CN103248888A - Sample adaptive offset (SAO) in accordance with video coding - Google Patents

Abstract

Sample adaptive offset (SAO) in accordance with video coding. SAO filtering can be performed before e-blocking processing (e.g., in accordance with video signal decoding and/or encoding). For example, a receiver and/or decoder communication device can receive signaling from a transmitter and/or encoder communication device that includes various band offsets. Corresponding band indices can be determined via analysis of the received video signal (e.g., received from the transmitter and/or encoder communication device), inferentially without requiring signaling of such band indices from the transmitter and/or encoder communication device. Upon appropriate analysis of one or more largest coding units (LCUs) generated from the video signal to determine a pixel value distribution (e.g., which may be using a histogram in one instance), then based on that pixel value distribution, the band indices are identified and the band offsets applied thereto.

Description

Sampled point self adaptation skew (SAO) according to the video coding

The cross reference of relevant patent/patent application

The application requires the U.S. Provisional Patent Application the 61/597th of submission on February 10th, 2012, the U.S. Provisional Patent Application the 61/598th that on February 13rd, No. 683 1 submitted to, the U.S. Provisional Patent Application the 61/603rd that on February 24th, No. 326 1 submitted to, the U.S. Patent application the 13/623rd that on September 20th, No. 190 1 submitted to, the U.S. Patent application the 13/758th that No. 765 and on February 4th, 2013 submit to, No. 169 priority, it is included in this for your guidance.

Technical field

The present invention relates in general to Digital Video Processing; More specifically, relate to processing and operation according to this Digital Video Processing.

Background technology

The communication system that operation is used for transmitting digital media (for example, image, video, data etc.) has continued to have developed for many years.For the communication system of the video data of this certain form of employing, export or show that a plurality of digital pictures are to realize being applicable to the vision signal of output and consumption with certain frame rate (for example, number of pictures per second).In a plurality of such communication system of utilizing video data to operate, between the video of throughput (for example, can transfer to the quantity of the picture frame of the second place from primary importance) and the signal that finally will be output or show and/or picture quality, exist compromise.Current techniques can't be fully or following means is provided acceptably: namely, it can be according to providing sufficient or acceptable video and/or picture quality transfer to the second place with video data from primary importance, thereby the expense of the relatively small amount that assurance is associated with communication, and the relatively low complexity of the communication equipment of each communication chain terminal etc.

Summary of the invention

(1) a kind of device comprises:

At least one input is used for:

From at least one extra means receiving video signals; And

Receive a plurality of band bendings via signaling from described at least one extra means; And

Processor is used for:

Analyze a plurality of pixels relevant with at least one maximum coding unit (LCU) and distribute with the identification pixel value, wherein, described maximum coding unit is relevant with described vision signal;

Identify a plurality of band indexs inferentially based on described pixel value distribution;

Handle according to described vision signal or based on the filtering of the signal of described vision signal described a plurality of band bendings are applied to described a plurality of band index;

Handle to generate first filtering signal to described vision signal or based on signal execution sampled point self adaptation skew (SAO) filtering of described vision signal, wherein, described sampled point self adaptation skew filtering is handled described a plurality of band bendings is applied to described a plurality of band index; And

Described first filtering signal is carried out block elimination filtering handle to generate second filtering signal.

(2) according to (1) described device, wherein, described processor is used for:

Analyze a plurality of pixels that are associated with at least one maximum coding unit and represent the pixel value histogram that described pixel value distributes with generation, wherein, described maximum coding unit is relevant with described vision signal; And

Based on described pixel value histogram, identification will be employed described a plurality of band indexs of described a plurality of band bendings.

(3) according to (1) described device, wherein:

Described a plurality of band index has discontinuously arranged, makes at least two continuous band indexs at least one band index value separated from one another in described a plurality of band index.

(4) according to (1) described device, wherein:

The described pixel value indication a plurality of subclass of the described a plurality of pixels relevant with at least a portion in described a plurality of band indexs respectively that distribute; And

Be employed described a plurality of band indexs of described a plurality of band bendings corresponding at least one subclass in a plurality of subclass of described a plurality of pixels, wherein, other subclass in a plurality of subclass of this at least one subclass and described a plurality of pixels are compared the pixel with relatively large or maximum quantity.

(5) according to (1) described device, wherein:

Described device is the communication equipment of operating in satellite communication system, wireless communication system, wired communication system, optical fiber telecommunications system and mobile communication system at least one.

(6) a kind of device comprises:

Input is used for from least one extra means receiving video signals and a plurality of band bending; And

Processor is used for:

Analyze a plurality of pixels relevant with at least one maximum coding unit (LCU) and distribute with the pixel value that identification is used for a plurality of band indexs of identification, described maximum coding unit is relevant with described vision signal; And

Handle according to described vision signal or based on the filtering of the signal of described vision signal described a plurality of band bendings are applied to described a plurality of band index.

(7) according to (6) described device, wherein, described processor is used for:

Handle to generate first filtering signal to described vision signal or based on signal execution sampled point self adaptation skew (SAO) filtering of described vision signal, wherein, described sampled point self adaptation skew filtering is handled described a plurality of band bendings is applied to described a plurality of band index; And

Described first filtering signal is carried out block elimination filtering handle to generate second filtering signal.

(8) according to (6) described device, wherein, described processor is used for:

Analyze a plurality of pixels relevant with at least one maximum coding unit to generate the pixel value histogram that the described pixel value of expression distributes, wherein, described maximum coding unit is relevant with described vision signal; And

Based on described pixel value histogram, identification will be employed a plurality of band indexs of described a plurality of band bendings.

(9) according to (6) described device, wherein:

Described a plurality of band index has discontinuously arranged, makes at least two continuous band indexs at least one band index value separated from one another in described a plurality of band index.

(10) according to (6) described device, wherein:

The described pixel value indication a plurality of subclass of the described a plurality of pixels relevant with at least a portion in described a plurality of band indexs respectively that distribute; And

Be employed described a plurality of band indexs of described a plurality of band bendings corresponding at least one subclass in a plurality of subclass of described a plurality of pixels, wherein, other subclass in a plurality of subclass of this at least one subclass and described a plurality of pixels are compared the pixel with relatively large or maximum quantity.

(11) according to (6) described device, wherein:

Described a plurality of band bending is received from described at least one extra means via signaling by described device; And

Described processor is used for identifying described a plurality of band index inferentially based on described pixel value distribution.

(12) according to (6) described device, wherein:

Described device is the receiver communication equipment that comprises Video Decoder;

Described at least one extra means is the reflector communication equipment that comprises video encoder; And

Described receiver communication equipment and described reflector communication equipment connect via at least one communication channel or couple communicatedly.

(13) according to (6) described device, wherein:

Described device is the communication equipment of operating in satellite communication system, wireless communication system, wired communication system, optical fiber telecommunications system and mobile communication system at least one.

(14) a kind of method of operation of communication equipment, described method comprises:

Via the input of described communication equipment, from least one additional communication equipment receiving video signals and a plurality of band bending;

Analyze a plurality of pixels relevant with at least one maximum coding unit (LCU) and distribute with the pixel value that identification is used for a plurality of band indexs of identification, wherein, described maximum coding unit is relevant with described vision signal; And

Handle according to described vision signal or based on the filtering of the signal of described vision signal described a plurality of band bendings are applied to described a plurality of band index.

(15) according to (14) described method, further comprise:

Handle to generate first filtering signal to described vision signal or based on signal execution sampled point self adaptation skew (SAO) filtering of described vision signal, wherein, described sampled point self adaptation skew filtering is handled and is comprised described a plurality of band bendings are applied to described a plurality of band index; And

Described first filtering signal is carried out block elimination filtering handle to generate second filtering signal.

(16) according to (14) described method, further comprise:

Analyze a plurality of pixels relevant with described at least one maximum coding unit to generate the pixel value histogram that the described pixel value of expression distributes, wherein, described maximum coding unit is relevant with described vision signal; And

Based on described pixel value histogram, identification will be employed a plurality of band indexs of described a plurality of band bendings.

(17) according to (14) described method, wherein:

Described a plurality of band index has discontinuously arranged, makes at least two continuous band indexs at least one band index value separated from one another in described a plurality of band index.

(18) according to (14) described method, wherein:

The described pixel value indication a plurality of subclass of the described a plurality of pixels relevant with at least a portion in described a plurality of band indexs respectively that distribute; And

Be employed described a plurality of band indexs of described a plurality of band bendings corresponding at least one subclass in a plurality of subclass of described a plurality of pixels, wherein, other subclass in a plurality of subclass of this at least one subclass and described a plurality of pixels are compared the pixel with relatively large or maximum quantity.

(19) according to (14) described method, further comprise:

Receive described a plurality of band bending via signaling from described at least one additional communication equipment; And

Identify described a plurality of band index inferentially based on described pixel value distribution.

(20) according to (14) described method, wherein:

Operate at least one in satellite communication system, wireless communication system, wired communication system, optical fiber telecommunications system and mobile communication system of described communication equipment.

Description of drawings

Fig. 1 and Fig. 2 show each execution mode of communication system.

Fig. 3 A shows the execution mode of computer.

Fig. 3 B shows the execution mode of notebook computer.

Fig. 3 C shows the execution mode of high definition (HD) TV.

Fig. 3 D shows the execution mode of single-definition (SD) TV.

Fig. 3 E shows the execution mode of hand-held media units.

Fig. 3 F shows the execution mode of set-top box (STB).

Fig. 3 G shows the execution mode of digital video disk (DVD) player.

Fig. 3 H shows the execution mode of general digital image and/or video processing equipment.

Fig. 4, Fig. 5 and Fig. 6 show the diagram of each execution mode of video coding framework.

Fig. 7 shows the diagram of the execution mode of intra-prediction process.

Fig. 8 shows the diagram of the execution mode of inter prediction processing.

Fig. 9 and Figure 10 show the diagram of each execution mode of video decode framework.

Figure 11 shows the execution mode that band bending sampled point self adaptation skew (SAO) filtering is handled.

Figure 12 shows the replaceable execution mode of video coding framework.

Figure 12 shows the execution mode of slicer grammer.

Figure 13 shows each execution mode of the indication (self adaptation and/or explicit signaling) of the transmission band skew under sampled point self adaptation skew (SAO) band bending pattern.

Figure 14 shows each execution mode of the indication (self adaptation and/or explicit signaling) of the frequency band granularity under the SAO band bending pattern.

Figure 15 shows the execution mode of implicit expression band index signaling.

Figure 16 shows the replaceable execution mode of implicit expression band index signaling.

Figure 17 shows the execution mode of band bending coding.

Figure 18 and Figure 19 show each execution mode for the method for the one or more equipment of operation (for example, communication equipment, receiver and/or decoder apparatus, reflector and/or encoder device etc.).

Embodiment

In a plurality of equipment that use Digital Media (such as digital video), utilize pixel to represent each image (it is in the nature numeral).In some communication system, Digital Media can transfer to the second place that can export or show this medium from primary importance.The purpose of digital communication system (comprising that operation is used for the communication system of communication digital video) is error free ground or with acceptable low error rate numerical data is transferred to another position from a position or subsystem.As shown in Figure 1, data can be transmitted by each communication channel in various communication systems: the media of magnetic media, wired, wireless, optical fiber, copper and/or other types.

Fig. 1 and Fig. 2 show each execution mode of communication system 100 and 200 respectively.

With reference to Fig. 1, the execution mode of communication system 100 is communication channels 199, and the communication equipment 110(that this communication channel will be positioned at communication channel 199 1 ends comprises the reflector 112 with encoder 114 and comprises the receiver 116 with decoder 118) be coupled to another communication equipment 120(that is positioned at communication channel 199 other ends communicatedly and comprise the reflector 126 with encoder 128 and comprise the receiver 122 with decoder 124).In some embodiments, communication equipment 110 and one of 120 may only comprise reflector or receiver.Exist several dissimilar media that communication channel 199 can be by its realization (for example, use dish 132 and 134 satellite communication channel 130, use tower 142 and 144 and/or radio communication channel 140, the wire communication channel 150 of local antenna 152 and 154 and/or use the fiber optic communication channel 160 of electric light (E/O) interface 162 and photoelectricity (O/E) interface 164).In addition, can realize more than one type media, and can interface together, thereby form communication channel 199.

Be noted that under the situation that does not deviate from scope and spirit of the present invention such communication equipment 110 and/or 120 can be static equipment or mobile device.For example, communication equipment 110 and one of 120 or both can realize in the fixed position maybe can be have with more than one Network Access Point (for example, different access point (AP) separately in comprising the mobile communication system environment of one or more wireless lan (wlan)s, different satellite separately in comprising the mobile communication system environment of one or more satellites or comprising different Network Access Point separately in the mobile communication system environment of one or more Network Access Points usually can utilize communication equipment 110 and/or 120 to realize communication by these network nodes) be associated and/or the mobile communication equipment of the ability that communicates.

Usually adopt error correction and channel coding schemes to reduce the inevitable transmission error that suffers in the communication system.Generally speaking, these error corrections and channel coding schemes relate in the emitter terminals of communication channel 199 and use encoder and use decoder at the receiver end of communication channel 199.

In any so required communication system (for example, comprise those variation of describing at Fig. 1), any information storing device (for example, hard disk drive (HDD), network information memory device and/or server etc.) in any application of maybe needing information is encoded and/or decoded, can adopt described any polytype ECC code.

In general, when considering video data from a position or subsystem when transferring to the communication system of another position, can think that generally the transmission ends that video data is coded in communication channel 199 carries out, can think that generally video data decoding carries out at the receiving terminal of communication channel 199.

Similarly, can two-way communication although the execution mode of this diagram shows between communication equipment 110 and 120, certainly be noted that, in some embodiments, communication equipment 110 may include only the video data encoding function, communication equipment 120 may include only the video data decoding function, and vice versa the unidirectional execution mode of video broadcasting execution mode (for example, according to).

Communication system 200 with reference to Fig. 2, transmitting terminal in communication channel 299, for example provide information bit 201(to reflector 297, especially corresponding to the video data in the execution mode), this reflector can be operated be used to utilizing encoder and symbol mapper 220(can be considered as separate functional blocks 222 and 224 respectively) carry out the decoding of these information bits 201, thereby generate series of discrete value modulation symbol 203, this series of discrete value modulation symbol is offered emission driver 230, and this emission driver use digital to analog converter (DAC) 232 generates to transmit signal 204 continuous time and utilize emission filter 234 to generate roughly and transmits signal 205 the filtered continuous time that matches with communication channel 299.Receiving terminal in communication channel 299, to receive signal 206 continuous time and offer AFE (analog front end) (AFE) 260, this AFE (analog front end) comprise receiving filter 262(its generate receive signal 207 filtered continuous time) and analog to digital converter (ADC) 264(its generate discrete time and receive signal 208).Tolerance maker 270 calculate decoders 280 be used for to the centrifugal pump modulation symbol and wherein 210 information encoded positions carry out the best-estimated metric 209(for example, based on symbol and/or).

In each of reflector 297 and receiver 298, can realize any required being integrated in wherein such as each assembly, piece, functional block, circuit.For example, this illustrates processing module 280a and comprises encoder and symbol mapper 220 and all corresponding components that are associated wherein, processing module 280 is shown comprises tolerance maker 270 and decoder 280 and all corresponding components that are associated wherein.Such processing module 280a and 280b can be corresponding integrated circuits.Certainly, under the situation that does not deviate from scope and spirit of the present invention, can replacedly carry out other borders and grouping.For example, all parts in the reflector 297 can be comprised in first processing module or the integrated circuit, all parts in the receiver 298 can be comprised in second processing module or the integrated circuit.Replacedly, in other embodiments can combination in any reflector 297 and each of receiver 298 in parts.

The same with aforementioned embodiments, such communication system 200 can be used for video data from a position or subsystem transfers to the communication of another position (for example, via communication channel 299 from reflector 297 to receiver 298).

Can realize the Video processing of digital picture and/or digital picture and/or medium (comprising each image in the digital video signal) by in each equipment shown in following Fig. 3 A to Fig. 3 H any one, thereby allow the user to watch this digital picture and/or video.These plurality of devices do not comprise the equipment list of limit, wherein can realize image described herein and/or Video processing, be noted that, under the situation that does not deviate from scope and spirit of the present invention, can realize that any general digital image and/or video processing equipment are in order to carry out processing as herein described.

Fig. 3 A shows the execution mode of computer 301.This computer 301 can be desktop computer, or enterprise's memory device (such as server) of main frame, it is attached to storage array (such as Redundant Array of Independent Disks, storage router, edge router, storage switch and/or storage director).The user can utilize computer 301 to check static number image and/or video (for example a series of digital pictures).Often comprise on the computer 301 that multiple image and/or video-see program and/or media player program watch this image (comprising video) to allow the user.

Fig. 3 B shows the execution mode of notebook computer 302.Can find and use such notebook computer 302 under any situation in multiple situation.In recent years, along with the continuous growth of disposal ability and the function of notebook computer, formerly use this notebook computer of many situation down-samplings of high-end and more competent desktop computer.The same with computer 301, notebook computer 302 can comprise that each image-watching program and/or media player program are in order to allow the user to watch this image (comprising video).

Fig. 3 C shows the execution mode of high definition (HD) TV 303.Many HD TVs 303 comprise and allow to receive thereon, handle and the integrated tuner of decoded media content (for example television broadcasting signal).Replacedly, sometimes, HD TV 303 is from source (such as digital video disk (DVD) player, set-top box (the STB)) receiving media content of another reception, processing and decoding cable and/or satellite television broadcasting signal.No matter the specific implementation mode how, HD TV 303 can be implemented to carry out image and/or Video processing as described herein.In general, HD TV 303 has the ability that shows the HD media content and is implemented as the aspect ratio with 16:9 widescreen often.

Fig. 3 D shows the execution mode of single-definition (SD) TV 304.Certainly, SD TV 304 is similar to HD TV 303 to a certain extent, and at least one difference is that SD TV 304 does not comprise the ability that shows the HD media content, and SD TV 304 is implemented as the full frame aspect ratio with 4:3 often.However, or even SD TV 304 also can be implemented to carry out image and/or Video processing as described herein.

Fig. 3 E shows the execution mode of hand-held media units 305.Hand-held media units 305 can be operated for to for image/video content information (such as JPEG (joint photographic experts group) (JPEG) file, tagged image file format (TIFF), bitmap, Motion Picture Experts Group (MPEG) file, the Windows Media(WMA/WMV) file of user's playback, such as the video file of the other types of MPEG4 file etc.) and/or can carry out common store or storage with the other types information of number format storage.On historical viewpoint, so hand-held media units is mainly used in storage and plays back audio medium; Yet so hand-held media units 305 can be used for storage and any virtual media of playback (for example, audio frequency media, video media, image media etc.).And so hand-held media units 305 can also comprise other functions, such as the integrated communicaton circuit that is used for wired and radio communication.Hand-held media units 305 like this can be implemented to carry out image and/or Video processing as described herein.

Fig. 3 F shows the execution mode of set-top box (STB) 306.As mentioned above, STB306 can be implemented sometimes to receive, handle and decode and will offer cable and/or the satellite television broadcasting signal of any suitable demonstration available devices (such as SD TV 304 and/or HD TV 303).Such STB306 can operate in order to carry out image and/or Video processing as described herein with the demonstration available devices independently or synergistically.

Fig. 3 G shows the execution mode of digital video disk (DVD) player 307.Under the situation that does not deviate from scope and spirit of the present invention, such DVD player can be that blue-ray DVD player, HD can use DVD player, SD can use DVD player, up-sampling can use DVD player (for example from SD to HD etc.).DVD player can provide signal to any suitable demonstration available devices (such as SD TV 304 and/or HD TV 303).DVD player 305 can be implemented to carry out image and/or Video processing as described herein.

Fig. 3 H shows the execution mode of general digital image and/or video processing equipment 308.In addition, as mentioned above, above-mentioned these various devices do not comprise the equipment list of the limit that can realize image described herein and/or Video processing, be noted that, under the situation that does not deviate from scope and spirit of the present invention, can realize that any general digital image and/or video processing equipment 308 are in order to carry out image as herein described and/or Video processing.

Fig. 4, Fig. 5 and Fig. 6 show each execution mode 400 of video coding framework, 500 and 600 diagram respectively.

With reference to the execution mode 400 of Fig. 4, at this figure as can be seen, receive incoming video signal by video encoder.In some embodiments, incoming video signal is made of coding unit (CU) or macro block (MB).The variable size of this coding unit or macro block also can comprise common a plurality of pixels with arranged in squares.In one embodiment, the size of this coding unit or macro block is 16 * 16 pixels.Yet, be noted that usually macro block can have any required size, such as N * N pixel, wherein N is integer (for example, 16 * 16,8 * 8 or 4 * 4).Certainly, though adopted square coding unit or macro block in the preferred implementation, some implementation can comprise non-square coding unit or macro block.

Incoming video signal can be called as usually corresponding to primitive frame (or picture) view data.For example, primitive frame (or picture) view data can experience processing to generate brightness and chroma sample.In some embodiments, the luma samples group in the macro block is a specific arrangements (for example 16 * 16), and the chroma sample group is different specific arrangements (for example 8 * 8).According to execution mode described herein, video encoder is that unit handles these samples with the block-by-block.

Incoming video signal experiences model selection then, and incoming video signal experiences in frame according to this model selection ground and/or inter prediction is handled.In general, incoming video signal is along the compression of compression path experience.When not having feedback ground and operate (for example, not according to inter prediction, also not according to infra-frame prediction) thus incoming video signal provides experience map function (for example, according to discrete cosine transform (DCT)) via the compression path.Certainly, in replaceable execution mode, can adopt other conversion.Under this operator scheme, incoming video signal itself is exactly the signal that compresses.The compression path can utilize the sensitivity of human eye shortage high-frequency to compress.

Yet, by optionally using in the frame or inter prediction video coding, can feed back along the compression path.According to feedback or predicted operation pattern, the compression path produces (relatively low energy) redundant (for example difference) and operates deduct the current macro predicted value from current macro.According in given instance, adopting which kind of prediction form, generate current macro and at least based on the part of same number of frames (or picture) or at least based on redundancy or difference between the macroblock prediction value of the part of at least one other frame (or picture).

The vision signal of consequent modification is then along the map function of compression path experience.In one embodiment, discrete cosine transform (DCT) operates to calculate each the coefficient value in the basic mode formula of predetermined quantity to one group of video sample (for example, brightness, colourity, redundancy etc.).For example, an execution mode comprises 64 basic functions (for example, for 8 * 8 samples).In general, different execution modes can adopt the basic function (for example, different conversion) of varying number.The combination in any of these each basic functions (comprising basic function suitable and the selectivity weighting) can be used for one group of given video sample of expression.The extra details relevant with each mode of carrying out map function with technical literature that the video coding that comprises aforesaid those standard/draft standard of incorporating into by reference is associated in be described.The output of conversion process comprises coefficient value separately.This output is offered quantizer.

Generally speaking, most of image blocks can produce coefficient (for example, according to the DCT coefficient in the execution mode of discrete cosine transform (DCT) operation) usually, make that the frequency of most of associated DC T coefficients is lower.For this reason and human eye relatively poor relatively to the sensitivity of high frequency visual effect, quantizer can be operated for the coefficient that major part is not quite relevant and be converted to null value.That is to say, can eliminate those coefficients that the Relative Contribution rate is lower than certain predetermined value (for example certain threshold value) according to quantification treatment.Quantizer can also be operated the value that can encode more effectively than the value of change process generation for significant coefficient is converted to.For example, quantification treatment can by with each corresponding coefficient divided by integer value and abandon any remainder and operate.When typical coding unit or macro block are operated; this handles the nonzero coefficient can produce relatively small amount usually, and these nonzero coefficients are transferred to then that entropy coder carries out lossless coding and for according to selecting in the frame and/or the feedback path of inter prediction processing uses according to the video coding.

Entropy coder is handled according to lossless compression-encoding and is operated.By contrast, quantization operation is normally lossy.The entropy coding is handled the coefficient that quantification treatment is provided and is operated.Those coefficients can be represented each feature (for example, brightness, colourity, redundancy etc.).Entropy coder can adopt various types of codings.For example, entropy coder can carry out context adaptive binary arithmetic coding (CABAC) and/or CAVLC (CAVLC).For example, according at least a portion of entropy encoding scheme, data are converted to (operation, rank) pairing (run-level paring) (for example, with data 14,3,0,4,0,0 ,-3 are converted to each (operation, rank) to (0,14), (0,3), (1,4), (2 ,-3)).In advance, work out the table that the variable length code value of being dispensed to is right, so that it is right that relative shorter length coding is distributed to common relatively value, and it is right that relative longer length coding is distributed to rare relatively value.

As the reader understanding, the operation of re-quantization and inverse transformation corresponds respectively to the operation of quantification and conversion.For example, at the execution mode of DCT for map function, inverse DCT (IDCT) is the conversion of adopting in the inverse transformation operation.

Picture buffer (perhaps being called digital picture buffer or DPB) receives signal from the IDCT module; The picture buffer operation is used for storage present frame (or picture) and/or one or more other frames (or picture), such as operating the frame (or picture) that uses as the infra-frame prediction that carries out according to the video coding and/or inter prediction.Be noted that according to infra-frame prediction the storage of relatively small amount is just enough, because perhaps there is no need present frame (or picture) or any other frame (or picture) in storage frame (or picture) sequence.When carrying out inter prediction according to the video coding, institute's canned data can be used for carrying out motion compensation and/or estimation.

In a possible execution mode, for estimation, to compare from each buffering counter pair in other frames (or picture) in the corresponding luma samples group (for example, 16 * 16) of present frame (or picture) and frame (or picture) sequence (for example according to inter prediction).In a possible implementation, locate matching area (for example prediction reference) and produce vector shift (for example motion vector).In single frame (or picture), can find a plurality of motion vectors, but not every motion vector all must point to identical direction.One or more operating operation ground that carry out according to estimation are used for generating one or more motion vectors.

Operation of motion compensation ground adopts can be according to one or more motion vectors of estimation generation.Identification and the prediction reference group of paying sample be in order to deduct from original incoming video signal, attempts to wish to produce relatively (for example, desirable, a plurality of) lower energy redundancy.If this generic operation does not cause producing lower energy redundancy, just needn't carry out motion compensation, map function can only be operated original incoming video signal, and redundancy (is not for example operated, according to incoming video signal is directly offered map function, in order to do not carry out infra-frame prediction, also do not carry out the operator scheme of inter prediction), maybe can use infra-frame prediction and the redundancy that infra-frame prediction produces is carried out map function.Similarly, if estimation and/or operation of motion compensation success, motion vector can also be sent to entropy coder with corresponding redundant coefficient, is used for can't harm the entropy coding.

Output from the whole video encoding operation is the output bit stream.Be noted that the output bit stream can necessarily handle according to generating continuous time signal (this continuous time signal can via traffic channel) certainly.For example, some execution mode is operated in wireless communication system.In this case, the output bit stream can be used for generate and can carry out suitable digital-to-analogue conversion, frequency translation, convergent-divergent, filtration, modulation, sign map and/or any other operation via the Wireless Telecom Equipment of the continuous time signal of traffic channel etc.

With reference to the execution mode 500 of Fig. 5, at this figure as can be seen, incoming video signal is received by video encoder.In some embodiments, incoming video signal constitutes (and/or can be divided into coding unit (CU)) by coding unit or macro block.The size of coding unit or macro block can change and can comprise and be arranged to foursquare a plurality of pixel usually.In one embodiment, the size of coding unit or macro block is 16 * 16 pixels.Yet, be noted that usually macro block can have any required size, such as N * N pixel, wherein N is integer.Certainly, though adopted square coding unit or macro block in the preferred implementation, some implementation can comprise non-square coding unit or macro block.

Incoming video signal can be called as usually corresponding to primitive frame (or picture) view data.For example, primitive frame (or picture) view data can handle to generate brightness and chroma sample.In some embodiments, the luma samples group in the macro block is a specific arrangements (for example 16 * 16), and the chroma sample group is different specific arrangements (for example 8 * 8).According to execution mode described herein, video encoder is that unit handles these samples with the block-by-block.

Incoming video signal carries out model selection then, and incoming video signal optionally carries out in the frame according to this pattern and/or inter prediction is handled.In general, incoming video signal compresses along the compression path.When not operating (for example, not according to inter prediction, also not according to infra-frame prediction), incoming video signal provides to carry out map function (for example, according to discrete cosine transform (DCT)) via the compression path with having feedback.Certainly, in replaceable execution mode, can adopt other conversion.Under this operator scheme, incoming video signal itself is exactly the signal that compresses.The compression path can utilize the sensitivity of human eye shortage high-frequency to compress.

Yet, by optionally using in the frame or inter prediction video coding, can feed back along the compression path.According to feedback or predicted operation pattern, the compression path causes (relatively low energy) redundant (for example difference) to be operated to deduct the current macro predicted value from current macro.According in given instance, adopting which kind of prediction form, generate current macro and at least based on the part of same number of frames (or picture) or at least based on redundancy or difference between the macroblock prediction value of the part of at least one other frame (or picture).

Consequent modification vision signal is carried out map function along the compression path then.In one embodiment, discrete cosine transform (DCT) operates to calculate each the coefficient value in the basic mode formula of predetermined quantity to one group of video sample (for example, brightness, colourity, redundancy etc.).For example, an execution mode comprises 64 basic functions (for example, for 8 * 8 samples).In general, different execution modes can adopt the basic function (for example, different conversion) of varying number.The combination in any that comprises these basic functions of suitable selectivity weighting can be used for one group of given video sample of expression.The extra details relevant with each mode of carrying out map function with technical literature that the video coding that comprises aforesaid those standard/draft standard of incorporating into by reference is associated in be described.The output of conversion process comprises coefficient value separately.This output is offered quantizer.

Generally speaking, most of image blocks can produce coefficient (for example, according to the DCT coefficient in the execution mode of discrete cosine transform (DCT) operation) usually, make that the frequency of most of associated DC T coefficients is lower.For this reason and human eye relatively poor relatively to the sensitivity of high frequency visual effect, quantizer can be operated for the coefficient that major part is not quite relevant and be converted to null value.That is to say, can eliminate those coefficients that Relative Contribution is lower than certain predetermined value (for example certain threshold value) according to quantizing process.Quantizer can also be operated the value that can encode more effectively than the value of change procedure generation for significant coefficient is converted to.For example, the coefficient that quantizing process can be by separately is divided by integer value and abandon any remainder and operate.When typical coding unit or macro block are operated; this process can produce the nonzero coefficient of relatively small amount usually, and these nonzero coefficients are transferred to entropy coder then to carry out lossless coding and use according to the feedback path that video coding is selected in the frame and/or inter prediction is handled.

Entropy coder is operated according to the lossless compression-encoding process.By contrast, quantization operation normally diminishes.The entropy cataloged procedure is operated the coefficient that quantizing process provides.Those coefficients can be represented each feature (for example, brightness, colourity, redundancy etc.).Entropy coder can adopt various types of codings.For example, entropy coder can carry out context adaptive binary arithmetic coding (CABAC) and/or CAVLC (CAVLC).For example, according at least a portion of entropy encoding scheme, data are converted to (operation, rank) pairing (run-level paring) (for example, with data 14,3,0,4,0,0 ,-3 are converted to each (operation, rank) to (0,14), (0,3), (1,4), (2 ,-3)).In advance, work out the table that the variable-length code (VLC) value of being dispensed to is right, so that it is right that relative shorter length coding is distributed to common relatively value, and it is right that relative longer length coding is distributed to rare relatively value.

As the reader understanding, the operation of re-quantization and inverse transformation corresponds respectively to the operation of quantification and conversion.For example, at the execution mode of DCT for map function, inverse DCT (IDCT) is the conversion of adopting in the inverse transformation operation.

Self-adaption loop filter (ALF) is implemented to handle the output from inverse transform block.Be stored in picture buffer (being sometimes referred to as DPB, digital picture buffer) before at decoding picture, self-adaption loop filter (ALF) is applied to decoding picture.Self-adaption loop filter (ALF) is implemented to reduce the coding noise of decoded picture, can optionally filter brightness and colourity respectively by ground, no matter whether self-adaption loop filter (ALF) is with the sheet rank or with the piece level applications.In the application of self-adaption loop filter (ALF), can use two-dimentional 2-D finite impulse response (FIR) (FIR) to filter.The coefficient of filter can design in encoder piecewise, then this information is passed to decoder (for example, pass to from the reflector communication equipment that comprises video encoder (perhaps being called encoder) comprise the Video Decoder receiver communication equipment of (perhaps being called decoder)).

An execution mode is by operating according to Wiener filtering design generation coefficient.In addition, handle also whether according to quad-tree structure this decision (is for example passed to decoder no matter whether carry out filtering, pass to from the reflector communication equipment that comprises video encoder (perhaps being called encoder) and to comprise the Video Decoder receiver communication equipment of (perhaps being called decoder)), can use one by one in encoder, wherein block size decides according to rate-distortion optimization.Be noted that the implementation of utilizing 2D filtering can introduce complexity according to Code And Decode.For example, by using 2D filtering according to self-adaption loop filter (ALF) and implementation, in encoder (it is realized) and decoder (it is realized), may there be the complexity of certain increase in the reflector communication equipment in the receiver communication equipment.

In some optional execution mode, will offer from the output of de-blocking filter and be implemented as processing from filter (for example realizing according to self-adaption loop filter (ALF), sampled point self adaptation skew (SAO) filter and/or any other filter type) in one or more other rings of the output of inverse transform block.For example, before, ALF is applied to decoding picture to be stored in picture buffer (being sometimes referred to as DPB, digital picture buffer) at decoding picture.ALF is implemented to reduce the coding noise of decoded picture, can selectively filter brightness and colourity respectively piecewise, no matter whether ALF is with the sheet rank or with the piece level applications.In the application of ALF, can use two-dimentional 2-D finite impulse response (FIR) (FIR) to filter.The coefficient of filter can design in encoder piecewise, then this information is passed to decoder (for example, pass to from the reflector communication equipment that comprises video encoder (perhaps being called encoder) comprise the Video Decoder receiver communication equipment of (perhaps being called decoder)).

An execution mode is used for generating coefficient according to the Wiener filtering design operation.In addition, handle also whether according to quad-tree structure this decision (is for example passed to decoder no matter whether carry out filtering, pass to from the reflector communication equipment that comprises video encoder (perhaps being called encoder) and to comprise the Video Decoder receiver communication equipment of (perhaps being called decoder)), can use block by block in encoder, wherein block size decides according to rate-distortion optimization.Be noted that the implementation of utilizing 2D filtering can introduce complexity according to Code And Decode.For example, by using 2D filtering according to ALF and implementation, in encoder (it is realized) and decoder (it is realized), may there be the complexity of certain increase in the reflector communication equipment in the receiver communication equipment.

As described at other execution modes, use ALF that one of a series of improvement according to this Video processing can be provided, comprise by quantizing peak signal noise ratio (PSNR) the improvement objective quality measurement that denoising causes at random.In addition, the subjective quality of Bian Ma vision signal can be realized by illumination compensation subsequently, handles according to ALF, and illumination compensation can be introduced according to carrying out migration processing and convergent-divergent processing (for example, according to using gain).

For filter in one type the ring, use self-adaption loop filter (ALF) that one of a series of improvement according to this Video processing can be provided, comprise by quantizing peak signal noise ratio (PSNR) the improvement objective quality measurement that denoising causes at random.In addition, the subjective quality of Bian Ma vision signal can be realized by illumination compensation subsequently, handles according to self-adaption loop filter (ALF), and illumination compensation can be introduced according to carrying out migration processing and convergent-divergent processing (for example, according to using gain).

Picture buffer (perhaps being called digital picture buffer or DPB) receives from the signal of ALF output; The picture buffer operation is used for storage present frame (or picture) and/or one or more other frames (or picture), operates the frame (or picture) that uses such as what carry out according to the video coding according to infra-frame prediction and/or inter prediction.Be noted that according to infra-frame prediction the storage of relatively small amount is just enough, because perhaps there is no need any other frame (or picture) or present frame (or picture) in storage frame (or picture) sequence.When carrying out inter prediction according to the video coding, institute's canned data can be used for carrying out motion compensation and/or estimation.

In a possible execution mode, for estimation, to compare from each buffering counter pair in other frames (or picture) in the corresponding luma samples group (for example, 16 * 16) of present frame (or picture) and frame (or picture) sequence (for example according to inter prediction).In a possible implementation, locate matching area (for example prediction reference) and produce vector shift (for example motion vector).In single frame (or picture), can find a plurality of motion vectors, but not every motion vector all must point to identical direction.One or more operating operation ground that carry out according to estimation generate one or more motion vectors.

Operation of motion compensation ground adopts can be according to one or more motion vectors of estimation generation.Identification and the prediction reference group of paying sample be in order to deduct from original incoming video signal, attempts to wish to produce relatively (for example desirable a plurality of) lower energy redundancy.If this generic operation does not cause producing lower energy redundancy, just needn't carry out motion compensation, map function can only be operated original incoming video signal, and redundancy (is not for example operated, according to incoming video signal is directly offered map function, in order to do not carry out infra-frame prediction, also do not carry out the operator scheme of inter prediction), maybe can use infra-frame prediction and the redundancy that infra-frame prediction produces is carried out map function.Similarly, if estimation and/or operation of motion compensation success, motion vector can also be sent to entropy coder with corresponding redundant coefficient, is used for can't harm the entropy coding.

Output from the whole video encoding operation is the output bit stream.Be noted that the output bit stream can necessarily handle according to generating continuous time signal (its can via traffic channel) certainly.For example, some execution mode is operated in wireless communication system.In this case, the output bit stream can carry out suitable digital-to-analogue conversion, frequency translation, convergent-divergent, filtration, modulation, sign map and/or any other operation at Wireless Telecom Equipment (its be used for to generate can via the continuous time signal of traffic channel etc.).

With reference to the execution mode 600 of Fig. 6, at the replaceable execution mode that this Figure illustrates a kind of video encoder, this video encoder is predicted, is changed and encodes and handle in order to produce the output bit stream of compression.This video encoder can be according to one or more video coding protocols, standard and/or recommended practice (such as the 10th part of ISO/IEC14496-10-MPEG-4, the AVC(advanced video coding) (perhaps be called H.264/MPEG-4 the 10th part or AVC(advanced video coding), ITU is H.264/MPEG4-AVC) operate and compatible with it.

Be noted that, corresponding Video Decoder (such as the Video Decoder of the equipment that is positioned at the communication channel other end) operation is used for decoding, this sequence (ideally) expression incoming video signal is handled in order to produce corresponding decoding video sequence in the complementation of inverse transformation and reconstruction.

In the time will scheming to compare with previous figure, the signal path output from re-quantization and inverse transformation (for example IDCT) piece that will offer intra-frame prediction block equally offers de-blocking filter.To offer from the output of de-blocking filter and be implemented as processing from filter (for example realizing according to self-adaption loop filter (ALF), sampled point self adaptation skew (SAO) filter and/or any other filter type) in one or more other rings of the output of inverse transform block.For example, in a possible execution mode, before, the SAO filter is applied to decoding picture to be stored in picture buffer (being sometimes referred to as DPB, digital picture buffer) at decoding picture.

With regard to any video encoder framework that is implemented to generate carry-out bit stream, be noted that in any one that this class framework can be in multiple communication equipment and realize.The output bit stream can experience extra processing (comprising self-correcting code (ECC), forward error correction (FEC) etc.), thereby generates the modification output bit stream that wherein has extra redundancy transaction.Equally, can be applicable to or the continuous time signal that is suitable for transmitting via communication channel carries out any suitable processing according to generation as being appreciated that about this digital signal.That is to say that such video encoder framework can be realized via the communication equipment of the one or more signals of one or more traffic channel being used for.Can carry out extra processing to the carry-out bit stream that is generated by this video encoder framework, thereby generate the continuous time signal that can be transmitted in the communication channel.

Fig. 7 shows the diagram of the execution mode 700 of intra-prediction process.At this figure as can be seen, the current block of video data (for example be square usually and generally include N * N pixel) is handled in order to wherein each pixel estimated.Adopt the prior encoded pixels on the top and the left side that are positioned at current block according to infra-frame prediction.From certain angle, intra prediction direction can be regarded as the vector corresponding to the reference pixel that extends to the top that is positioned at current pixel or the left side from current pixel.Details according to the infra-frame prediction that H.264/AVC is applied to encode have carried out (for example stipulating in the corresponding standard of above incorporating into by reference, International Telecommunications Union, ITU-T, TELECOMMUNICATION STANDARDIZATION SECTOR OF ITU, H.264 (03/2010), SERIES H:AUDIOVISUAL AND MULTIMEDIA SYSTEMS, Infrastructure of audiovisual services – Coding of moving video, Advanced video coding for generic audiovisual services, H.264(or be called the 10th part of International Telecomm ISO/IEC14496-10 – MPEG-4 Recommendation ITU-T, the AVC(advanced video coding), H.264/MPEG-4 the 10th part or AVC(advanced video coding), ITU H.264/MPEG4-AVC or equivalent document).

Redundant (it is the difference between current pixel and reference or the predict pixel) is the redundancy through coding.At this figure as can be seen, infra-frame prediction utilizes the pixel in the common frame (or picture) to operate.Certainly be noted that given pixel can have different component separately associated with it, and may there be separately different sample groups in each component.

Fig. 8 shows the diagram of the execution mode 800 of inter prediction processing.Different with infra-frame prediction, inter prediction is used for identifying motion vector (for example, inter prediction direction) based on the current pixel groups in the present frame (or picture) and one or more groups reference or the predict pixel that are positioned at one or more other frames (or picture) of frame (or picture) sequence.As can be seen, motion vector extends to another frame (or picture) from the present frame (or picture) in frame (or picture) sequence.Inter prediction can use sub pixel interpolation, makes the function of predicted pixel values corresponding to a plurality of pixels in reference frame or the picture.

Can handle to calculate redundancy according to inter prediction, the redundancy of even now is different from the redundancy of calculating according to intra-prediction process.Handle according to inter prediction, the redundancy of each pixel is poor corresponding between current pixel and the predicted pixel values again.Yet, to handle according to inter prediction, current pixel and reference or predict pixel are not positioned at same number of frames (or picture).Although the figure shows the inter prediction that adopts about one or more previous frames or picture, be noted that equally replaceable execution mode can utilize corresponding to before the present frame and/or frame afterwards operate.For example, according to suitable buffering and/or memory management, can store a plurality of frames.When given frame is operated, can according to before given frame and/or other frames afterwards generate reference.

In conjunction with CU, elementary cell can be used for prediction partition mode (that is, predicting unit or PU).Be noted that equally only to be last degree of depth CU restriction PU, and corresponding size is confined to the size of CU.

Fig. 9 and Figure 10 show each execution mode 900 of video decode framework and 1000 diagram respectively.

In general, this video decode framework is operated incoming bit stream.Certainly be noted that this incoming bit stream can generate from the signal that communication channel receives according to communication equipment.Can carry out various operations to the continuous time signal that receives from communication channel, comprise such as can be suitably according to the digital sample of generation incoming bit stream, demodulation, convergent-divergent, filtering etc.And, can realize that some execution mode of one or more self-correcting codes (ECC), forward error correction (FEC) etc. can carry out suitable decoding according to ECC, FEC etc., thereby generate incoming bit stream.That is to say, (for example exporting bit stream according to generating correspondence, the output bit stream that can launch from the transmitter portion of reflector communication equipment or transceiver communications equipment) carries out some execution mode of extra redundancy, can carry out suitable processing according to generating incoming bit stream.On the whole, such video decode framework regrettably will be handled incoming bit stream, thereby as far as possible closely and fully in the ideal case generate the outputting video signal corresponding to original incoming video signal, be used for exporting to one or more videos and show available devices.

With reference to the execution mode 900 of Fig. 9, in general, handle incoming bit stream such as the decoder of entropy decoder (for example, its can according to realizations such as CABAC, CAVLC) according to the complementation of encode (as in the video encoder framework, carrying out).Incoming bit stream can be regarded as the compression output bit stream that (as far as possible closely and fully in the ideal case) generated by the video encoder framework.Certainly, in actual applications, in the signal via one or more communication link transmission, may suffer a part of error.Entropy decoder handles and extracts suitable coefficient to incoming bit stream, such as DCT coefficient (for example, information such as expression colourity, brightness), and coefficient offered re-quantization and inverse transform block.If the employing dct transform, re-quantization and inverse transform block just can be implemented to carry out inverse DCT (IDCT) operation so.Subsequently, the A/D block filter is implemented each frame and/or the picture that become next life corresponding to outputting video signal.These frames and/or picture can be offered picture buffer or digital picture buffer (DPB), in order to be used for comprising other operations of motion compensation.In general, such operation of motion compensation can be regarded as the inter prediction that is associated corresponding to the video coding.Equally, can also carry out inter prediction to the signal from re-quantization and inverse transform block output.Be similar to the video coding, this Video Decoder framework can be implemented to carry out model selection, according to incoming bit stream is decoded, is not undertaken by infra-frame prediction or inter prediction, undertaken or undertaken by infra-frame prediction by inter prediction, thereby generate outputting video signal.

Execution mode 1000 with reference to Figure 10, in some optional execution mode, such as filter in the one or more rings that can encode to realize according to the video that be used for to generate the output bit stream (for example, realize according to self-adaption loop filter (ALF), sampled point self adaptation skew (SAO) filter and/or any other filter type), filter can be realized in the Video Decoder framework in corresponding one or more rings.In one embodiment, after de-blocking filter, suitably realize filter in one or more rings.

According to some possible execution mode, (for example, according to the SAO filtering that realizes in the filter in other rings in Fig. 6) can carry out sampled point self adaptation skew (SAO) and handle after the block elimination filtering of finishing decoding picture is handled.Carry out this processing based on the zone that is defined as one or more complete maximum coding units (LCU).

Figure 11 shows the execution mode 1100 that band bending sampled point self adaptation skew (SAO) filtering is handled.The figure shows the concept of band bending SAO.After using each skew, consequent pixel is tailored to effective 8 pixel coverages [0,255].In the figure, with offset applications to four continuous, an active band; Do not revise remaining band.Certainly, in other embodiments, these skews can be applied to discontinuous band.

Figure 12 shows the replaceable execution mode 1200 of video coding framework.In this execution mode 1200, any one or the interior filter of a plurality of other rings are (for example, the interior filter of ring according to self-adaption loop filter (ALF), sampled point self adaptation skew (SAO) filter and/or the realization of any other filter type) can be implemented as the output of handling re-quantization and inverse transform block (for example, before de-blocking filter).In other words, in these execution modes, going piece can use filter in one or more other rings (for example, the SAO filter in execution mode) before handling.In replaceable execution mode, before going the piece processing, can realize filter in these rings (for example, the interior filter of realizing according to self-adaption loop filter (ALF), sampled point self adaptation skew (SAO) filter and/or any other filter type of ring).Yet, before going the piece processing, each aspect of the present invention, execution mode and/or its equivalent (are for example operatively used the interior filter of these rings, according to filter in the ring of self-adaption loop filter (ALF), sampled point self adaptation skew (SAO) filter and/or the realization of any other filter type), as shown in figure 12.

According to some execution mode, the false shadow of some unacceptable bulks may appear in (for example, at the execution mode of realizing this SAO from the output of going piece to handle) when SAO can operate and open.In these execution modes, mainly be because two adjacent LCU are using different band deviants.In order to alleviate this problem, going piece can use this SAO before handling, and removing piece to handle to be used under these situations, to reduce any not desirable and block false shadow of occurring.In this case, go the boundary intensity that uses in the piece processing procedure and variable β and tC also to be determined by the SAO parameter.

From some angle, it is correcting filter (for example, the histogram correcting filter in some execution mode) that band skew SAO can be considered basically.Come classified pixels to produce the distribution of pixel based on intensity level.For example, at the histogram execution mode, based on intensity level pixel (for example, the pixel of one or more maximum coding units (LCU)) is classified as histogram bin or " band ".Whole pixel coverage (0-255) is divided into 32 bands uniformly, and particular offset is added on all pixels in each band.The skew that encoder will be used from scope [7,7] selection.

Although skew can be applicable to whole 32 bands, in order to simplify the band migration processing and to reduce expense, can utilize the reduced set of band skew SAO among any LCU (for example, only 4 continuous bands) to carry out actual modification.Encoder is selected four continuous bands, and skew is transmitted at these four continuous bands.Do not revise remaining 28 bands (zero offset).Owing to may there be 32 bands, first band that therefore has non-zero offset is represented with bit stream.The band_position parameter carries this information.Remaining three active bands can be determined by (band_position+i) %32 of i ∈ [1,3].Note modular arithmetic herein, if first band is 29,30 or 31, remaining band is with wraparound 0.

Figure 13 shows each execution mode 1300 of the indication (self adaptation and/or explicit signaling) of the transmission band skew under sampled point self adaptation skew (SAO) the band modes of deflection.Can finish such operation according to the quantity that indicates the transmission band skew under the SAO band bending pattern adaptively.For example, the quantity of the transmission band under SAO band bending pattern skew may be relevant with the size of LCU (quantity that for example, makes transmission band under the SAO band bending pattern be offset can be the function of LCU size).For example, if the LCU size reduces, the quantity of transmission band also reduces.For another example, 4 transmission band skews can be used for 64 * 64LCU, and 3 transmission band skews can be used for 32 * 32LCU, and 2 transmission band skews can be used for 16 * 16LCU.Generally speaking, according to LCU different size separately, can point out the different quantity separately of the transmission band skew under the SAO band bending pattern.

The quantity of the transmission band of each LCU size skew also can be according to the SPS(sequence parameter set), PPS(image parameters size), APS(auto-adaptive parameter collection), macro block bar head, LCU data and/or use other part explicitlies to notify with signal.

Figure 14 shows each execution mode 1400 of the indication (self adaptation and/or explicit signaling) of the frequency band granularity under the SAO band bending pattern.Can finish such operation according to the granularity that indicates adaptively under the SAO band bending pattern.

In some embodiments, whole pixel coverage (0-255) is divided into 32 uniform frequency bands.Can utilize the band bending SAO among any LCU that 4 frequency bands are only carried out actual modification.Encoder (for example, the reflector communication equipment) is selected four continuous frequency bands, and skew is transmitted at these four continuous frequency bands.Do not revise remaining 28 frequency bands (for example, zero offset).In each frequency band, particular offset is added on all pixels.

Because can change the size (for example, 64 * 64,32 * 32 or 16 * 16) of LCU, the granularity of frequency band can be adaptive.For example, the size of LCU is more little, and granularity is just more coarse.For another example, if LCU is of a size of 32 * 32, then [0,255] evenly can be divided into 16 frequency bands, each frequency band is contained 16 continuous intensity levels.In general, can according to LCU separately different size indicate different frequency band granularity separately under the SAO band bending pattern.

The frequency band granularity of each LCU size also can be according to the SPS(sequence parameter set), PPS(image parameters size), APS(auto-adaptive parameter collection), macro block bar head, LCU data and/or use other parts clearly to notify with signal.

Figure 15 shows the execution mode 1500 of implicit expression band index signaling.For example, can infer based on the pixel value of current LCU (for example, determine based on the analysis of LCU, determine with inferring, etc.) this information, and do not have explicitly signal notice band index.For example, by generating the pixel value histogram of LCU, band bending can be used for the prevailing frequency band of quantity of pixel.These band indexs are continuous (for example, band index can be so that have discontinuous distribution, so that at least two continuous band indexs at least one band index value at interval, in other words, band index is not necessarily continuous each other) not necessarily.

Figure 16 shows the replaceable execution mode 1600 of implicit expression band index signaling.In the utmost point reduced graph that shows utmost point simplification execution mode, there is the LCU that only has two gray scales.Histogram (for example, possible a kind of mode of the description pixel distribution of understanding as the reader) illustrates that the pixel that 50% pixel has a gray scale 25,50% has gray scale 205.Therefore, two band bendings are just enough, replaced original four.

Figure 17 shows the execution mode 1700 of band bending coding.Under the band bending pattern, because sao_band_position represents to have the beginning of the band bending of non-zero offset, therefore the first deviant sao_offset[cldx] [saoDepth] [x0] [y0] [0] (for example be necessary for non-zero, in some cases, the value of possible minimum can be to get 1 value).So, and direct coding sao_offset[cldx not] [saoDepth] [x0] [y0] [0], bs (sao_offset[cldx] [saoDepth] [x0] [y0] [0]) 1 and sao_offset[cldx] sign bit of [saoDepth] [x0] [y0] [0] can encode separately, and wherein abs is the function that calculates absolute value.

Figure 18 and Figure 19 show each execution mode of the method for operation of one or more equipment (for example, communication equipment, receiver and/or decoder apparatus, reflector and/or encoder device etc.).

With reference to the method 1800 of Figure 18, method 1800 is at first passed through the input of communication equipment from least one additional communication equipment receiving video signals and a plurality of band bending, shown in square frame 1810.

Method 1800 continues to analyze a plurality of pixels that are associated with at least one maximum coding unit (LCU) (this maximum coding unit is associated with vision signal) and distributes with the pixel value that identification is used for a plurality of band indexs of identification, shown in square frame 1820.

Method 1800 is handled according to vision signal or based on this filtering of signal then a plurality of band bendings is applied to a plurality of band indexs, shown in square frame 1830.

With reference to the method 1900 of Figure 19, this method 1900 is at first passed through the input of communication equipment from least one additional communication equipment receiving video signals and a plurality of band bending, shown in square frame 1910.

Method 1900 continues to analyze a plurality of pixels that are associated with at least one maximum coding unit (LCU) (this maximum coding unit is associated with vision signal) and distributes with the pixel value that identification is used for a plurality of band indexs of identification, shown in square frame 1920.

Method 1900 continues vision signal or carries out sampled point self adaptation skew (SAO) filtering based on this signal and handle to generate first filtering signal, makes SAO filtering processing comprise a plurality of band bendings are applied to a plurality of band indexs, shown in square frame 1930.

Method 1900 continues that first filtering signal is carried out block elimination filtering and handles to generate second filtering signal, shown in square frame 1940.

Be noted that baseband processing module and/or processing module and/or the miscellaneous part wherein wherein realized such as utilizing equally, carry out in the literary composition at each method described each operation and function in any one that can be in multiple communication equipment.For example, this baseband processing module and/or processing module can generate these signals and carry out these operations as described herein, processing etc., can also carry out each operation and analysis as described herein, or any other operation as described herein and function etc., or equivalent separately.

In some embodiments, this baseband processing module and/or processing module (it can be realized in identical device or distinct device) can be carried out this processing according to each aspect of the present invention, operation etc., and/or any other operation as described herein and function etc., or equivalent separately.In some embodiments, this processing is carried out collaboratively by first processing module in first equipment and second processing module in second equipment.In other embodiments, this processing, operation etc. are all carried out by the baseband processing module in the given equipment and/or processing module.In addition other execution modes in, use at least the first processing module and second processing module in the individual equipment to carry out this processing, operation etc.

Equally, as used herein term " basically " and " approx " provide the tolerance of industry reception and/or the correlation between article for its corresponding term.The scope of the tolerance that industry receives is less than 1%-50% and corresponding to (but being not limited to) component value, integrated circuit technology variable, temperature variable, rising and fall time and/or thermal noise.The difference of the correlation between the article be several percentage points to the order of magnitude.As the same employed term of this paper " operationally be coupled to ", " being coupled to " and/or " coupling " comprise directly couple between the article and/or article between (for example couple indirectly by article between two parties, article include, but is not limited to parts, element, circuit and/or module), wherein, for coupling indirectly, article are not revised signal message but can be regulated its current level, voltage level and/or power level between two parties.Further employed supposition couples (that is, when an element couple with another element by inferring) and comprises in the mode identical with " being coupled to " and directly and indirectly coupling between two article as this paper.Further employed term " can operate for " or " operationally being coupled to " show that article comprise one or more power supply connections, input, output etc. in order to carry out one or more corresponding functions and the supposition that may further include with one or more other article couples when activating as this paper.As this paper further employed term " with ... be associated " comprise directly and/or indirectly coupling of the stand-alone item that embeds in another article and/or article.As used herein term " with ... match in excellence or beauty " show and relatively provide required relation between two or more article, the signal etc.For example, when to be signal 1 bigger than the order of magnitude of signal 2 in described pass, when the order of magnitude of signal 1 is bigger than the order of magnitude of signal 2 or when the order of magnitude of signal 2 than the order of magnitude of signal 1 hour, can realize favourable comparison.

Can be single treatment facility or a plurality for the treatment of facility as the same employed term of this paper " processing module ", " module ", " treatment circuit " and/or " processing unit " (for example, comprise to operate, realize and/or be used for the various modules of coding, decoding, Base-Band Processing and/or circuit etc.).This treatment facility can be microprocessor, microcontroller, digital signal processor, microcomputer, central processing unit, field programmable gate array, programmable logic device, state machine, logical circuit, analog circuit, digital circuit and/or based on the hard coded of circuit and/or any equipment of operational order operation signal (simulation and/or numeral).Processing module, module, treatment circuit and/or processing unit can have memory and/or the integrated memory element that is associated, and this memory and/or integrated memory element can be embedding circuit, module, treatment circuit and/or the processing units of single memory device, a plurality of memory device and/or processing module.This memory device can be any apparatus of read-only memory (ROM), random-access memory (ram), volatile memory, nonvolatile memory, static memory, dynamic memory, flash memory, cache memory and/or storing digital information.Be noted that, if processing module, module, treatment circuit and/or processing unit comprise more than one treatment facility, so treatment facility just can centralized positioning (for example, directly be coupled in together by wired and/or wireless bus structure) or can Distributed localization (for example, coupling cloud computing indirectly by what local area network (LAN) and/or wide area network carried out).Be furthermore noted that, if processing module, module, treatment circuit and/or processing unit are realized one or more functions by state machine, analog circuit, digital circuit and/or logical circuit, can will store that the memory of corresponding operating instruction and/or memory component are embedded in the circuit that comprises state machine, analog circuit, digital circuit and/or logical circuit or outside.Be furthermore noted that memory component can be stored, processing module, module, treatment circuit and/or processing unit are carried out hard coded and/or the operational order corresponding at least a portion step shown in one or more figure and/or function.Can include sort memory equipment or memory component in the goods.

Above invention has been described under the help of the method step of the performance that shows specified function and relation thereof.For convenience of description, this paper defines boundary and the order of these formation function pieces and method step arbitrarily.As long as suitably carry out appointed function and relation, just can limit and substitute boundary and order.Any alternative boundary and order are all in claimed scope and spirit scope of the present invention.In addition, for convenience of description, define the boundary of these formation function pieces arbitrarily.As long as suitably carry out some critical function, just can limit alternative boundary.Similarly, this paper also defines FB(flow block) arbitrarily in order to some critical function is shown.From used degree, stipulated FB(flow block) boundary and order in addition, and still carried out some critical function.The alternative definitions of formation function piece and flow chart block and order is in claimed scope and spirit scope of the present invention.Those of ordinary skill in the art will understand that also formation function piece herein and other illustrative pieces, module and parts may be implemented as and utilize processor or its combination in any of discreet component, application-specific integrated circuit (ASIC), the suitable software of execution etc. to illustrate.

May invention has been described at one or more execution modes at least in part.Embodiments of the present invention are used for illustrating the present invention in this article, its one side, its characteristics, its concept and/or the example.The physical embodiments that embodies device of the present invention, goods, machine and/or technology can comprise one or more aspects of describing with reference to one or more execution modes of discussing in the literary composition, characteristics, concept, example etc.In addition, from figure to figure, execution mode can be incorporated function, step, module of the identical or similar name that can use identical or different Ref. No. etc. into, just because of this, described function, step, module etc. can be that identical or similar function, step, module etc. maybe can be different function, step, module etc.

Unless specify from the negative, pass to the signal of the element among any one shown figure herein, can be analog signal or digital signal, continuous time signal or discrete-time signal and single-ended signal or differential signal from signal and/or the signal between the element of this element.For example, if signal path is shown as single-ended path, then also represent differential signal path.Similarly, if signal path is shown as differential path, then also represent single ended signal paths.Although this paper is described one or more certain architectures, but can realize other frameworks equally, other frameworks use coupling indirectly between one or more data/address buss (clearly not illustrating), the direct connection between the element and/or other elements of approving as those of ordinary skill in the art.

Term " module " is used for each execution mode of the present invention is described.Module comprises via the hardware realization in order to carry out one or more functions of modules, such as handling one or more input signals to produce the functional block of one or more output signals.The hardware of realization module itself can be operated in conjunction with software and/or firmware.As used herein module can comprise one or more submodules, and each submodule itself is exactly module.

Although this paper has clearly described the particular combination of each function of the present invention and characteristics, other combinations of these characteristics and function are possible equally.The present invention is not subjected to the restriction of particular instance disclosed herein and clearly makes up in conjunction with other.

Claims (10)

1. device comprises:

At least one input is used for:

From at least one extra means receiving video signals; And

Receive a plurality of band bendings via signaling from described at least one extra means; And

Processor is used for:

Analyze a plurality of pixels relevant with at least one maximum coding unit (LCU) and distribute with the identification pixel value, wherein, described maximum coding unit is relevant with described vision signal;

Identify a plurality of band indexs inferentially based on described pixel value distribution;

Handle according to described vision signal or based on the filtering of the signal of described vision signal described a plurality of band bendings are applied to described a plurality of band index;

Handle to generate first filtering signal to described vision signal or based on signal execution sampled point self adaptation skew (SAO) filtering of described vision signal, wherein, described sampled point self adaptation skew filtering is handled described a plurality of band bendings is applied to described a plurality of band index; And

Described first filtering signal is carried out block elimination filtering handle to generate second filtering signal.

2. device according to claim 1, wherein, described processor is used for:

Analyze a plurality of pixels that are associated with at least one maximum coding unit and represent the pixel value histogram that described pixel value distributes with generation, wherein, described maximum coding unit is relevant with described vision signal; And

Based on described pixel value histogram, identification will be employed described a plurality of band indexs of described a plurality of band bendings.

3. device according to claim 1, wherein:

Described a plurality of band index has discontinuously arranged, makes at least two continuous band indexs at least one band index value separated from one another in described a plurality of band index.

4. device according to claim 1, wherein:

The described pixel value indication a plurality of subclass of the described a plurality of pixels relevant with at least a portion in described a plurality of band indexs respectively that distribute; And

Be employed described a plurality of band indexs of described a plurality of band bendings corresponding at least one subclass in a plurality of subclass of described a plurality of pixels, wherein, other subclass in a plurality of subclass of this at least one subclass and described a plurality of pixels are compared the pixel with relatively large or maximum quantity.

5. device according to claim 1, wherein:

Described device is the communication equipment of operating in satellite communication system, wireless communication system, wired communication system, optical fiber telecommunications system and mobile communication system at least one.

6. device comprises:

Input is used for from least one extra means receiving video signals and a plurality of band bending; And

Processor is used for:

Analyze a plurality of pixels relevant with at least one maximum coding unit (LCU) and distribute with the pixel value that identification is used for a plurality of band indexs of identification, described maximum coding unit is relevant with described vision signal; And

Handle according to described vision signal or based on the filtering of the signal of described vision signal described a plurality of band bendings are applied to described a plurality of band index.

7. device according to claim 6, wherein, described processor is used for:

Handle to generate first filtering signal to described vision signal or based on signal execution sampled point self adaptation skew (SAO) filtering of described vision signal, wherein, described sampled point self adaptation skew filtering is handled described a plurality of band bendings is applied to described a plurality of band index; And

Described first filtering signal is carried out block elimination filtering handle to generate second filtering signal.

8. device according to claim 6, wherein, described processor is used for:

Analyze a plurality of pixels relevant with at least one maximum coding unit to generate the pixel value histogram that the described pixel value of expression distributes, wherein, described maximum coding unit is relevant with described vision signal; And

Based on described pixel value histogram, identification will be employed a plurality of band indexs of described a plurality of band bendings.

9. device according to claim 6, wherein:

Described a plurality of band index has discontinuously arranged, makes at least two continuous band indexs at least one band index value separated from one another in described a plurality of band index.

10. the method for operation of a communication equipment, described method comprises:

Via the input of described communication equipment, from least one additional communication equipment receiving video signals and a plurality of band bending;

Analyze a plurality of pixels relevant with at least one maximum coding unit (LCU) and distribute with the pixel value that identification is used for a plurality of band indexs of identification, wherein, described maximum coding unit is relevant with described vision signal; And

Handle according to described vision signal or based on the filtering of the signal of described vision signal described a plurality of band bendings are applied to described a plurality of band index.

Images