JP3360942B2 - Video encoding device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION [0001] BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to transmission of moving images and the like.
The present invention relates to a moving picture coding apparatus to be used. [0002] 2. Description of the Related Art Conventionally, techniques in such a field include:
For example, there is one described in the following literature.     Reference 1: Optronics (1992-5) Optronics, Okubo           “Encoding in Videophone / Conference” P.74-79,86-98     Reference 2: IEEE 1993 CUSTOM CIRCUITS CONFERENCE (1993),             Subroto Bose et al. “A Single Chip Multistandard              Video Codec ”P.11.4.1-11.4.4 In the coding of moving images, as described in Reference 2,
An image of one frame to be obtained is, for example, 16 × 16 pixels.
Small block called a macro block (hereinafter referred to as MB)
The following processes (1) to (5) are sequentially performed for each lock,
Perform encoding. (1) Motion vector detection The frame to be encoded, that is, the code of the encoded frame
The region having the largest correlation with the MB to be encoded
Detect from the illumination frame. The encoded reference frame at this time
Frames are generally used before and after the encoded frame.
Can be In the region detection, in the coded reference frame
The vicinity of the same spatial position as the MB to be encoded is a search area.
The largest correlation of the MB to be coded in the search area.
A critical area is required. And the area with the largest correlation
Position in the reference frame and the spatial position of the MB to be encoded
The difference from the position is detected as a motion vector. (2) Discrete cosine transform (hereinafter referred to as DCT) A kind of orthogonal transformation is performed on the MB to be coded, and
MB data is frequency spectrum data from spatial data
Is converted to The input data of DCT is
Whether the encoding target MB data of the
Or the region to be referred to by the encoding target MB and the motion vector
Of the reference frame is used. That judgment
As a standard, when the difference value is large, the encoding
Frame data is used as is
Data is used. (3) Quantization DCT output data is converted to a predetermined representative value. (4) Variable length coding Larger for less frequently occurring values of quantization output data
Is given a code of a certain length, and for frequently occurring values
Codes of small length are allocated and encoded
Is output. (5) Inverse quantization, inverse discrete cosine transform (hereinafter IDCT)
That) Inverse quantization is the inverse of quantization, and IDCT is the inverse of DCT.
Perform each processing. With these processes, spatial data
The image data is played back as
This is the reference frame data for the system. Where quantization
Is not a 1: 1 transform, so quantization input and inverse quantization
Not equivalent to output. Therefore, the original image and the reproduced image
Will be different. The motion vector detection in (1) is bright
Is performed on the degree component data, and the processing of (2) to (5)
Performed on the luminance component and the color component. Brightness in 1MB
The component is 16 × 16 = 256 pixels, and the color component is half of 1
There are 28 pixels. FIG. 2 shows an example of a conventional moving picture coding apparatus.
FIG. 2 is a block diagram showing the configuration, with reference to FIG.
The flow of the encoding process for one MB data will be described.
You. This video encoding apparatus 100 encodes an encoded MB.
Image data to store image data necessary for
Buffer 101, and the image data buffer 101
On the output side, a motion vector detection unit 102 and a DCT unit 103
Is connected. Output side of the motion vector detection unit 102
Is connected to the control unit 104. Out of DCT103
The IDCT unit 105 and the quantization unit 106 are connected to the input side.
The output data of the IDCT unit 105 is
The connection is output to the buffer 101. Quantization unit 106
Are output to the inverse quantization unit 107 and the variable-length encoding unit 108
Is connected. The output signal of the inverse quantization unit 107 is also ID
Input to the CT unit 105 and output from the variable length coding unit 108
Is the encoded data DT_outIs output.
The control unit 104 controls the encoding of the moving image encoding device 100.
It controls the processing sequence, and includes a control signal s1
04a, s104b, and s104c
ing. The control signal s104a from the control unit 104 is DC
T section 103, IDCT section 105, quantization section 106, inverse amount
Input to the child unit 107 and the variable length coding unit 108,
The control signal s104b is input to the image data buffer 101
It is a configuration that is performed. The control signal s104c is
The connection is supplied to an image memory 150 (not shown).
The motion vector detection unit 102 includes the image data buffer 101
Input the data required for motion vector detection from
The motion vector d102 of the result of the vector detection is
Output to DCT103 is the frame data to be encoded
Reference of (current image data) and position indicated by motion vector
Frame data (reference image data) and image data
File 101 and the current image as DCT input data
Use the data or use the current image data and reference image data
To use the data.DCT section103 is the judgment result
The result is a selection result s103 of intra / interframe prediction.
Output to the IDCT unit 105 and perform DCT.
DCT result data d103 is quantized106Send to
You. [0005] The quantization unit 106 is a DCT.output datad10
3 is input and quantized, and converted to a representative value by this quantization.
The supplied data d106 is supplied to the variable length coding unit 108.
It is. The variable length coding unit 108 is a quantization unit106De from
Assign a code corresponding to the frequency of occurrence to
Coded data DT_outIs generated and output. Inverse quantization unit 1
07 performs inverse quantization on the quantized data d106,
The data d107 of the inverse quantization result is supplied to the IDCT unit 105.
Pay. In the IDCT unit 105, the data d107
IDCT is performed, and the
The intra-frame prediction selection result s103 indicates intra-frame prediction.
If selected, the IDCT unit 105 outputs the IDCT result
The image data back as it is as the reproduced image data d105.
Output to the camera 101. In the DCT section 103,
When the inter-frame prediction is selected, the IDCT unit 10
Reference numeral 5 denotes input of reference image data from the image data buffer 101.
To add to the IDCT result. Reference image in IDCT result
The data to which the image data has been added is the reproduced image data d105.
Is output to the image data buffer 101. Image
Data buffer101Stores the playback image data.
Played image data storedIs outsideOutput to the image memory
You. The control unit 104 includes a motion vector detection unit 102 and a DC
T unit 103, IDCT unit 105,Quantization unit106
And the inverse quantization unit 107 and the variable length encoding unit 108
The encoding control signal s104a is transmitted, and the encoding process is performed.
Control the cans. The content of the encoding control signal s104a is
It consists of an address and data.
The specified part performs the calculation of the encoding process. Coding
The data part of the control signal s104a is the specified part
Parameters to be given to In addition, the control unit 104
Buffer for controlling input / output to / from data buffer 101
Sends the control signal s104b to the image data buffer 101.
External image memory 150 and image data buffer 10
Image memory control for controlling data input / output between one
The signal s104c is sent to the external image memory 150. FIG. 3 shows an image data buffer shown in FIG.10
1FIG. 4 is a diagram for explaining data transfer. The following (i) to (ii)
i), external image memory 1 required for 1 MB encoding process
The figure shows the data transfer amount between the image data buffer 50 and the image data buffer 101.
This will be described with reference to FIG. (I) Reference image data read Reference to perform motion vector detection for luminance component
Images for the area used as the search range in the frame
Data must be read from the image memory 150. example
For example, ± 16 pixels in the horizontal direction and the vertical
If the search range is ± 16 pixels in the direction, an image of 48 × 48 pixels
It is necessary to read out image data. But usually the spatial position
Encoding is performed according to the order of
Part overlapping the search range between the MB and the MB processed immediately before
Will exist. Image data of this overlapping part
Is already stored in the image data buffer 101
In the motion vector detection for the encoding target MB,
If only the data to be used is read, as shown in FIG.
× 48 pixels are external image memory150Read from
You. On the other hand, for the color components,
Since the pixels are read out in the same manner, they are 16 × 8 pixels. (Ii) Current image data 16 × 16 pixels for luminance component and 16 × 8 color component
PixelButIs read. (iii) Playback image data The luminance component is 16 × 16 pixels and the color component is 16 × 8 pixels.
When the above (i) to (iii) are summed up, (16 × 48 + 16 × 8) + (16 × 16 + 16 × 8)
+ (16 × 16 + 16 × 8) = 1664 pixels Becomes FIG. 4 is a block diagram showing the moving picture coding apparatus and the picture memory shown in FIG.
FIG. FIG. 4 shows the moving picture coding of FIG.
An image memory 150 connected to the device 100 is shown.
You. Input data DT of the image memory 150_inIs in chronological order
The input moving image data has been converted from analog to digital
The data is stored in the image memory 150.image dataD
T_inIs written. Control unit in video encoding device 100
104 outputs an image memory control signal S104c;
Based on the image memory control signal s104c, image data
d150 is the moving picture coding apparatus 100 and the image memory 150
Transferred between and. Refer to the contents of image data d150
It consists of image data, current image data, and reproduced image data.
It is. [0008] SUMMARY OF THE INVENTION However, the conventional
The moving image coding apparatus has the following problems. processing
When the size of the image increases, the moving image encoding apparatus 100
Processing capacity, video encoding device 100 and image memory 15
When the data transfer speed between 0 is limitedUThere is a problem
Was. For example, in the standard video standard, the size of one frame is
720 x 480 pixels, frame frequency 30 Hz
It is. The number of MBs in one frame is 720 × 480 / (16 × 16) = 1350 MB And the maximum allowable processing time for 1 MB is 1/30 sec / 1350 = 14.6 μs And in a very short time. As a result, the processing
The performance is not satisfied with one video encoding device
Also occurs. In addition, the moving image encoding device 100 and the image memo
Cycle per pixel in data transfer between memory 150
Is 24.6 μs / 1664 pixels = 14.78 ns Becomes 4 pixels to increase transfer capacity per pixel
Even if they are transferred at the same time, they are transferred at a period of 59.1 ns
There is a need. Consider transfer overhead at about 25%
And data at a period of 47.3 ns, that is, 21.1 MHz.
I need to input and output. Code in video coding device
Cannot satisfy either the image processing performance or the image transfer performance.
When the image data is divided and processed.
A system with an appropriate amount of encoding calculation and image data transfer
Is described in reference 2. FIG. 5 shows a moving image obtained by dividing image data.
It is a block diagram which shows an encoding device. This image coding device
The moving image encoding apparatus shown in FIG.100When
Two encoding processing units 100A having the same configuration,100BTo
Have. Each of the encoding units 100A and 100B includes:
Image memories 150A and 150B are connected respectively.
You. 720 images are stored in the two image memories 150A and 150B.
× 240 pixel image data DT_inAre written respectively
You. eachEncoding processing unit100A, 100B and image memory 1
Between the data d150A and d150B between 50A and 150B.
Each transfer is performed, and each of the encoding processing units 100A, 10A
0B respectively perform an encoding process. Figure 6 shows the frame
3 is a diagram showing an MB of FIG. FIG. 6 shows a part of one frame.
A0 to G4 of a plurality of MBs are shown. For example, M
Image data DT corresponding to A0 to D4 of B_inIs an image memo
And the encoding processing unit 100A writes the MB
A0 to D4, that is, a code corresponding to the upper MB of the frame.
Perform encryption. Similarly, it corresponds to E0 to G4 of MB.
Image data DT_inIs written to the image memory 150B,
The encryption processing means 100B determines that the MBs E0 to G4,
Of theLowerIs performed corresponding to the MB. This
According to such an encoding process, a negative
Since the load is divided, the concentration of processing volume can be avoided.
You. However, the MB at the top and bottom border of the frame
The motion vector detection range cannot be secured enough
Motion vector detection performance is degraded and encoding efficiency is reduced.
I was letting it down. For example, for the upper MB D2
For motion vector detection, it corresponds to the lower MB of E0 to E4
Data could not be referenced. [0010] The present invention solves the above problems.
To decideDivide current image data into multiple block rows
And a plurality of blocks corresponding to each divided block
A moving image encoding device including an encoding processing unit,
Each encoding unit is connected to the image memory,
Image data temporary input / output controlled by
Synchronization with the image data buffer for storage and the encoding processing means
It has a control means and a control means. The encoding process
The means is connected to the image data buffer, and includes an encoding system.
The sequence of the encoding process is controlled by the control signal,
Location in the current image stored in the image data buffer
Image data of a block of a fixed size and the reference block
Perform motion vector detection using the image data of
Encodes the current image in block units and encodes the encoded data.
Output. The synchronous control means includes a synchronous control
When an interface signal is input, the image memory
Encodes the synchronization control signal for access right arbitration
And the synchronization control signal from the other encoding processing unit.
Receiving the synchronization control interface signal.
Things. The control unit is configured to control the buffer control signal
And a function of outputting the encoding control signal.
Automatically by transmitting and receiving the synchronous control interface signal.
Own encoding processing unit has access right to the image memory
Is determined. And before
The control means has an access right to the image memory.
Place where you are Output an image memory control signal to
Access the image memory and use it in the motion vector detection
Common among the image data of the blocks of the reference image
Read the image data to be used from the image memory and
The image data buffer and the image in the other encoding processing unit.
The data is stored in the image data buffer. [0011] According to the present invention, as described above,
Each encoding processing unit controls each control means.
Block rows in the current image adjacent to each other
To be encoded. At this time,A control hand in an encoding unit
The stage receives a synchronization control interface signal from the synchronization control means.
Upon receipt, it outputs an image memory control signal to
Access and block the reference image used for motion vector detection.
Lock image data shared with other encoding processing units
Read from the image memory
Image data buffer and other encoding in the encoding processing unit
The image data is stored in the image data buffer in the processing unit. Reference image
Of the image data of the block
Unused block image data and current image block image
Image data means that each encoding processing unit
Read from the memory and stored in the image data buffer. this
Thus, each encoding processing unit records the data in the image data buffer.
Using the stored image data, a motion vector is detected and
Encoding is performed in lock units.Therefore, the above problem can be solved.
You can. [0012] FIG. 1 is a moving picture coding system showing an embodiment of the present invention.
It is a block diagram of an apparatus. This moving picture coding apparatus 200
Are two components configured using a CPU (central processing unit)
Are provided with encoding processing sections 200A and 200B of
The image processing units 200A and 200B input the images input in time series.
Image data DT_inIs connected to the image memory 250 for storing
ing. Each of the encoding processing units 200A and 200B
Perform motion vector detection for each block and perform encoding
Each encoding processing unit 200A, 200
B and data transfer between the image memory 250 and encoding
Data encoded from the output side of the processing units 200A and 200B
Data DT_outIs outputIt has become.Each encoding
Processing units 200A and 200B share image memory 250
And the image memory 2 of each of the encoding units 200A and 200B.
Against 50Mediation of access rights (ie, access conflicts)
Avoidance arbitration) is performed by the synchronization control signal s200.
You.FIG. 7 is a block diagram showing the inside of the encoding processing unit in FIG.
It is a lock figure. The encoding units 200A and 200B are
And the image data buffer 201 is
Have. The image data buffer 201 is the image memory 2
50, and each of the encoding processing units 200A, 20A
Function to store image data necessary for encoding in 0B
have.The image data buffer 201 has a motion database.
Vector detection unit 202, DCT unit 203, IDCT unit 20
5, quantization unit 206, inverse quantization unit 207, and variable length code
Encoding means comprising an encoding unit 208 is connected.
You. That is,The output side of the image data buffer 201
Vector detection unit 202 and DCT unit 203 are connected
I have. On the output side of the motion vector detection unit 202,,Control hand
The control unit 204, which is a stage, is connected.DCT section20
3 is connected to the IDCT unit 205 and the quantization unit 206.
The IDCT unit 205Reproduction image data d20
5Is a connection output to the image data buffer 201.
You. On the output side of the child unit 206, an inverse quantization unit 207 is
The variable length coding unit 208 is connected. Inverse quantization unit 20
7 output datad207Is also input to the IDCT unit 205.
The output of the variable length coding unit 208 is the coded data DT
_outIs output. The control unit 204
Control the encoding sequence of the encoding unit
And the control signals s204a, s204b, s204
It has a function to send out c. Control unit204System from
The control signal s204a is transmitted to the DCT unit 203 and the IDCT unit 20.
5, quantization unit 206, inverse quantization unit 207, and variable length code
The control signal s204b is input to the
This is a configuration input to the data buffer 201. Control signal
s204cIs a connection supplied to the image memory 250.
You. Each of the encoding units 200A and 200B has, Synchronous control
Synchronous control means for outputting the signal s200Sync
Control unit220Are respectively provided, and the synchronization control unit is provided.22
0Is connected to the control unit 204. Next, the operation of the moving picture coding apparatus shown in FIG. 1 will be described.
This will be described with reference to FIG. Encoding processing units 200A, 200B
Share the image memory 250 and access in time divisionToline
, Respectively. Synchronization control signal s200
Which encoding processing unit200A, 200BIs an image memo
It is determined whether to input / output data to / from the memory 250. Motion vector
The torque detection unit 202 moves from the image data buffer 201.
Enter the data required for vector detection and
The motion vector data d202 of the output result is stored in the control unit 204.
Output toDCT section203 is the current image data and motion vector
The reference image data at the position indicated by the
Input from buffer 201 and present image as DCT input data
Use image data or current image data and reference image
Use dataTojudge.DCT section203 is
The judgment result is the selection result s2 of intra-frame / inter-frame prediction.
03 to the IDCT unit 205 and DCT
The DCT result data d203 is sent to the quantization unit 206.
Send out. The quantization unit 206Output data of DCT section 203
TaQuantization is performed by inputting d203.
The data d206 converted into the value is used as the variable length coding unit 208
Supplied to The variable length coding unit 208 is a quantization unit206
A code corresponding to the frequency of occurrence of data d206 from
Coded data DT_outIs generated and output. Inverse quantum
The quantization unit 207 performs inverse quantization on the quantized data d206.
And the data d207 of the inverse quantization result is stored in the IDCT unit 20.
5 The IDCT unit 205 converts the data d207
IDCT is performed on the frame, and the frame in DCT section 203 is subjected to IDCT.
Selection result s203 of intra-frame / inter-frame prediction is within a frame
When prediction is selected, the IDCT unit 205
The image data is directly used as the reproduced image data d205.
Output to the data buffer 201. Also, the DCT unit 203
If the inter-frame prediction is selected in the
The unit 205 stores the reference image data from the image data buffer 201.
And adds it to the IDCT result. IDCT results
The data to which the reference image data has been added is the reproduced image data d.
205 is output to the image data buffer 201.
The image data buffer stores the reproduced image data,
The stored reproduced image data is displayed as image data d250.
Output to the image memory 250. The control unit 204 includes a motion vector detection unit 202
, DCT section 203, IDCT section 205,Quantization unit
206, an inverse quantization unit 207, and a variable length encoding unit 208
At the same time, the encoding control signal s204a is transmitted, and the encoding process is performed.
Control sequence. Encoding control signal s204a
The content consists of an address and data, and the address
The part specified by performs the calculation of the encoding process.
The data of the encoding control signal s204a is,Identified
Parameters to be given to the part. The control unit 204
Is,Controls input / output to / from image data buffer 201
BufferControl signal s204bTo the image data buffer 2
To 01Send it out,External image memory 250 and image
Controls data input / output between data buffers 201
Image memoryControl signals204c is an external image memory
250WhatSend out. This image memoryControl signals204c
Each of the encoding processing units 200A and 200B performs processing.
Block to be selected. Each encoding processing unit 20 of the present embodiment
0A and 200B are adjacent blocks in the current image.
Encode each column in parallel. For example, the encoding processing unit
200A is,A0 to A4, C0 to C4, and E in FIG.
Encoding is performed on the MBs 0 to E4, and the encoding processing unit 2
00B is the MB of B0 to B4, D0 to D4, and F0 to F4
Is encoded. MB A0 to A4 and B0 to B0
B4, each C0-C4 and each D0-D4, and each E0-E4
And the processing of F0 to F4 is slightly shifted in timing.
, And are encoded simultaneously. FIG. 8 is a data transfer diagram of the encoding processing unit shown in FIG.
FIG. This figure shows MBs C1 and D1
Is shown. Each encoding processing unit 2
00A, 200BToOf the processing time
The detection uses about 1/3 and DCT and quantization etc.
2/3 are used. Reference data of luminance component for C1
If the data is ± 16 pixels in both the horizontal and vertical directions,
The reference image MBs to be processed are B0, B1, B2, C0,
C1, C2, D0, D1, D2
B0, B1, C0, C1, D
Since 0 and D1 are used, the image data buffer
Is stored in the key 201. Therefore, newly required M
The image data of B is the image data of B2, C2, and D2.
You. Similarly, newly required image data for D1 is
Image data of MBs C2, D2 and E2. here,
The image data of C2 and D2 is,Each encoding processing unit 200A,
Since it is necessary to detect both 200B motion vectors,rear
As described above, for example, the encoding processing unit 200 on the master side
A image memory control signal s output from the control unit 204 of A
204c, the code read from the image memory 250
Data buffer of both the image processing units 200A and 200B
Written in 201. The reading of the luminance data of the reference image is performed as shown in FIG.
Time t₀And then read out the current image data
Is performed. In reading the current image data, the current image of C1 is read.
Data is read out first, D1 current image data is read out later
Is done. After reading the current image data of C1, the encoding process
The control unit 200A starts detecting a motion vector for C1.
You. After reading the current image data of D1, the encoding processing unit
200B starts motion vector detection for D1.
Also, after reading of the current image data of D1 is completed,
C stored in each image data buffer 201
0 and D0 reproduced image data in the image memory 250 in order.
Write. At the end of writing the reproduced image data of D0, C
Since the motion vector detection for 1 has been completed,
Of the reference area of the color component indicated by the
The data C1c is an image data buffer of the encoding processing unit 200A.
Written in the key 201. Subsequently, the color component for D1
The data D1c in the reference area is read out and
0B is written to the image data buffer 201. Two
In the encoding process for C1 and D1 of the MB, the image memory 2
50 and the data transfer amount between each image data buffer 201
Is 281 between time t1 and time t2 shown in FIG.
This is 6 pixel data, which is 14 per MB.
It becomes data of 08 pixels. This value is equivalent to the data
It is reduced to 11/13 for the transmission amount of 1664 pixels.
As a result, the processing time can be reduced. FIG. 9 shows the access between the encoding units in FIG.
FIG. 8 is a diagram showing the change in
Each of the encoding processing units 200A from time t1 to time t2,
The access to the 200B image memory 250 will be described.
You. One of the encoding units 200A and 200B is
And the other becomes a slave. 3 or more encoding units
In this case, one encoding unit becomes the master and
All the encryption processing units become slaves. Here, the encoding process
The processing unit 200A is the master, and the encoding unit 200B is the slave.
And The encoding processing unit 200A,200BBoth
Data required for communication from the image memory 250
At this time, the image memory control signal s204c is
Output from the encryption processing unit 200A and output from the slave side.
I'm not forced.Thus, the slave-side encoding processing unit 2
In the control unit 204 of 00B, an image for reading data is
Since there is no need to output the memory control signal s204c,
Processing efficiency can be improved.Which encoding processing unit200
A, 200BIs a master or a slave,System
Synchronizes the synchronization control interface signal sI from the control unit 204
controlDepartment220 and set. At time t1,
The encryption processing unit 200A accesses the image memory 250.
Abandon the right. Control unit 20 in encoding processing unit 200A
4, the synchronization control interface signal sI
The synchronization control unit 220 supplies the synchronization control
From the section 220 to the synchronization control section 22 in the encoding section 200B
The synchronization control signal s200 is sent to 0. Encoding processing unit
From the synchronization control unit 220 in the synchronization control interface 200B.
Of the control unit 20 in the encoding processing unit 200B.
4 and the control unit 204 stores it in the image memory 250.
Recognize that you have access to the Control unit 204
Is an image for reading the current image data of MB D1.
Outputs image memory control signal S204c to image memory 250
Then, the corresponding data is read from the image memory 250.
You. After reading the current image data of D1, the codeConversionProcessing unit
The control unit 204 of the 200B
Give up access rights. That is, the signConversionProcessing unit 200B
Your part 204,Signs synchronous control interface signal sI
ConversionIt is supplied to the synchronization control unit 220 of the processing unit 200B. Sign
ConversionFrom the synchronization control unit 220 of the processing unit 200B to the encoding processing unit
200A synchronization control unit 220What,Synchronization control signal s200
Is sent. In the same procedure, the encoding processing unit 200A
Of the reproduction image data of C0 by the encoding unit 2
Writing and encoding of the reproduced image data of D0 by 00B
C1c color component C1c reference image data by the processor 200A
The color component D1 of D1 by the reading and encoding processing unit 200B
c Reference image data reading and encoding processing unit 200A
The reference image data of the luminance component of C3 is read.
You. The luminance component that is the reference image data of MB C3 is,Picture
Read from image memory 250EachEncoding processing unit 200
A and 200B in the image data buffer 201 respectively.
Be included. Here, the writing procedure of the reference image data of C3
Will be described. Control of coding processing section 200A as master
The unit 204 synchronizes the common data access request signal
Control interfacesignalEncoding unit 200 as sI
A to the synchronization control unit 220 of FIG. This synchronization control unit 22
0 to the synchronization control unit 220 of the encoding processing unit 200B
The synchronization control signal s200 is sent out. Synchronous control signal s
After a predetermined cycle delay from the output of the image memory 200, the image memory 2
Reading from 50 is started. Predetermined
The number of cycles is determined by the two encoding processing units 200A and 200A.
00B is set in advance. SignConversionOf the processing unit 200B
The control unit 204,The common data address via the synchronization control unit 220
Access signal. Simultaneously with the passage of the predetermined cycle
Then, data transfer to the image data buffer 201 is started.
Begun. Control unit of encoding processing unit 200A as master
Reference numeral 204 denotes an image memory control signal s204c
250 and a buffer control signal s204b
Is supplied to the image data buffer 201. Meanwhile, coding
The control unit 204 of the processing unit 200B,Buffer control signal s
204b to the image data buffer 201
Do just that. Luminance component which is reference image data of MB C3
After the reading of the reference image, the luminance component of D3 is
It is read out as image data. Subsequently, the encoding processing unit 2
E3 reference image data of MB according to 00B, encoding process
Each reading of current image data of C2 of MB by unit 200A
Is performed. As described above, in this embodiment, the image memory
Two encoding processing units 200A and 200 sharing the same 250
B constitutes a moving picture encoding apparatus,
There is no center, and the load on the encoding processing units 200A and 200B is low.
Is reduced. Also, each of the encoding units 200A, 200B
Is,Parallel MB rows adjacent to each other in the current image
To be encoded. Therefore, the area of the reference image is limited.
Use common reference image data
Function, preventing the image quality from deteriorating.AboveImage memory 250
The amount of data transfer between the video encoding device 200 and the
I can do it.In particular, the encoding processing units 200A, 200B
The data that is required in common for both is stored in the image memory 250.
When reading from the master, the image memory control signal s204c is
Is output from the encoding processing unit 200A
Is not output. Thereby, the coding on the slave side is performed.
The control unit 204 of the processing unit 200B reads the data
It is not necessary to output the image memory control signal s204c for
Therefore, processing efficiency can be improved.Note that the present invention
The present invention is not limited to the embodiment, and various modifications are possible.Its deformation
As an example, for example, the following (a) and (b)
is there. (a)   The number of encoding processing units 200A and 200B is two.
Or three or more, in which case
In addition, the burden on each encoding unit is reduced.example
For example, the encoding unit is provided in parallel with three columns, and the current image is divided into three parts.
And read and process (ie, three encoding units
To process C2 / D2 / E2 of MB shown in FIG. 6 in parallel.
The C3, D3, and E3 image data
It is used for processing in the encoding processing unit, and is almost the same as the above embodiment.
Works in the same way, reducing the burden on each encoding unit.
You. (b)   MB size can be changed according to the purpose of image processing
And the above-mentioned operation is performed for any size MB.
The same effect as the example is exerted. [0020] As described in detail above, according to the present invention,
If multiple image processing units share the image memory,Niso
A synchronization control means and a control means are provided respectively, and control of the control means is provided.
Control, each encoding processing unit
Encode columns in parallel, and encode
Image memory control signal output from the control means in the
Than,Commonly used reference image dataImage memory
Image from each encoding unit that uses the
The data is stored in the data buffer.That
Because the area of the reference image is not limited,
Image data can be used, and image memory andDuplicate
Between the number encoding unitReduce the amount of data transferred
Can be.In particular, in the encoding unit that does not have access rights
In the control means, the data of the commonly used reference image is
Outputs image memory control signal for reading from image memory
It is not necessary to perform
You.

BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a block diagram of a moving picture coding apparatus according to an embodiment of the present invention. FIG. 2 is a configuration block diagram illustrating an example of a conventional video encoding device. FIG. 3 is a diagram for explaining data transfer of an image data buffer in FIG. 2; FIG. 4 is a block diagram showing a moving picture coding apparatus and a picture memory of FIG. 2; FIG. 5 is a block diagram illustrating a moving image encoding device that divides image data. FIG. 6 is a diagram showing MBs in a frame. FIG. 7 is a configuration block diagram showing the inside of an encoding processing unit in FIG. 1; FIG. 8 is a diagram illustrating data transfer of the encoding processing unit in FIG. 1; FIG. 9 is a diagram showing a change in access right between encoding processing units in FIG. 1; [Description of Codes] 200 Video coding devices 200A, 200B Coding processing unit 250 Image memory s200 Synchronization control signal s204c Image memory control signal s250 Transfer data DT _out Encoded data 201 Image data buffer 202 Motion vector detection unit 203 DCT unit 204 control unit 220 synchronization control unit s204b buffer control signal sI synchronization control interface signal

Claims (1)

(57) [Claims] 1. The current image data is divided into a plurality of block strings.
And a plurality of blocks corresponding to each divided block
A moving image encoding device including an encoding processing unit , wherein each of the encoding processing units includes a part in the current image.
Image data of block of fixed size and reference block
Shared image that stores reference image data including
Are connected to the image memory, each encoding unit is connected to the image memory, input by the buffer control signal
Image data buffer for temporarily storing image data whose output is controlled
Buffer and the image data buffer, and
The sequence of the encoding process is controlled by the
A predetermined size of the current image stored in the data buffer.
Image data of the reference block and the image data of the reference block.
And perform motion vector detection using the
Encode in blocks and output encoded data
Encoding processing means, when a synchronization control interface signal is input, the image
Synchronous control signals for arbitrating access rights to memory
Given to an encoding processing section, and transmitted from the other encoding processing section to the same
Receiving the synchronization control interface signal,
Synchronization control means for outputting the buffer control signal and the encoding control signal.
And a synchronous control interface signal.
The self-encoding processing unit is provided with the image memory
To determine whether or not you have access to
Means, wherein the control means sets an access right to the image memory.
If it has, output the image memory control signal and
Access image memory and use for motion vector detection
Common among the image data of the blocks of the reference image
Image data to be read from the image memory
In the image data buffer and the other encoding processing unit
Moving image characterized by being stored in an image data buffer
Image coding device.