding means at the same time as inputting to the second coding means. 2. The image coding apparatus according to claim 1, wherein the first coding means has a configuration in which the second coding means is simplified. 3. The second encoding means compresses continuous image data, quantizing means for quantizing the compressed signal, and variable length code for the quantized signal. Variable-length coding means for dequantizing, dequantizing means for dequantizing the quantized signal, decompressing means for decompressing the dequantized signal, and an image interval based on the decompressed signal. Motion detection means for detecting the motion of the image, motion compensation means for performing motion compensation based on the detected motion, and each image of the continuous image data used for the compression based on the motion-compensated signal Arithmetic means for obtaining a prediction error of data, the first encoding means, the dequantizing means of the second encoding means, the decompressing means, the motion detecting means, the motion compensating means, And in a configuration that does not substantially include the arithmetic means The image coding apparatus according to claim 2. 4. The first encoding means and the second encoding means compress a signal based on each image data of continuous image data, and quantize the compressed signal. Quantization means, variable length coding means for variable length coding the quantized signal, dequantization means for dequantizing the quantized signal, and decompressing the dequantized signal Decompressing means, motion detecting means for detecting a motion vector by searching a predetermined range for each pixel forming the image data based on the decompressed signal, and motion compensation based on the detected motion vector. Encoding means for calculating a prediction error of each image data of the continuous image data used for the compression based on the motion-compensated signal, 1 encoding means The motion detecting means includes the second
3. The image coding apparatus according to claim 2, wherein the motion vector is detected by searching a range substantially narrower than a search range of the motion vector in the motion detection unit of the coding unit. 5. The first encoding means and the second encoding means compress a signal based on each image data of continuous image data, and quantize the compressed signal. Quantization means, variable length coding means for variable length coding the quantized signal, dequantization means for dequantizing the quantized signal, and decompressing the dequantized signal Decompressing means, motion detecting means for detecting motion between images based on the decompressed signal, motion compensating means for compensating motion based on the detected motion, and motion compensating signal based on the motion compensated signal. Encoding means for obtaining a prediction error of each image data of the continuous image data used for the compression, which is any one of the respective means of the first encoding means, or In any plurality or all The image coding apparatus according to claim 2, wherein the calculation to be performed is configured to be easily performed with a lower accuracy than the calculation in each corresponding unit of the second encoding unit.