CN101304520A - Image decoding system and self-adapting fetching-rapidly method for motion compensation thereof - Google Patents
Image decoding system and self-adapting fetching-rapidly method for motion compensation thereof Download PDFInfo
- Publication number
- CN101304520A CN101304520A CN 200710097481 CN200710097481A CN101304520A CN 101304520 A CN101304520 A CN 101304520A CN 200710097481 CN200710097481 CN 200710097481 CN 200710097481 A CN200710097481 A CN 200710097481A CN 101304520 A CN101304520 A CN 101304520A
- Authority
- CN
- China
- Prior art keywords
- macro block
- fast delivery
- delivery formula
- motion level
- motion
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F12/00—Accessing, addressing or allocating within memory systems or architectures
- G06F12/02—Addressing or allocation; Relocation
- G06F12/08—Addressing or allocation; Relocation in hierarchically structured memory systems, e.g. virtual memory systems
- G06F12/0802—Addressing of a memory level in which the access to the desired data or data block requires associative addressing means, e.g. caches
- G06F12/0875—Addressing of a memory level in which the access to the desired data or data block requires associative addressing means, e.g. caches with dedicated cache, e.g. instruction or stack
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F12/00—Accessing, addressing or allocating within memory systems or architectures
- G06F12/02—Addressing or allocation; Relocation
- G06F12/08—Addressing or allocation; Relocation in hierarchically structured memory systems, e.g. virtual memory systems
- G06F12/0802—Addressing of a memory level in which the access to the desired data or data block requires associative addressing means, e.g. caches
- G06F12/0862—Addressing of a memory level in which the access to the desired data or data block requires associative addressing means, e.g. caches with prefetch
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Compression Or Coding Systems Of Tv Signals (AREA)
Abstract
The invention relates to a self-adaptable fast-taking method for motion compensation, which is suitable for an image decoding system. The method comprises the steps: firstly, solving a macro-block and a motion vector between the macro-block and the following macro-block; secondly, determining a fast-taking mode according to the motion vector, wherein, the fast-taking mode corresponds to a fast-taking scope on a reference picture frame, the image data of which is pre-solved and stored in a DRAM (dynamic random access memory); thirdly, accessing the dram in order to store the image data in the fast-taking scope to a high speed cache; then, accessing the DRAM so as to store the image data of the fast-taking scope in a high-speed cache, and then accessing the high-speed cache in order to solve whether or not the reference area needed by the following macro-block is located within the fast-taking scope; finally, determining whether the fast-taking model is to be kept or not based on the result obtained from the former step that whether or not the reference area needed by the following macro-block is located within the fast-taking scope.
Description
Technical field
The present invention relates to a kind of image-decoding system, refer to a kind of self-adapting fetching-rapidly method of image-decoding system motion compensation especially.
Background technology
At present, digital video product is grown up fast, and its application has comprised categories such as image communication, guard monitor and entertainment applications.Video compression technology is the key of emerging video product technology, and video compression is a prerequisite to keep the video image quality mainly, as far as possible finishes video coding with minimum position, wherein, MPEG-2, MPEG-4, H.264/AVC and standard such as AVS the most common.Along with the raising of mobile phone operation efficiency, the evolution of battery power management technology, and the appearance of high-speed radio binding, increasing end product has action property and Video Applications function concurrently.For video product, how to utilize the limited calculation resources of hardware, designing practical coding/decoding system is an important topic.
Video bit stream (Bitstream) is typically formed by a succession of video picture frame (Frame), and each picture frame is divided into several macro blocks (Macroblock) again, and each macro block has generally included four 8 * 8 luma blocks and two 8 * 8 colourity blocks.Video coding carries out the compressed encoding computing of image data based on motion compensated prediction, conversion and quantification and entropy coding (Entropy coding).The picture frame that the motion prediction of image coding system (Motion estimation) utilizes last coding is as the basis, thereby calculates the motion vector (Motion vector) of macro block, with to the image data compressed encoding.
On the contrary, the motion compensation of (Inter-frame) decoding technique (Motion compensation) then is the motion vector that solves present macro block between the picture of image-decoding system, and in the reference picture frame, find out the most similar reference zone, and then solve each macro block.In order to carry out above-mentioned action, decoder must repeatedly carry out access to the dynamic random access memory that stores with reference to the picture frame image data, and therefore, motion compensation is for accounting for most the part of system's frequency range resource.And repeatedly to the dynamic random memory access data except the occupation problem that causes the frequency range resource, access action also causes power consumption and system effectiveness to reduce.Application at portable apparatus such as present mobile phone integration video capability if the power loss that access action caused is not improved, will be unfavorable for the electrical management of application apparatus.
In view of random access memory is carried out the access of single mass data, have more efficient compared to the access of low volume data repeatedly, thereby propose the present invention, in the hope of reducing the occupancy of motion compensation, and reduce power consumption and improve system effectiveness for system's frequency range.
Summary of the invention
Technical problem to be solved by this invention is, the self-adapting fetching-rapidly method of a kind of image-decoding system and motion compensation thereof is provided, it is by the motion vector of foundation macro block, determine a fast delivery formula, the pairing image data of fast delivery formula is accessed to high-speed cache (Cache memory) by dynamic random access memory stores, can reduce the demand of motion compensation (Motion compensation) for system's frequency range.
In order to realize above-mentioned purpose, the invention provides the self-adapting fetching-rapidly method that discloses a kind of motion compensation, it is applicable to an image-decoding system, this image-decoding system is decoded to a video bit stream (Bitstream).The step of this method is at first, to solve the motion vector (Motion vector) of next macro block of a macro block (Macroblock) and this macro block and this macro block.Secondly, motion vector according to next macro block of this macro block and this macro block, determine a fast delivery formula, wherein this fast delivery formula is got scope corresponding to one soon with reference to one on the picture frame (Reference frame), this image data with reference to picture frame is solved in advance, and is stored in the dynamic random access memory.Subsequently, this dynamic random access memory of access is stored to a high-speed cache (Cache memory) will be somebody's turn to do with reference to this image data of getting scope soon on the picture frame.Then, whether this high-speed cache of access is positioned at this with the required reference zone of next macro block that solves this macro block and gets scope soon.At last, according to previous step rapid in, whether next macro block of this macro block is positioned at the result that this gets scope soon, determines whether to keep this fast delivery formula.
In order to realize above-mentioned purpose, the present invention more discloses a kind of image-decoding system, and it is applicable to decodes to a video bit stream.This image-decoding system comprises a decoder, a dynamic random access memory, a fast delivery formula decision module and a high-speed cache.This decoder is to the decoding of this video bit stream, and producing one in advance with reference to picture frame, this decoder solves the motion vector of next macro block of a macro block of a present picture frame and this macro block and this macro block.This dynamic random access memory cooperates this image data with reference to picture frame of storage.This fast delivery formula decision module is according to the motion vector of next macro block of this macro block and this macro block, determine a fast delivery formula, this fast delivery formula is got scope corresponding to this soon with reference to one on the picture frame, should be with reference to this gets the image data of scope soon on the picture frame in this dynamic randon access memory of this fast delivery formula decision module and access.This high-speed cache cooperate store this fast delivery formula decision module institute access this with reference to this gets the image data of scope soon on the picture frame.This high-speed cache of this decoder accesses to utilize this image data of getting scope soon, solves next macro block of this macro block.
The present invention has following useful effect: the self-adapting fetching-rapidly method of disclosed image-decoding system and motion compensation thereof is according to the motion vector of representing dependence between macro block, determining a fast delivery formula, so with the pairing image data of scope of getting soon of this fast delivery formula in advance from dynamically being accessed to cache stores in the random asccess memory.Be with, the present invention can reduce in the motion compensation computing, to the number of times of dynamic random access memory access data, thereby reduce the demand of system's frequency range, improving system works usefulness, and reduce data access for power consumption, be beneficial to the electrical management of image-decoding system institute application apparatus.
For enabling further to understand feature of the present invention and technology contents, see also following about detailed description of the present invention and accompanying drawing, yet accompanying drawing only provide with reference to and the explanation usefulness, be not to be used for the present invention is limited.
Description of drawings
Fig. 1 is the system architecture schematic diagram of an image-decoding system;
Fig. 2 is the flow chart of steps of the self-adapting fetching-rapidly method of disclosed motion compensation;
Fig. 3 is the scope of getting the soon schematic diagram of the one dimension direction motion level of disclosed interdependent motion level;
Fig. 4 is the scope of getting the soon schematic diagram of the two-dimensional directional motion level of disclosed interdependent motion level;
Fig. 5 is the scope of getting the soon schematic diagram of the compound movement level of disclosed interdependent motion level;
Fig. 6 is the scope of getting the soon schematic diagram of disclosed non-interdependent motion level;
Fig. 7 is the disclosed schematic diagram of getting the data displacement soon; And
Fig. 8 is the system architecture schematic diagram of a specific embodiment of disclosed image-decoding system.
Wherein, Reference numeral:
10,20: image-decoding system
11: processor
15,25: dynamic random access memory
17,27: high-speed cache
21: decoder
23: fast delivery formula decision module
231: motion level decision unit
233: caching data amount decision unit
235: the caching data permute unit
237: get tactful adjustment unit soon
30,40,50,60,70: with reference to picture frame
300,400,500,600,700: the reference position
310,410,510,610,710: get scope soon
7100: non-term of reference
A: position
Embodiment
The present invention is directed in motion compensation (Motioncompensation) calculating process of (Inter-frame) decoding between image frame, the image data that single is a large amount of to the dynamic random access memory access, compared in batches to dynamic random access memory accessing image data, more save frequency range, thereby utilize the motion vector (M0tion vector) of representing dependence between macro block, dynamically determine a fast delivery formula, so that single is from a large amount of in advance accessing image data of dynamic random access memory, and image data is stored in the high-speed cache (Cachememory), so that video bit stream (Bitstream) De Xie Code computing to be provided.
At first, see also Fig. 1, this figure is the system architecture schematic diagram of an image-decoding system.As shown in Figure 1, image-decoding system 10 has comprised a processor 11, a dynamic random access memory 15 and a high-speed cache 17, and this is the typical image-decoding hardware structure of general video terminal product.Processor 11 is as the main mechanism of computing that video bit stream is decoded, wherein Xiang Guan decoding computing has comprised that each macro block (Macroblock) with each picture frame (Frame) in the video bit stream solves, with each drawing frame data of reconstruction picture group, and then image is reappeared.Image-decoding system 10 can more comprise the digital signal processing module that is not shown in Fig. 1, to cooperate processor 11 computing of decoding.The picture frame image data that dynamic random access memory 15 cooperates storage processor 11 to be solved.At the motion compensation computing, the macro block image data of the reference picture frame relevant with the computing of present picture frame macro block is accessed to high-speed cache 17 in advance and is stored, processor 11 accessing caches 17 stored image datas are decoded, to reduce the time of delay of data access.
In one embodiment, described processor 11 is a reduced instruction set computer (RISC, a Reducedinstruction set computing) processor, in order to the execution of each instruction of image-decoding process.
Then, see also Fig. 2, this figure is the flow chart of steps of the self-adapting fetching-rapidly method of disclosed motion compensation, and this motion compensation process is applicable to image-decoding system 10.As shown in Figure 2, the method comprises the following steps:
At first, processor 11 solve-at present on the picture frame-motion vector (step S200) of the following-macro block of macro block and this macro block and this macro block;
Secondly, motion vector according to next macro block of this macro block and this macro block determines a fast delivery formula, wherein this fast delivery formula corresponding to one with reference to one getting scope soon on the picture frame, this is solved by processor 11 in advance with reference to picture frame, and is stored in (step S202) in the dynamic random access memory 15;
Then, processor 11 access dynamic random access memorys 15 are stored to high-speed cache 17 (step S204) will be somebody's turn to do with reference to this image data of getting scope soon on the picture frame;
Subsequently, processor 11 accessing caches 17, solving next macro block of this macro block, and then whether the required reference zone of next macro block that solves this macro block is positioned at this and gets scope (step S206) soon;
Come, according to the result of step S206, whether decision keeps this fast delivery formula (step S208) again;
If for keeping this fast delivery formula, repeating step S206 then is to continue utilizing these high-speed cache 17 stored image datas to solve next macro block again of this next macro block according to the decision of step S208;
If for not keeping this fast delivery formula, repeating step S202 then determines another fast delivery formula to utilize next macro block and the motion vector of next macro block again according to the decision of step S208.
In step S206, if processor 11 in this gets the image data of scope soon, is not found out the reference zone identical image data required with next macro block of this macro block, then processor 11 is an access dynamic random access memory 15, to solve next macro block of this macro block.
Fast delivery formula proposed by the invention has comprised a motion level, a caching data amount and caching data displacement (Swap) pattern.
The decision of the motion level of just fast delivery formula at first, earlier explains.Described motion level determined by the dependence between adjacent macroblocks, that is utilizes the represented direction of motion of motion vector of this macro block of present picture frame and next macro block to decide the motion level.The motion level has comprised an interdependent motion level and a non-interdependent motion level, described interdependent motion level has comprised one dimension direction motion level, two-dimensional directional motion level and a compound movement level, below respectively each motion layer level is illustrated by Fig. 3 to Fig. 6.
See also Fig. 3, this figure is the scope of getting the soon schematic diagram of the one dimension direction motion level of interdependent motion level.As shown in Figure 3, reference position 300 is the position of the macro block of present picture frame corresponding to reference picture frame 30.If motion vector according to this macro block and next macro block, next macro block is along positive x direction displacement, then next macro block position that may occur should be the position of the positive x direction of contiguous reference position 300, thereby judging that the required reference zone of next macro block should drop on gets in the scope 310 soon.Therefore, processor 11 access dynamic random access memorys 15 will be accessed to high-speed cache 17 with reference to the image data of getting scope 310 on the picture frame 30 soon and store, to offer next time macro block of processor 11 computings.For instance, described-Wei direction motion level can be moving horizontally of still frame or image etc.
See also Fig. 4, this figure is the scope of getting the soon schematic diagram of the two-dimensional directional motion level of interdependent motion level.As shown in Figure 4, reference position 400 is the position of the macro block of present picture frame corresponding to reference picture frame 40.If motion vector according to this macro block and next macro block, next macro block is except the positive x direction displacement in edge, still along positive y direction or the displacement of negative y direction, then next macro block position that may occur should be the positive x direction of contiguous reference position 400 and the position of positive and negative y direction, thereby judging that the required reference zone of next macro block should drop on gets in the scope 410 soon.Therefore, the image data of the stored scope of getting soon 410 with reference to picture frame 40 of processor 11 access dynamic random access memorys 15 is accessed to high-speed cache 17 and stores, to offer next time macro block of processor 11 computings.For instance, described two-dimensional directional motion level can be image amplification, dwindle or rotation etc.
See also Fig. 5, this figure is the scope of getting the soon schematic diagram of the compound movement level of interdependent motion level.As shown in Figure 5, reference position 500 is the position of the macro block of present picture frame corresponding to reference picture frame 50.If according to the motion vector of this macro block and next macro block, the displacement of next macro block is bigger, the location about that next macro block position that may occur may reference position 500 then, thus judging that the required reference zone of next macro block should drop on gets in the scope 510 soon.Therefore, the image data of the stored scope of getting soon 510 with reference to picture frame 50 of processor 11 access dynamic random access memorys 15 is accessed to high-speed cache 17 and stores, to offer next time macro block of processor 11 computings.Described compound movement level refers to the compound movement of image in the same scene.
See also Fig. 6, this figure is the scope of getting the soon schematic diagram of non-interdependent motion level.Figure and the conversion of reference diagram occurrence scene before the feeling the pulse with the finger-tip of described non-interdependent motion level, therefore, in picture frame 60, it is irrelevant with reference position 600 to get scope 610 soon, at this moment, processor 11 can not carry out access to dynamic random access memory 15, to avoid taking unnecessary frequency range resource.
Do not drop on getting soon in the scope of prediction if solve the reference zone of next macro block, then represent the dependence between macro block to change.At this moment, must determine the motion level again, to guarantee the hit rate of caching data.
Secondly, the decision about the caching data amount of fast delivery formula is described.Based on utilizing single that dynamic random access memory 15 access mass data are replaced repeatedly access, to reduce taking and the notion that improves system effectiveness of frequency range, can further the image data amount of getting scope soon dynamically be adjusted at the correctness of caching data and the motion level between macro block.Just, the motion layer level when between macro block is the one dimension direction motion level of interdependent motion level continuously, then can heighten the image data amount of getting scope soon.For example, in Fig. 3, increase the macro block image data of the x direction of getting scope 310 soon.If the motion layer level between macro block is the two-dimensional directional motion level of interdependent motion level continuously, then can heighten the image data amount of getting scope soon.For example, in Fig. 4, increase the macro block image data of the x direction of getting scope 410 soon.If the motion layer level between macro block is the compound movement level of interdependent motion level continuously, it is on the low side then to have represented the required reference zone of next macro block to drop on the probability of the scope of getting soon of prediction, at this moment, can further reduce the image data amount of the scope of getting soon, even do not make data and get soon, to avoid taking unnecessary frequency range resource.Moreover if the motion level between macro block is non-interdependent motion level, the image data amount that can directly will get soon is adjusted into zero, does not make data and gets soon, to avoid taking of frequency range.
The caching data displacement patterns of fast delivery formula then, is described.Because the memory capacity of high-speed cache 17 is limited, therefore, the image data of computing reference value of must not having displaces from high-speed cache 17, with the increase storage area.Described caching data displacement patterns is relevant with the motion level, just, motion level when between macro block is the one dimension direction motion level or the two-dimensional directional motion level of interdependent motion level, then the image data opposite with the direction of motion can be displaced high-speed cache 17.See also Fig. 7, this figure is the disclosed schematic diagram of getting the data displacement soon, and as shown in Figure 7, macro block is done displacement along positive x direction, is stored in advance in the high-speed cache 17 with reference to an image data of getting scope 710 soon on the picture frame 70.When macro block continues to be displaced to reference position 700, then the image data of the non-term of reference 7100 of position A left can be displaced high-speed cache 17.
Then, see also Fig. 8, this figure is the system architecture schematic diagram of a specific embodiment of disclosed image-decoding system, and image-decoding system 20 is applicable to decodes to a video bit stream.As shown in Figure 8, image-decoding system 20 has comprised a decoder 21, one fast delivery formula decision module 23, a dynamic random access memory 25 and a high-speed cache 27.Wherein, decoder 21 is made of hardware such as software program collocation processors jointly with fast delivery formula decision module 23.
The decoding of the 21 pairs of video bit streams of decoder, producing one in advance with reference to picture frame, decoder 21 solves the motion vector of next macro block of a macro block of a present picture frame and this macro block and this macro block.Dynamic random access memory 25 cooperates the image data that stores with reference to picture frame.Fast delivery formula decision module 23 is according to the motion vector of next macro block of this macro block and this macro block, determine a fast delivery formula, this fast delivery formula is got scope soon corresponding to one on the reference picture frame, in fast delivery formula decision module 23 and the access dynamic random access memory 25 with reference to this gets the image data of scope soon on the picture frame.This gets the image data of scope soon on the reference picture frame of 23 accesses of the high-speed cache 27 cooperation fast delivery formula decision module of storage.Decoder 21 accessing caches 27 to utilize this image data of getting scope soon, solve next macro block of this macro block.
Based on above stated specification, fast delivery formula has comprised a motion level, a caching data amount and a caching data displacement patterns.In Fig. 8, fast delivery formula decision module 23 comprises that again motion level decision unit 231, a caching data amount determine unit 233, a caching data permute unit 235 and to get tactful adjustment unit 237 soon.Motion level decision unit 231 determines a motion level according to the motion vector of next macro block of this macro block and this macro block.Caching data amount decision unit 233 is according to this motion level, and data quantity is got in decision soon.The displacement patterns of caching data permute unit 235 decision caching datas is with the image data of judging that high-speed cache 27 should displace.Get the operation result of tactful adjustment unit 237 soon according to decoder 21, and then whether the required reference zone of next macro block that solves this macro block is positioned at this and gets scope soon, keep this fast delivery formula to control fast delivery formula decision module 23, or control fast delivery formula decision module 23 another fast delivery formulas of decision.That is, get tactful adjustment unit 237 may command motion levels decision unit 231 another motion levels of decision soon, and according to this motion level, control caching data amount decision unit 233 and caching data permute unit 235 determine another caching data amount and another caching data displacement patterns respectively.
What deserves to be mentioned is, according to simulation result of the present invention, when the motion level is the one dimension direction motion level of interdependent motion level and two-dimensional directional motion level, can save more than 20 percent for the frequency range demand of system.Moreover, if when image is compound movement level or non-interdependent motion level, can utilize the control of the caching data amount of fast delivery formula, with further reduction caching data amount, thereby reduce the shared system's frequency range of data access.
Describe in detail by above example, when the self-adapting fetching-rapidly method that can know disclosed image-decoding system and motion compensation thereof according to the motion vector of representing dependence between macro block, determining a fast delivery formula, so with the pairing image data of scope of getting soon of this fast delivery formula in advance from dynamically being accessed to cache stores in the random asccess memory.Be with, the present invention can reduce in the motion compensation computing, to the number of times of dynamic random access memory access data, thereby reduce the demand of system's frequency range, improving system works usefulness, and reduce data access for power consumption, be beneficial to the electrical management of image-decoding system institute application apparatus.
Certainly; the present invention also can have other various embodiments; under the situation that does not deviate from spirit of the present invention and essence thereof; being familiar with those of ordinary skill in the art ought can make various corresponding changes and distortion according to the present invention, but these corresponding changes and distortion all should belong to the protection range of the appended claim of the present invention.
Claims (10)
1, a kind of self-adapting fetching-rapidly method of motion compensation is characterized in that, is applicable to an image-decoding system, and this image-decoding system system decodes to a video bit stream, and this self-adapting fetching-rapidly method comprises the following steps:
(a) solve the motion vector of next macro block of a macro block and this macro block and this macro block;
(b) according to the motion vector of next macro block of this macro block and this macro block, determine a fast delivery formula, wherein this fast delivery formula is got scope corresponding to one soon with reference to one on the picture frame, and this image data with reference to picture frame is solved in advance and is stored in the dynamic random access memory;
(c) this dynamic random access memory of access is stored to a high-speed cache will be somebody's turn to do with reference to this image data of getting scope soon on the picture frame;
Whether (d) this high-speed cache of access is positioned at this with the required reference zone of next macro block that solves this macro block and gets scope soon; And
(e) according to the result of step (d), whether decision keeps this fast delivery formula.
2, self-adapting fetching-rapidly method according to claim 1 is characterized in that, if the decision of step (e) is for keeping this fast delivery formula, repeating step (d) then, if the decision of step (e) is for keeping this fast delivery formula, repeating step (b) then is to determine another fast delivery formula.
3, self-adapting fetching-rapidly method according to claim 1 is characterized in that, with reference to the reference position on the picture frame, it is relevant with this reference position that this gets scope soon corresponding to this for this macro block.
4, self-adapting fetching-rapidly method according to claim 1, it is characterized in that, this fast delivery formula comprises a motion level, the decision of this motion level is relevant with the motion vector of next macro block of this macro block and this macro block, this motion level comprises an interdependent motion level and a non-interdependent motion level, and wherein this interdependent motion level includes an one dimension direction motion level, a two-dimensional directional motion level and a compound movement level.
5, self-adapting fetching-rapidly method according to claim 4, it is characterized in that, this fast delivery formula more comprises a caching data amount and a caching data displacement patterns, this caching data displacement patterns decision should displace the image data of this high-speed cache, and wherein the decision of this caching data amount and this caching data displacement patterns system is relevant with this motion level.
6, a kind of image-decoding system is characterized in that, is applicable to a video bit stream is decoded, and this image-decoding system comprises:
One decoder, to the decoding of this video bit stream, producing one in advance with reference to picture frame, this decoder solves the motion vector of next macro block of a macro block of a present picture frame and this macro block and this macro block;
One dynamic random access memory cooperates to store this image data with reference to picture frame;
One fast delivery formula decision module, motion vector according to next macro block of this macro block and this macro block, determine a fast delivery formula, this fast delivery formula is got scope corresponding to this soon with reference to one on the picture frame, and this fast delivery formula decision module and this dynamic random access memory of access should be with reference to this gets the image data of scope soon on the picture frame; And
One high-speed cache, this that cooperate to store this fast delivery formula decision module institute access is with reference to this gets the image data of scope soon on the picture frame;
Wherein this high-speed cache of this decoder accesses to utilize this image data of getting scope soon, solves next macro block of this macro block.
7, image-decoding system according to claim 6 is characterized in that, with reference to the reference position on the picture frame, it is relevant with this reference position that this gets scope soon corresponding to this for this macro block.
8, image-decoding system according to claim 6, it is characterized in that, this fast delivery formula comprises a motion level, this motion level is relevant with the motion vector of next macro block of this macro block and this macro block, this fast delivery formula decision module comprises motion level decision unit, this motion level decision unit is according to the motion vector of next macro block of this macro block and this macro block, determine this motion level, this motion level comprises an interdependent motion level and a non-interdependent motion level, and wherein this interdependent motion level includes an one dimension direction motion level, an one two-dimensional directional motion level and a compound movement level.
9, image-decoding system according to claim 8, it is characterized in that, this fast delivery formula more comprises a caching data amount and a caching data displacement patterns, this fast delivery formula decision module more comprises a caching data amount decision unit and a caching data permute unit, this caching data amount decision unit determines this caching data amount, this caching data permute unit determines this caching data displacement patterns, displace the image data of this high-speed cache with decision, wherein the decision of this caching data amount and this caching data displacement patterns is relevant with this motion level.
10, image-decoding system according to claim 6, it is characterized in that, this fast delivery formula decision module comprises that one gets tactful adjustment unit soon, this is got tactful adjustment unit soon and whether is positioned at this according to next macro block of this macro block and gets scope soon, keep this fast delivery formula with this fast delivery formula decision module of control, or determine another fast delivery formula.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN 200710097481 CN101304520A (en) | 2007-05-09 | 2007-05-09 | Image decoding system and self-adapting fetching-rapidly method for motion compensation thereof |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN 200710097481 CN101304520A (en) | 2007-05-09 | 2007-05-09 | Image decoding system and self-adapting fetching-rapidly method for motion compensation thereof |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN101304520A true CN101304520A (en) | 2008-11-12 |
Family
ID=40114193
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN 200710097481 Pending CN101304520A (en) | 2007-05-09 | 2007-05-09 | Image decoding system and self-adapting fetching-rapidly method for motion compensation thereof |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN101304520A (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101990102A (en) * | 2010-11-09 | 2011-03-23 | 福州瑞芯微电子有限公司 | Data reading method and device on embedded equipment in video decoding process |
| US20150055707A1 (en) * | 2013-08-26 | 2015-02-26 | Amlogic Co., Ltd. | Method and Apparatus for Motion Compensation Reference Data Caching |
| CN105430410B (en) * | 2014-09-17 | 2019-01-11 | 联发科技股份有限公司 | Motion compensation unit and motion compensation process |
| CN113170103B (en) * | 2018-12-07 | 2024-03-15 | 松下电器（美国）知识产权公司 | Encoding device, decoding device, encoding method, and decoding method |
-
2007
- 2007-05-09 CN CN 200710097481 patent/CN101304520A/en active Pending
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101990102A (en) * | 2010-11-09 | 2011-03-23 | 福州瑞芯微电子有限公司 | Data reading method and device on embedded equipment in video decoding process |
| CN101990102B (en) * | 2010-11-09 | 2013-07-24 | 福州瑞芯微电子有限公司 | Data reading method and device on embedded equipment in video decoding process |
| US20150055707A1 (en) * | 2013-08-26 | 2015-02-26 | Amlogic Co., Ltd. | Method and Apparatus for Motion Compensation Reference Data Caching |
| US9363524B2 (en) * | 2013-08-26 | 2016-06-07 | Amlogic Co., Limited | Method and apparatus for motion compensation reference data caching |
| CN105430410B (en) * | 2014-09-17 | 2019-01-11 | 联发科技股份有限公司 | Motion compensation unit and motion compensation process |
| CN113170103B (en) * | 2018-12-07 | 2024-03-15 | 松下电器（美国）知识产权公司 | Encoding device, decoding device, encoding method, and decoding method |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103650509B (en) | Adaptive configuration of reference frame buffer based on camera and background motion | |
| CN102246522B (en) | Intelligent decoded picture buffering | |
| CN100484246C (en) | Pixel prefetching device of motion compensating module in AVS video hardware decoder | |
| CN101115208A (en) | Picture coding apparatus | |
| US9351004B2 (en) | Multiview video coding reference picture selection under a one reference picture constraint | |
| CN101193306A (en) | Motion vector detecting apparatus and motion vector detecting method | |
| KR20160094308A (en) | Image processing device and semiconductor device | |
| WO2023142716A1 (en) | Encoding method and apparatus, real-time communication method and apparatus, device, and storage medium | |
| CN101304520A (en) | Image decoding system and self-adapting fetching-rapidly method for motion compensation thereof | |
| CN101459842A (en) | Decoding method and apparatus for space desampling | |
| US20060088097A1 (en) | Moving picture encoding apparatus having increased encoding speed and method thereof | |
| CN101557518B (en) | Method and device for compensating motion, method and device for replacing cache | |
| JP2024514348A (en) | Video encoding methods, equipment and electronic equipment | |
| US7881367B2 (en) | Method of video coding for handheld apparatus | |
| US9363524B2 (en) | Method and apparatus for motion compensation reference data caching | |
| CN111355962A (en) | Video decoding caching method suitable for multiple reference frames, computer device and computer readable storage medium | |
| WO2022206217A1 (en) | Method and apparatus for performing image processing in video encoder, and medium and system | |
| CN101977317B (en) | Intra-frame prediction method and device | |
| US20130127887A1 (en) | Method for storing interpolation data | |
| CN100586188C (en) | A kind of hardware implementation method of calculating based on the infra-frame prediction of AVS | |
| CN100496126C (en) | Image coder and coding method | |
| CN104041037B (en) | Dynamic image encoding method, device | |
| Silveira et al. | Reference frame context-adaptive variable-length coder: a real-time hardware-friendly approach for lossless external memory bandwidth reduction in current video-coding systems | |
| WO2022206199A1 (en) | Method and apparatus for performing image processing in video decoding apparatus, and system | |
| WO2023065890A1 (en) | Multimedia data processing method and apparatus, computer device, computer readable storage medium, and computer program product |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C02 | Deemed withdrawal of patent application after publication (patent law 2001) | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Open date: 20081112 |