CN113170178A

CN113170178A - Modification of the construction of a merging candidate list

Info

Publication number: CN113170178A
Application number: CN201980074650.1A
Authority: CN
Inventors: 王祥林; 陈漪纹
Original assignee: Beijing Dajia Internet Information Technology Co Ltd
Current assignee: Beijing Dajia Internet Information Technology Co Ltd
Priority date: 2018-09-12
Filing date: 2019-09-12
Publication date: 2021-07-23
Also published as: WO2020056143A1

Abstract

A computing device performing a method of decoding video data by generating a merge list, comprising: inserting the first set of spatial merge candidates into a merge list; appending the second set of temporal merging candidates to a merge list; in response to determining that there are at least two merge candidates in the merge list, appending a third set of average merge candidates to the merge list based on the at least two merge candidates; responsive to determining that there are less than the maximum allowed number of merge candidates in the merge list, appending a fourth set of history-based motion vector prediction (HMVP) merge candidates to the merge list; and receiving a merge index from the bitstream identifying a merge candidate in the merge list; and generating a prediction block using the identified merge candidates in the merge list.

Description

Modification of the construction of a merging candidate list

Technical Field

The present application relates generally to video data encoding and decoding and, in particular, to methods and systems for modifying the construction of merge candidate lists during video data encoding and decoding.

Background

Various electronic devices, such as digital televisions, laptop or desktop computers, tablet computers, digital cameras, digital recording devices, digital media players, video game consoles, smart phones, video teleconferencing devices, video streaming devices, and the like, support digital video. Electronic devices transmit, receive, encode, decode, and/or store digital video data by implementing video compression/decompression standards as defined by the MPEG-4, ITU-T H.263, ITU-T H.264/MPEG-4, Part 10, Advanced Video Coding (AVC), High Efficiency Video Coding (HEVC), and general video coding (VVC) standards. Video compression typically includes performing spatial (intra) prediction and/or temporal (inter) prediction to reduce or remove redundancy inherent in video data. For block-based video coding, a video frame is partitioned into one or more slices, each slice having a plurality of video blocks, which may also be referred to as Coding Tree Units (CTUs). Each CTU may contain one Coding Unit (CU) or be recursively split into smaller CUs until a predefined minimum CU size is reached. Each CU (also referred to as a leaf CU) contains one or more Transform Units (TUs) and each CU also contains one or more Prediction Units (PUs). Each CU may be encoded in intra, inter, or IBC mode. Video blocks in an intra-coded (I) slice of a video frame are encoded using spatial prediction with respect to reference samples in neighboring blocks within the same video frame. Video blocks in an inter-coded (P or B) slice of a video frame may use spatial prediction with respect to reference samples in neighboring blocks within the same video frame or temporal prediction with respect to reference samples in other previous and/or future reference video frames.

A prediction block for a current video block to be encoded is derived based on spatial prediction or temporal prediction of a reference block (e.g., a neighboring block) that has been previously encoded. The process of finding the reference block may be accomplished by a block matching algorithm. Residual data representing pixel differences between the current block to be encoded and the prediction block is referred to as a residual block or prediction error. The inter-coded block is encoded according to the residual block and a motion vector pointing to a reference block forming a prediction block in a reference frame. The process of determining motion vectors is commonly referred to as motion estimation. And encoding the intra-coded block according to the intra-frame prediction mode and the residual block. For further compression, the residual block is transformed from the pixel domain to a transform domain (e.g., frequency domain), resulting in residual transform coefficients, which may then be quantized. The quantized transform coefficients, initially arranged in a two-dimensional array, may be scanned to produce a one-dimensional vector of transform coefficients, and then entropy encoded into a video bitstream to achieve even greater compression.

The encoded video bitstream is then saved in a computer readable storage medium (e.g., flash memory) for access by another electronic device having digital video capabilities or for direct transmission to the electronic device, either wired or wirelessly. The electronic device then performs video decompression (which is the inverse of the video compression described above), e.g., by parsing the encoded video bitstream to obtain semantic elements from the bitstream and reconstructing the digital video data from the encoded video bitstream into its original format based at least in part on the semantic elements obtained from the bitstream, and the electronic device renders the reconstructed digital video data on a display of the electronic device.

As the digital video quality changes from high definition to 4K × 2K or even 8K × 4K, the amount of video data to be encoded/decoded grows exponentially. It is a long-standing challenge how to encode/decode video data more efficiently while maintaining the image quality of the decoded video data.

Disclosure of Invention

The present application describes embodiments relating to video data encoding and decoding, and more particularly, to systems and methods for parallel processing of video data during video encoding and decoding using history-based motion vector prediction.

According to a first aspect of the present application, a method of decoding video data is performed at a computing device having one or more processors and a memory storing a plurality of programs to be executed by the one or more processors. A computing device performs a method of decoding video data by generating a merge list, comprising: inserting the first set of spatial merge candidates into a merge list; appending the second set of temporal merging candidates to a merge list; in response to determining that there are at least two merge candidates in the merge list, appending a third set of average merge candidates to the merge list based on the at least two merge candidates; responsive to determining that there are less than the maximum allowed number of merge candidates in the merge list, appending a fourth set of history-based motion vector prediction (HMVP) merge candidates to the merge list; and receiving a merge index from the bitstream identifying a merge candidate in the merge list; and generating a prediction block using the identified merge candidates in the merge list.

According to a second aspect of the application, a computing device includes one or more processors, memory, and a plurality of programs stored in the memory. The programs, when executed by one or more processors, cause a computing device to perform operations as described above.

According to a third aspect of the application, a non-transitory computer readable storage medium stores a plurality of programs for execution by a computing device having one or more processors. The programs, when executed by one or more processors, cause a computing device to perform operations as described above.

Drawings

The accompanying drawings, which are included to provide a further understanding of the embodiments and are incorporated in and constitute a part of this specification, illustrate the embodiments described and together with the description serve to explain the principles. Like reference numerals designate corresponding parts.

Fig. 1 is a block diagram illustrating an exemplary video encoding and decoding system according to some embodiments of the present disclosure.

Fig. 2 is a block diagram illustrating an exemplary video encoder according to some embodiments of the present disclosure.

Fig. 3 is a block diagram illustrating an exemplary video decoder according to some embodiments of the present disclosure.

Fig. 4A-4D are block diagrams illustrating how a frame is recursively quadtree partitioned into multiple video blocks of different sizes according to some embodiments of the disclosure.

Fig. 5A is a block diagram illustrating spatially neighboring block positions and temporally co-located block positions of a current CU to be encoded according to some embodiments of the present disclosure.

Fig. 5B is a block diagram illustrating multi-thread encoding of multiple lines of a CTU of a picture using wavefront parallel processing, according to some embodiments of the present disclosure.

Fig. 6 is a block diagram illustrating an example modified insertion order of motion information in a merged motion vector prediction candidate list according to some embodiments of the present disclosure.

Fig. 7 is a flow diagram illustrating an exemplary process by which a video codec implements a technique to build a merged motion vector prediction candidate list according to some embodiments of the present disclosure.

Detailed Description

Reference will now be made in detail to the present embodiments, examples of which are illustrated in the accompanying drawings. In the following detailed description, numerous non-limiting specific details are set forth in order to provide an understanding of the subject matter presented herein. It will be apparent, however, to one skilled in the art that various alternatives may be used and the subject matter may be practiced without these specific details without departing from the scope of the claims. For example, it will be apparent to one of ordinary skill in the art that the subject matter presented herein may be implemented on many types of electronic devices having digital video capabilities.

Fig. 1 is a block diagram illustrating an example system 10 for encoding and decoding video blocks in parallel according to some embodiments of the present disclosure. As shown in fig. 1, system 10 includes a source device 12, source device 12 generating and encoding video data to be later decoded by a target device 14. Source device 12 and target device 14 may comprise any of a wide variety of electronic devices, including desktop or laptop computers, tablet computers, smart phones, set-top boxes, digital televisions, cameras, display devices, digital media players, video game machines, video streaming devices, and the like. In some embodiments, source device 12 and target device 14 are equipped with wireless communication capabilities.

In some embodiments, target device 14 may receive encoded video data to be decoded via link 16. Link 16 may include any type of communication medium or device capable of moving encoded video data from source device 12 to destination device 14. In one example, link 16 may include a communication medium that enables source device 12 to transmit encoded video data directly to target device 14 in real-time. The encoded video data may be modulated according to a communication standard, such as a wireless communication protocol, and transmitted to the target device 14. The communication medium may include any wireless or wired communication medium such as a Radio Frequency (RF) spectrum or one or more physical transmission lines. The communication medium may form part of a packet-based network (e.g., a local area network, a wide area network, or a global network such as the internet). The communication medium may include a router, switch, base station, or any other device that may facilitate communication from source device 12 to target device 14.

In some other implementations, the encoded video data may be sent from the output interface 22 to the storage device 32. Subsequently, the encoded video data in storage device 32 may be accessed by target device 14 via input interface 28. Storage device 32 may include any of a variety of distributed or locally accessed data storage media such as a hard drive, blu-ray discs, DVDs, CD-ROMs, flash memory, volatile or non-volatile memory, or any other suitable digital storage media for storing encoded video data. In another example, storage device 32 may correspond to a file server or another intermediate storage device that may hold encoded video data generated by source device 12. The target device 14 may access the stored video data from the storage device 32 via streaming or download. The file server may be any type of computer capable of storing encoded video data and transmitting the encoded video data to the target device 14. Exemplary file servers include web servers (e.g., for a website), FTP servers, Network Attached Storage (NAS) devices, or local disk drives. The target device 14 may access the encoded video data through any standard data connection suitable for accessing encoded video data stored on a file server, including a wireless channel (e.g., a Wi-Fi connection), a wired connection (e.g., DSL, cable modem, etc.), or a combination of both wireless and wired connections. The transmission of the encoded video data from the storage device 32 may be a streaming transmission, a download transmission, or a combination of both a streaming and a download transmission.

As shown in fig. 1, source device 12 includes a video source 18, a video encoder 20, and an output interface 22. Video source 18 may include sources such as the following or a combination of such sources: a video capture device (e.g., a video camera), a video archive containing previously captured video, a video feed interface for receiving video from a video content provider, and/or a computer graphics system for generating computer graphics data as source video. As one example, if video source 18 is a video camera of a security monitoring system, source device 12 and destination device 14 may form a camera phone or video phone. However, the embodiments described herein are generally applicable to video encoding/decoding and may be applied to wireless and/or wired applications.

Captured, pre-captured, or computer-generated video may be encoded by video encoder 20. The encoded video data may be sent directly to the target device 14 via the output interface 22 of the source device 12. The encoded video data may also (or alternatively) be stored on storage device 32 for later access by target device 14 or other devices for decoding and/or playback. The output interface 22 may further include a modem and/or a transmitter.

The target device 14 includes an input interface 28, a video decoder 30, and a display device 34. Input interface 28 may include a receiver and/or a modem and receives encoded video data over link 16. The encoded video data communicated over link 16 or provided on storage device 32 may include various semantic elements generated by video encoder 20 for use by video decoder 30 in decoding the video data. Such semantic elements may be included within encoded video data sent over a communication medium, stored on a storage medium, or stored on a file server.

In some embodiments, the target device 14 may include a display device 34, and the display device 34 may be an integrated display device and an external display device configured to communicate with the target device 14. Display device 34 displays the decoded video data to a user and may include any of a variety of display devices, such as a Liquid Crystal Display (LCD), a plasma display, an Organic Light Emitting Diode (OLED) display, or another type of display device.

Video encoder 20 and video decoder 30 may operate according to a proprietary or industry standard (e.g., VVC, HEVC, MPEG-4, Part 10, Advanced Video Coding (AVC)) or an extension of such a standard. It should be understood that the present application is not limited to a particular video encoding/decoding standard and may be applicable to other video encoding/decoding standards. It is generally recognized that video encoder 20 of source device 12 may be configured to encode video data according to any of these current or future standards. Similarly, it is also generally contemplated that video decoder 30 of target device 14 may be configured to decode video data in accordance with any of these current or future standards.

Video encoder 20 and video decoder 30 may each be implemented as any of a variety of suitable encoder circuits, such as one or more microprocessors, Digital Signal Processors (DSPs), Application Specific Integrated Circuits (ASICs), Field Programmable Gate Arrays (FPGAs), discrete logic, software, hardware, firmware, or any combinations thereof. When implemented in part in software, the electronic device may store instructions for the software in a suitable non-transitory computer-readable medium and execute the instructions in hardware using one or more processors to perform the video encoding/decoding operations disclosed in this disclosure. Each of video encoder 20 and video decoder 30 may be included in one or more encoders or decoders, either of which may be integrated as part of a combined encoder/decoder (CODEC) in the respective device.

Fig. 2 is a block diagram illustrating an exemplary video encoder 20 according to some embodiments described in the present application. Video encoder 20 may perform intra-prediction encoding and inter-prediction encoding of video blocks within video frames. Intra-prediction coding relies on spatial prediction to reduce or remove spatial redundancy in video data within a given video frame or picture. Inter-prediction coding relies on temporal prediction to reduce or remove temporal redundancy in video data within adjacent video frames or pictures of a video sequence.

As shown in fig. 2, video encoder 20 includes a video data memory 40, a prediction processing unit 41, a Decoded Picture Buffer (DPB)64, an adder 50, a transform processing unit 52, a quantization unit 54, and an entropy encoding unit 56. Prediction processing unit 41 further includes a motion estimation unit 42, a motion compensation unit 44, a partition unit 45, an intra prediction processing unit 46, and an intra Block Copy (BC) unit 48. In some embodiments, video encoder 20 also includes an inverse quantization unit 58, an inverse transform processing unit 60, and an adder 62 for video block reconstruction. A deblocking filter (not shown) may be located between adder 62 and DPB64 to filter block boundaries to remove blockiness from the reconstructed video. In addition to a deblocking filter, an in-loop filter (not shown) may be used to filter the output of adder 62. Video encoder 20 may take the form of, or be dispersed among, one or more of the fixed or programmable hardware units illustrated.

Video data memory 40 may store video data to be encoded by components of video encoder 20. The video data in video data storage 40 may be obtained, for example, from video source 18. DPB64 is a buffer that stores reference video data for use by video encoder 20 in encoding video data (e.g., in intra or inter prediction encoding modes). Video data memory 40 and DPB64 may be formed from any of a variety of memory devices. In various examples, video data memory 40 may be on-chip with other components of video encoder 20, or off-chip with respect to those components.

As shown in fig. 2, upon receiving the video data, a partition unit 45 within prediction processing unit 41 partitions the video data into video blocks. This partitioning may also include partitioning the video frame into slices, partitions (tiles), or other larger Coding Units (CUs) according to a predefined splitting structure, such as a quadtree structure, associated with the video data. A video frame may be divided into a plurality of video blocks (or a set of video blocks referred to as partitions). Prediction processing unit 41 may select one of a plurality of possible prediction encoding modes, such as one of one or more inter-prediction encoding modes of a plurality of intra-prediction encoding modes, for the current video block based on the error results (e.g., encoding rate and distortion level). Prediction processing unit 41 may provide the resulting intra-predicted or inter-predicted encoded blocks to adder 50 to generate a residual block, and to adder 62 to reconstruct the encoded block for subsequent use as part of a reference frame. The prediction processing unit 41 also provides semantic elements such as motion vectors, intra mode indicators, partition information, and other such syntax information to the entropy encoding unit 56.

To select a suitable intra-prediction encoding mode for the current video block, intra-prediction processing unit 46 within prediction processing unit 41 may perform intra-prediction encoding of the current video block in relation to one or more neighboring blocks in the same frame as the current block to be encoded to provide spatial prediction. Motion estimation unit 42 and motion compensation unit 44 within prediction processing unit 41 perform inter-prediction encoding of the current video block in relation to one or more prediction blocks in one or more reference frames to provide temporal prediction. Video encoder 20 may perform multiple encoding passes, for example, to select an appropriate encoding mode for each block of video data.

In some implementations, motion estimation unit 42 determines the inter-prediction mode for the current video frame by generating motion vectors according to predetermined patterns within the sequence of video frames, the motion vectors indicating the displacement of Prediction Units (PUs) of video blocks within the current video frame relative to prediction blocks within the reference video frame. Motion estimation performed by motion estimation unit 42 is the process of generating motion vectors that estimate motion for video blocks. For example, a motion vector may indicate the displacement of a PU of a video block within a current video frame or picture relative to a prediction block (or other coding unit) within a reference frame that is related to a current block (or other coding unit) being encoded within the current frame. The predetermined pattern may designate video frames in the sequence as P-frames or B-frames. Intra BC unit 48 may determine vectors (e.g., block vectors) for intra BC encoding in a similar manner as motion vectors determined by motion estimation unit 42 for inter prediction, or block vectors may be determined using motion estimation unit 42.

In terms of pixel differences, which may be determined by Sum of Absolute Differences (SAD), Sum of Squared Differences (SSD), or other difference metrics, a prediction block is a block of the reference frame that is considered to closely match a PU of the video block to be encoded. In some implementations, video encoder 20 may calculate values for sub-integer pixel positions of reference frames stored in DPB 64. For example, video encoder 20 may interpolate values for a quarter-pixel position, an eighth-pixel position, or other fractional-pixel positions of the reference frame. Thus, motion estimation unit 42 may perform a motion search with respect to the full pixel position and the fractional pixel position and output a motion vector with fractional pixel precision.

Motion estimation unit 42 calculates motion vectors for PUs of video blocks in the inter-prediction coded frame by: the location of the PU is compared to locations of prediction blocks of reference frames selected from a first reference frame list (list 0) or a second reference frame list (list 1), each of which identifies one or more reference frames stored in the DPB 64. The motion estimation unit 42 sends the calculated motion vector to the motion compensation unit 44 and then to the entropy coding unit 56.

The motion compensation performed by motion compensation unit 44 may involve extracting or generating a prediction block based on the motion vector determined by motion estimation unit 42. Upon receiving the motion vector for the PU of the current video block, motion compensation unit 44 may locate the prediction block to which the motion vector points in one of the reference frame lists, retrieve the prediction block from DPB64, and forward the prediction block to adder 50. Adder 50 then forms a residual video block of pixel difference values by subtracting the pixel values of the prediction block provided by motion compensation unit 44 from the pixel values of the current video block being encoded. The pixel difference values forming the residual video block may comprise a luminance difference component or a chrominance difference component or both. Motion compensation unit 44 may also generate semantic elements associated with video blocks of the video frame for use by video decoder 30 in decoding video blocks of the video frame. The semantic elements may include, for example, semantic elements defining motion vectors used to identify prediction blocks, any flag indicating a prediction mode, or any other syntax information described herein. It should be noted that motion estimation unit 42 and motion compensation unit 44 may be highly integrated, but are illustrated separately for conceptual purposes.

In some embodiments, intra BC unit 48 may generate vectors and extract prediction blocks in a manner similar to that described above in connection with motion estimation unit 42 and motion compensation unit 44, but in the same frame as the current block being encoded, and these vectors are referred to as block vectors rather than motion vectors. In particular, intra BC unit 48 may determine the intra prediction mode to be used for encoding the current block. In some examples, intra BC unit 48 may encode current blocks using various intra prediction modes, e.g., during separate encoding passes, and test their performance through rate-distortion analysis. Next, intra BC unit 48 may select an appropriate intra prediction mode among the various tested intra prediction modes to use and generate an intra mode indicator accordingly. For example, intra BC unit 48 may calculate rate-distortion values for various tested intra-prediction modes using rate-distortion analysis, and select the intra-prediction mode having the best rate-distortion characteristics among the tested modes to use as the appropriate intra-prediction mode. Rate-distortion analysis generally determines the amount of distortion (or error) between an encoded block and the original, unencoded block that was encoded to generate the encoded block, as well as the bit rate (i.e., number of bits) used to produce the encoded block. Intra BC unit 48 may calculate ratios from the distortion and rate for various encoded blocks to determine which intra prediction mode exhibits the best rate-distortion value for the block.

In other examples, intra BC unit 48 may use, in whole or in part, motion estimation unit 42 and motion compensation unit 44 to perform such functions for intra BC prediction according to embodiments described herein. In either case, for intra block copying, the prediction block may be a block that is considered to closely match the block to be encoded in terms of pixel differences, which may be determined by Sum of Absolute Differences (SAD), Sum of Squared Differences (SSD), or other difference metrics, and the identification of the prediction block may include calculating values for sub-integer pixel locations.

Whether the prediction block is from the same frame according to intra prediction or from a different frame according to inter prediction, video encoder 20 may form a residual video block by subtracting pixel values of the prediction block from pixel values of the current video block being encoded to form pixel difference values. The pixel difference values forming the residual video block may include both luminance component differences and chrominance component differences.

As an alternative to inter prediction performed by motion estimation unit 42 and motion compensation unit 44 or intra block copy prediction performed by intra BC unit 48 as described above, intra prediction processing unit 46 may intra predict the current video block. In particular, intra-prediction processing unit 46 may determine an intra-prediction mode for encoding the current block. To this end, intra-prediction processing unit 46 may encode the current block using various intra-prediction modes, e.g., during separate encoding passes, and intra-prediction processing unit 46 (or, in some examples, a mode selection unit) may select an appropriate intra-prediction mode from the tested intra-prediction modes for use. Intra-prediction processing unit 46 may provide information to entropy encoding unit 56 indicating the intra-prediction mode selected for the block. Entropy encoding unit 56 may encode information indicating the selected intra-prediction mode into a bitstream.

After prediction processing unit 41 determines a prediction block for the current video block via inter prediction or intra prediction, adder 50 forms a residual video block by subtracting the prediction block from the current video block. The residual video data in the residual block may be included in one or more Transform Units (TUs) and provided to transform processing unit 52. The transform processing unit 52 transforms the residual video data into residual transform coefficients using a transform such as Discrete Cosine Transform (DCT) or a conceptually similar transform.

Transform processing unit 52 may send the resulting transform coefficients to quantization unit 54. The quantization unit 54 quantizes the transform coefficients to further reduce the bit rate. The quantization process may also reduce the bit depth associated with some or all of the coefficients. The quantization level may be modified by adjusting a quantization parameter. In some examples, quantization unit 54 may then perform a scan of a matrix including quantized transform coefficients. Alternatively, entropy encoding unit 56 may perform scanning.

After quantization, entropy encoding unit 56 entropy encodes the quantized transform coefficients into a video bitstream using, for example, Context Adaptive Variable Length Coding (CAVLC), Context Adaptive Binary Arithmetic Coding (CABAC), syntax-based context adaptive binary arithmetic coding (SBAC), Probability Interval Partition Entropy (PIPE) coding, or another entropy encoding method or technique. The encoded bitstream may then be transmitted to video decoder 30, or archived in storage device 32 for later transmission to video decoder 30 or retrieval by video decoder 30. Entropy encoding unit 56 may also entropy encode the motion vectors and other semantic elements for the current video frame being encoded.

Inverse quantization unit 58 and inverse transform processing unit 60 apply inverse quantization and inverse transform, respectively, to reconstruct the residual video block in the pixel domain for use in generating reference blocks for predicting other video blocks. As noted above, motion compensation unit 44 may generate a motion compensated prediction block from one or more reference blocks of a frame stored in DPB 64. Motion compensation unit 44 may also apply one or more interpolation filters to the prediction blocks to calculate sub-integer pixel values for use in motion estimation.

Adder 62 adds the reconstructed residual block to the motion compensated prediction block produced by motion compensation unit 44 to produce a reference block for storage in DPB 64. The reference block may then be used by intra BC unit 48, motion estimation unit 42, and motion compensation unit 44 as a prediction block to inter-predict another video block in a subsequent video frame.

Fig. 3 is a block diagram illustrating an exemplary video decoder 30 according to some embodiments of the present application. The video decoder 30 includes a video data memory 79, an entropy decoding unit 80, a prediction processing unit 81, an inverse quantization unit 86, an inverse transform processing unit 88, an adder 90, and a DPB 92. Prediction processing unit 81 further includes a motion compensation unit 82, an intra prediction processing unit 84, and an intra BC unit 85. Video decoder 30 may perform a decoding process that is substantially reciprocal to the encoding process described above with respect to video encoder 20 in connection with fig. 2. For example, motion compensation unit 82 may generate prediction data based on motion vectors received from entropy decoding unit 80, while intra-prediction unit 84 may generate prediction data based on intra-prediction mode indicators received from entropy decoding unit 80.

In some examples, the units of video decoder 30 may be tasked to perform embodiments of the present application. Furthermore, in some examples, embodiments of the present disclosure may be dispersed in one or more of the plurality of units of video decoder 30. For example, intra BC unit 85 may perform embodiments of the present application alone or in combination with other units of video decoder 30, such as motion compensation unit 82, intra prediction processing unit 84, and entropy decoding unit 80. In some examples, video decoder 30 may not include intra BC unit 85, and the functions of intra BC unit 85 may be performed by other components of prediction processing unit 81 (such as motion compensation unit 82).

Video data memory 79 may store video data to be decoded by other components of video decoder 30, such as an encoded video bitstream. The video data stored in video data storage 79 may be obtained, for example, from storage device 32, from a local video source (such as a camera), via wired or wireless network communication of the video data, or by accessing a physical data storage medium (e.g., a flash drive or hard disk). Video data memory 79 may include a Coded Picture Buffer (CPB) that stores coded video data from a coded video bitstream. Decoded Picture Buffer (DPB)92 of video decoder 30 stores reference video data for use by video decoder 30 when decoding the video data (e.g., in intra or inter prediction encoding modes). Video data memory 79 and DPB 92 may be formed from any of a variety of memory devices, such as Dynamic Random Access Memory (DRAM) (including synchronous DRAM (sdram)), magnetoresistive ram (mram), resistive ram (rram), or other types of memory devices. For illustrative purposes, video data memory 79 and DPB 92 are depicted in fig. 3 as two different components of video decoder 30. It will be apparent to those skilled in the art that video data memory 79 and DPB 92 may be provided by the same memory device or separate memory devices. In some examples, video data memory 79 may be on-chip with other components of video decoder 30, or off-chip with respect to those components.

During the decoding process, video decoder 30 receives an encoded video bitstream representing video blocks and associated semantic elements of an encoded video frame. Video decoder 30 may receive semantic elements at the video frame level and/or the video block level. Entropy decoding unit 80 of video decoder 30 entropy decodes the bitstream to generate quantized coefficients, motion vectors or intra prediction mode indicators, and other semantic elements. Then, the entropy decoding unit 80 forwards the motion vector and other semantic elements to the prediction processing unit 81.

When a video frame is encoded as an intra-prediction encoded (I) frame or as an intra-coded prediction block for use in other types of frames, intra-prediction processing unit 84 of prediction processing unit 81 may generate prediction data for a video block of the current video frame based on the signaled intra-prediction mode and reference data from previously decoded blocks of the current frame.

When a video frame is encoded as an inter-prediction encoded (i.e., B or P) frame, motion compensation unit 82 of prediction processing unit 81 generates one or more prediction blocks for the video block of the current video frame based on the motion vectors and other semantic elements received from entropy decoding unit 80. Each of the prediction blocks may be generated from a reference frame within one of the reference frame lists. Video decoder 30 may use a default construction technique to construct reference frame lists, list 0 and list 1, based on the reference frames stored in DPB 92.

In some examples, when encoding a video block according to the intra BC mode described herein, intra BC unit 85 of prediction processing unit 81 generates a prediction block for the current video block based on the block vector and other semantic elements received from entropy decoding unit 80. The prediction block may be within a reconstruction region of the same picture as the current video block defined by video encoder 20.

Motion compensation unit 82 and/or intra BC unit 85 determine prediction information for the video block of the current video frame by parsing the motion vectors and other semantic elements and then use the prediction information to generate a prediction block for the current video block being decoded. For example, motion compensation unit 82 uses some of the received semantic elements to determine a prediction mode (e.g., intra-prediction or inter-prediction) for encoding a video block of a video frame, an inter-prediction frame type (e.g., B or P), construction information for one or more of a list of reference frames for the frame, a motion vector for each inter-prediction encoded video block of the frame, an inter-prediction state for each inter-prediction encoded video block of the frame, and other information for decoding a video block in the current video frame.

Similarly, some of the received semantic elements, such as flags, may be used by intra BC unit 85 to determine that the current video block is predicted using an intra BC mode, build information for which video blocks of the frame are within the reconstruction region and should be stored in DPB 92, a block vector for each intra BC predicted video block of the frame, intra BC prediction status for each intra BC predicted video block of the frame, and other information for decoding the video blocks in the current video frame.

Motion compensation unit 82 may also perform interpolation using interpolation filters as used by video encoder 20 during encoding of video blocks to calculate interpolated values for sub-integer pixels of a reference block. In this case, motion compensation unit 82 may determine interpolation filters used by video encoder 20 from the received semantic elements and use these interpolation filters to generate prediction blocks.

Inverse quantization unit 86 inverse quantizes the quantized transform coefficients provided in the bitstream and entropy decoded by entropy decoding unit 80 using the same quantization parameter calculated by video encoder 20 for each video block in the video frame to determine the degree of quantization. Inverse transform processing unit 88 applies an inverse transform (e.g., an inverse DCT, an inverse integer transform, or a conceptually similar inverse transform process) to the transform coefficients in order to reconstruct the residual block in the pixel domain.

After motion compensation unit 82 or intra BC unit 85 generates a prediction block for the current video block based on the vector and other semantic elements, adder 90 reconstructs the decoded video block for the current video block by adding the residual block from inverse transform processing unit 88 to the corresponding prediction block generated by motion compensation unit 82 and intra BC unit 85. An in-loop filter (not shown) may be located between adder 90 and DPB 92 to further process the decoded video block. The decoded video blocks in a given frame are then stored in DPB 92, and DPB 92 stores reference frames for subsequent motion compensation of subsequent video blocks. DPB 92, or a memory device separate from DPB 92, may also store decoded video for later presentation on a display device (e.g., display device 34 of fig. 1).

In a typical video encoding process, a video sequence typically comprises an ordered set of frames or pictures. Each frame may include three arrays of samples, denoted SL, SCb, and SCr. SL is a two-dimensional array of brightness samples. SCb is a two-dimensional array of Cb chroma samples. SCr is a two-dimensional array of Cr chroma samples. In other cases, the frame may be monochromatic, and thus include only one two-dimensional array of luminance samples.

As shown in fig. 4A, video encoder 20 (or, more specifically, segmentation unit 45) generates an encoded representation of a frame by first segmenting the frame into a set of Coding Tree Units (CTUs). A video frame may include an integer number of CTUs ordered sequentially from left to right and top to bottom in raster scan order. Each CTU is the largest logical coding unit, and the width and height of the CTUs are signaled by video encoder 20 in a sequence parameter set such that all CTUs in a video sequence have the same size of one of 128 × 128, 64 × 64, 32 × 32, and 16 × 16. It should be noted, however, that the present application is not necessarily limited to a particular size. As shown in fig. 4B, each CTU may include one Coding Tree Block (CTB) of luma samples, two corresponding coding tree blocks of chroma samples, and semantic elements for encoding samples of the coding tree blocks. The semantic elements describe the properties of the different types of units that encode the pixel blocks and how the video sequence may be reconstructed at video decoder 30, including inter or intra prediction, intra prediction modes, motion vectors, and other parameters. In a monochrome picture or a picture with three separate color planes, a CTU may comprise a single coding tree block and semantic elements for coding samples of the coding tree block. The coding tree block may be an N × N block of samples.

To achieve better performance, video encoder 20 may recursively perform tree partitioning, e.g., binary tree partitioning, quadtree partitioning, or a combination of both, on the coding tree blocks of the CTUs and divide the CTUs into smaller Coding Units (CUs). As depicted in fig. 4C, the 64 × 64 CTU 400 is first divided into four smaller CUs, each having a block size of 32 × 32. Of the four smaller CUs, CU 410 and CU 420 are divided into four CUs with block sizes of 16 × 16, respectively. The two 16 × 16 CUs 430 and the CU 440 are further divided into four CUs having block sizes of 8 × 8, respectively. Fig. 4D depicts a quadtree data structure showing the final result of the segmentation process of the CTU 400 as depicted in fig. 4C, each leaf node of the quadtree corresponding to one CU of various sizes ranging from 32 x 32 to 8 x 8. Similar to the CTU depicted in fig. 4B, each CU may include a Coded Block (CB) of luma samples and two corresponding coded blocks of chroma samples of the same size frame, and semantic elements for encoding the samples of the coded blocks. In a monochrome picture or a picture with three separate color planes, a CU may comprise a single coding block and syntax structures for coding the samples of the coding block.

In some implementations, video encoder 20 may further partition the coding block of the CU into one or more mxn Prediction Blocks (PBs). A prediction block is a block of rectangular (square or non-square) samples to which the same prediction (inter or intra) is applied. A Prediction Unit (PU) of a CU may include a prediction block of luma samples, two corresponding prediction blocks of chroma samples, and semantic elements for predicting the prediction block. In a monochrome picture or a picture with three separate color planes, a PU may include a single prediction block and syntax structures used to predict the prediction block. Video encoder 20 may generate predicted luma, predicted Cb, and predicted Cr blocks for the luma, Cb, and Cr predicted blocks for each PU of the CU.

Video encoder 20 may generate the prediction block for the PU using intra prediction or inter prediction. If video encoder 20 uses intra-prediction to generate the prediction block for the PU, video encoder 20 may generate the prediction block for the PU based on the decoding samples of the frame associated with the PU. If video encoder 20 uses inter-prediction to generate the prediction block for the PU, video encoder 20 may generate the prediction block for the PU based on decoding samples of one or more frames other than the frame associated with the PU.

After video encoder 20 generates the predicted luma block, the predicted Cb block, and the predicted Cr block for one or more PUs of the CU, video encoder 20 may generate a luma residual block for the CU by subtracting the predicted luma block of the CU from the original luma coding block of the CU, such that each sample in the luma residual block of the CU indicates a difference between a luma sample in one of the predicted luma blocks of the CU and a corresponding sample in the original luma coding block of the CU. Similarly, video encoder 20 may generate the Cb residual block and the Cr residual block for the CU, respectively, such that each sample in the Cb residual block of the CU indicates a difference between a Cb sample in one of the predicted Cb blocks of the CU and a corresponding sample in the original Cb coding block of the CU, and each sample in the Cr residual block of the CU may indicate a difference between a Cr sample in one of the predicted Cr blocks of the CU and a corresponding sample in the original Cr coding block of the CU.

Furthermore, as shown in fig. 4C, video encoder 20 may decompose the luma, Cb, and Cr residual blocks of the CU into one or more luma, Cb, and Cr transform blocks using quadtree partitioning. A transform block is a block of rectangular (square or non-square) samples to which the same transform is applied. A Transform Unit (TU) of a CU may include a transform block of luma samples, two corresponding transform blocks of chroma samples, and semantic elements for transforming the transform block samples. Thus, each TU of a CU may be associated with a luma transform block, a Cb transform block, and a Cr transform block. In some examples, the luma transform block associated with a TU may be a sub-block of a luma residual block of a CU. The Cb transform block may be a sub-block of a Cb residual block of the CU. The Cr transform block may be a sub-block of the Cr residual block of the CU. In a monochrome picture or a picture with three separate color planes, a TU may comprise a single transform block and syntax structures for transforming the samples of the transform block.

Video encoder 20 may apply one or more transforms to a luma transform block of a TU to generate a luma coefficient block for the TU. The coefficient block may be a two-dimensional array of transform coefficients. The transform coefficients may be scalars. Video encoder 20 may apply one or more transforms to Cb transform blocks of a TU to generate Cb coefficient blocks for the TU. Video encoder 20 may apply one or more transforms to a Cr transform block of a TU to generate a Cr coefficient block for the TU.

After generating the coefficient block (e.g., a luminance coefficient block, a Cb coefficient block, or a Cr coefficient block), video encoder 20 may quantize the coefficient block. Quantization generally refers to the process by which transform coefficients are quantized to possibly reduce the amount of data used to represent the transform coefficients, thereby providing further compression. After video encoder 20 quantizes the coefficient block, video encoder 20 may entropy encode semantic elements that indicate the quantized transform coefficients. For example, video encoder 20 may perform Context Adaptive Binary Arithmetic Coding (CABAC) on semantic elements that indicate quantized transform coefficients. Finally, video encoder 20 may output a bitstream that includes the bit sequence that forms a representation of the encoded frames and associated data, which is stored in storage device 32 or transmitted to target device 14.

Upon receiving the bitstream generated by video encoder 20, video decoder 30 may parse the bitstream to obtain semantic elements from the bitstream. Video decoder 30 may reconstruct the frames of video data based at least in part on semantic elements obtained from the bitstream. The process of reconstructing the video data is generally reciprocal to the encoding process performed by video encoder 20. For example, video decoder 30 may perform inverse transforms on coefficient blocks associated with TUs of the current CU to reconstruct residual blocks associated with the TUs of the current CU. Video decoder 30 also reconstructs the encoded block of the current CU by adding samples of the prediction block for the PUs of the current CU to corresponding samples of the transform blocks of the TUs of the current CU. After reconstructing the encoded blocks for each CU of a frame, video decoder 30 may reconstruct the frame.

As described above, video encoding mainly uses two modes, i.e., intra-frame prediction (or intra-frame prediction) and inter-frame prediction (or inter-frame prediction), to achieve video compression. It should be noted that IBC may be considered as intra prediction or third mode. Between the two modes, inter prediction contributes more to coding efficiency than intra prediction since the current video block is predicted from the reference video block using motion vectors.

But with ever-improving video data capture techniques and finer video block sizes for preserving details in the video data, the amount of data required to represent motion vectors for the current frame also increases substantially. One way to overcome this challenge is to benefit from the fact that: not only does a set of neighboring CUs in both the spatial and temporal domains have similar video data for prediction purposes, but the motion vectors between these neighboring CUs are also similar. Thus, it is possible to use the motion information of spatially neighboring CUs and/or temporally co-located CUs as an approximation of the motion information (e.g., motion vector) of the current CU by: their spatial and temporal correlation, also called "motion vector predictor" (MVP) of the current CU, is explored.

Instead of encoding the actual motion vector of the current CU into the video bitstream as determined by motion estimation unit 42 as described above in connection with fig. 2, the motion vector predictor of the current CU is subtracted from the actual motion vector of the current CU to generate a Motion Vector Difference (MVD) for the current CU. By doing so, it is not necessary to encode the motion vector determined by the motion estimation unit 42 for each CU of a frame into the video bitstream, and the amount of data used to represent motion information in the video bitstream can be significantly reduced.

Similar to the process of selecting a prediction block in a reference frame during inter prediction of a coding block, both video encoder 20 and video decoder 30 need to employ a set of rules for constructing a motion vector candidate list (also referred to as a "merge list") for the current CU using those potential candidate motion vectors associated with spatially neighboring CUs and/or temporally co-located CUs of the current CU, and then select one member from the motion vector candidate list as a motion vector predictor for the current CU. By doing so, the motion vector candidate list itself need not be sent between video encoder 20 and video decoder 30, and the index of the selected motion vector predictor within the motion vector candidate list is sufficient for video encoder 20 and video decoder 30 to use the same motion vector predictor within the motion vector candidate list to encode and decode the current CU.

In some embodiments, each inter-predicted CU has three motion vector prediction modes for constructing a motion vector candidate list, including an inter mode (which is also referred to as "advanced motion vector prediction" (AMVP)), a skip mode, and a merge mode. In each mode, one or more motion vector candidates may be added to the motion vector candidate list according to the algorithm described below. Finally, one of these motion vector candidates in the candidate list is used as the best motion vector predictor for the inter-predicted CU to be encoded into the video bitstream by video encoder 20 or decoded from the video bitstream by video decoder 30. In order to find the best motion vector predictor from the candidate list, a Motion Vector Competition (MVC) scheme is introduced to select a motion vector from a given candidate set of motion vectors (i.e., a motion vector candidate list) comprising spatial motion vector candidates and temporal motion vector candidates.

In addition to deriving motion vector predictor candidates from spatially neighboring CUs or temporally collocated CUs, motion vector predictor candidates may also be derived from a so-called "history-based motion vector prediction" (HMVP) table. The HMVP table accommodates a predefined number of motion vector predictors, each motion vector predictor having been used to encode/decode a particular CU in the same row of CTUs (or sometimes the same CTU). Due to the spatial/temporal proximity of these CUs, the probability that one of the motion vector predictors in the HMVP table can be reused for encoding/decoding different CUs within the CTU of the same row is high. Therefore, it is possible to achieve higher coding efficiency by including the HMVP table in the process of constructing the motion vector candidate list.

In some embodiments, the HMVP table has a fixed length (e.g., 5) and is managed in a first-in-first-out (FIFO) like manner. For example, when decoding one inter-coded block of a CU, a motion vector is reconstructed for the CU. Because the reconstructed motion vector may be a motion vector predictor of a subsequent CU, the HMVP table is updated on the fly using such motion vector. When updating the HMVP table, there are two scenarios: (i) the reconstructed motion vector is different from other existing motion vectors in the HMVP table or (ii) the reconstructed motion vector is the same as one of the existing motion vectors in the HMVP table. For the first scene, if the HMVP table is not full, the reconstructed motion vector is added as the latest motion vector to the HMVP table. If the HMVP table is full, the oldest motion vector in the HMVP table needs to be first removed from the HMVP table before the reconstructed motion vector is added as the newest motion vector. In other words, in this case, the HMVP table is similar to the FIFO buffer, such that motion information located at the head of the FIFO buffer and associated with another previous inter-coded block is shifted out of the buffer, such that the reconstructed motion vector is appended to the tail of the FIFO buffer as the latest member in the HMVP table. For the second context, prior to adding the reconstructed motion vector as the latest motion vector to the HMVP table, an existing motion vector in the HMVP table that is substantially the same as the reconstructed motion vector is removed from the HMVP table. If the HMVP table is also maintained in the form of a FIFO buffer, the motion vector predictor following the same motion vector in the HMVP table is shifted forward by one element to occupy the space left by the removed motion vector, and then the reconstructed motion vector is appended to the tail of the FIFO buffer as the latest member in the HMVP table.

The motion vectors in the HMVP table may be added to the motion vector candidate list under different prediction modes, such as AMVP, merge, skip, etc. It has been found that motion information of a previous inter-coded block stored in the HMVP table even if it is not adjacent to the current block can be used for more efficient motion vector prediction.

After selecting one MVP candidate within a given candidate set of motion vectors for a current CU, video encoder 20 may generate one or more semantic elements for the corresponding MVP candidate and encode them into a video bitstream such that video decoder 30 may retrieve the MVP candidate from the video bitstream using the semantic elements. Different modes (e.g., AMVP, merge, skip, etc.) have different sets of semantic elements depending on the particular mode used to construct the motion vector candidate set. For AMVP mode, the semantic elements include inter prediction indicators (list 0, list 1, or bi-prediction), reference indices, motion vector candidate indices, motion vector prediction residual signals, and the like. For skip mode and merge mode, only the merge index is encoded into the bitstream, since the current CU inherits other semantic elements including inter prediction indicators, reference indices, and motion vectors from the neighboring CUs referred to by the encoded merge index. In case of skip coding CU, the motion vector prediction residual signal is also omitted.

Fig. 5A is a block diagram illustrating spatially neighboring block positions and temporally co-located block positions of a current CU to be encoded/decoded according to some embodiments of the present disclosure. For a given mode, a Motion Vector Prediction (MVP) candidate list is constructed by first checking the availability of motion vectors associated with the left spatially neighboring block position and the above spatially neighboring block position and the availability of motion vectors associated with the temporally co-located block position, and then checking the motion vectors in the HMVP table. During the process of constructing the MVP candidate list, some redundant MVP candidates are removed from the candidate list and, if necessary, zero-valued motion vectors are added to make the candidate list have a fixed length (note that different modes may have different fixed lengths). After constructing the MVP candidate list, video encoder 20 may select the best motion vector predictor from the candidate list and encode a corresponding index into the video bitstream that indicates the selected candidate.

Using fig. 5A as an example and assuming that the candidate list has a fixed length of 2, a Motion Vector Predictor (MVP) candidate list for the current CU may be constructed by sequentially performing the following steps in AMVP mode:

1) selecting MVP candidates from spatially neighboring CUs

a) At most one un-scaled MVP candidate is derived from one of the two left-side spatially neighboring CUs starting with a0 and ending with a 1;

b) if no un-scaled MVP candidates from the left side are available in the previous step, at most one scaled MVP candidate is obtained from one of the two left-side spatially neighboring CUs starting with a0 and ending with a 1;

c) at most one un-scaled MVP candidate is derived from one of three above-spatially neighboring CUs starting with B0, then B1 and ending with B2;

d) if neither a0 nor a1 are available, or if they are coded in intra mode, at most one scaled MVP candidate is derived from one of three above spatially neighboring CUs starting with B0, then B1 and ending with B2;

2) removing one of the two MVP candidates from the MVP candidate list if they are found in the previous step and are the same;

3) selecting MVP candidates from temporally co-located CUs

a) If the MVP candidate list after the previous step does not include two MVP candidates, at most one MVP candidate is obtained from the temporally co-located CU (e.g., T0);

4) selection of MVP candidates from HMVP table

a) If the MVP candidate list after the previous step does not include two MVP candidates, obtaining at most two history-based MVPs from the HMVP table; and

5) if the MVP candidate list after passing through the previous steps does not include two MVP candidates, at most two zero-valued MVPs are added to the MVP candidate list.

Since there are only two candidates in the AMVP mode MVP candidate list constructed above, an associated semantic element (such as a binary flag) is encoded into the bitstream to indicate which of the two MVP candidates within the candidate list is used to decode the current CU.

In some embodiments, the MVP candidate list for the current CU in skip or merge mode may be constructed by performing a set of similar steps in order as above. It should be noted that a special type of merge candidate, called "pairwise merge candidate", is also included in the MVP candidate list for skip or merge mode. The pair-wise merge candidates are generated by averaging the MVs of the two previously obtained merge mode motion vector candidates. The size of the merged MVP candidate list (e.g., from 1 to 6) is signaled in the slice header of the current CU. For each CU in the merge mode, the index of the best merge candidate is encoded using truncated unary binarization (TU). The first bin of the merge index is context coded and for the other bins, bypass coding is used.

As mentioned above, the history-based MVP may be added to the AMVP mode MVP candidate list or the merged MVP candidate list after the spatial MVP and the temporal MVP. The motion information of the previous inter-coded CU is stored in the HMVP table and used as an MVP candidate for the current CU. The HMVP table is maintained during the encoding/decoding process. Whenever there is a non-sub-block inter-coded CU, the associated motion vector information is added to the last entry of the HMVP table as a new candidate (if the HMVP table is full and there is no identical copy of the associated motion vector information in the table) while the motion vector information stored in the first entry of the HMVP table is removed from the HMVP table. Optionally, the same copy of the associated motion vector information is removed from the table before adding it to the last entry of the HMVP table.

As noted above, Intra Block Copy (IBC) may significantly improve the encoding efficiency of screen content material. Since the IBC mode is implemented as a block-level coding mode, Block Matching (BM) is performed at video encoder 20 to find the best block vector for each CU. Here, the block vector is used to indicate a displacement from the current block to a reference block that has been reconstructed within the current picture. The IBC-encoded CU is considered as a third prediction mode other than the intra-prediction mode or the inter-prediction mode.

At the CU level, the IBC mode may be signaled as IBC AMVP mode or IBC skip/merge mode as follows:

-IBC AMVP mode: the Block Vector Difference (BVD) between the actual block vector of the CU and the block vector predictor of the CU selected from the block vector candidates of the CU is encoded in the same way as described above for the motion vector difference in AMVP mode. The block vector prediction method uses two block vector candidates as predictors, one from the left neighboring block and the other from the above neighboring block (if IBC coding). When any neighboring block is not available, the default block vector will be used as the block vector predictor. A binary flag is signaled to indicate the block vector predictor index. The IBC AMVP candidate list consists of spatial candidates and HMVP candidates.

IBC skip/merge mode: the merge candidate index is used to indicate which of the block vector candidates from the merge candidate list (also referred to as "merge list") of neighboring IBC coding blocks is used to predict the block vector for the current block. The IBC merge candidate list consists of spatial candidates, HMVP candidates, and pair candidates.

Another approach to improving the coding efficiency employed by prior art coding standards is to introduce parallel processing into the video encoding/decoding process using, for example, a multi-core processor. For example, Wavefront Parallel Processing (WPP) has been introduced into HEVC as a feature to encode or decode rows of CTUs in parallel using multiple threads.

Fig. 5B is a block diagram illustrating multi-thread encoding of CTUs for multiple lines of a picture using Wavefront Parallel Processing (WPP), according to some embodiments of the present disclosure. When WPP is enabled, it is possible to process multiple rows of CTUs in parallel in a wavefront manner, where there may be a delay of two CTUs between the start of two adjacent wavefronts. For example, to encode picture 500 using WPP, a video codec, such as video encoder 20 and video decoder 30, may divide a Coding Tree Unit (CTU) of picture 500 into a plurality of wavefronts, each wavefront corresponding to a respective row of CTUs in the picture. The video codec may begin encoding/decoding the top wavefront, e.g., using a first codec core or thread. After the video codec has encoded/decoded two or more CTUs of the top wavefront, the video codec may begin encoding/decoding the next top wavefront in parallel with encoding/decoding the top wavefront, e.g., using a second parallel codec core or thread. After the video codec has encoded/decoded two or more CTUs of a second top wavefront, the video codec may begin encoding/decoding a third wavefront from the top in parallel with encoding/decoding a higher wavefront, e.g., using a third parallel codec core or thread. This pattern may continue along the wavefront in picture 500. In the present disclosure, a group of CTUs that a video codec is simultaneously encoding/decoding using WPP is referred to as a "CTU group". Thus, when a video codec encodes/decodes a picture using WPP, each CTU in the group of CTUs may belong to a unique wavefront of the picture, and the CTUs may offset at least two columns of CTUs of the picture from CTUs in a respective upper wavefront.

The video codec may initialize a context for a current wavefront to perform Context Adaptive Binary Arithmetic Coding (CABAC) of the current wavefront based on data of first two blocks of the previous wavefront and one or more elements of a slice header of a slice including a first encoded block of the current wavefront. The video codec may perform CABAC initialization of a subsequent wavefront (or row of CTUs) using the context state after encoding/decoding two CTUs of the row of CTUs above the subsequent row of CTUs. In other words, assuming that the current wavefront is not the CTU of the top row of the picture, the video codec (or more specifically, a thread of the video codec) may encode/decode at least two blocks of the wavefront above the current wavefront before starting to encode/decode the current wavefront. The video codec may then initialize a CABAC context for the current wavefront after encoding/decoding at least two blocks of the wavefront above the current wavefront. In this example, each CTU row of picture 500 is a separate partition and has an associated thread (WPP thread 1, WPP thread 2, …) such that multiple CTU rows in picture 500 may be encoded in parallel.

Because the current implementation of the HMVP table uses a global Motion Vector (MV) buffer to store previously reconstructed motion vectors, the HMVP table cannot be implemented on the WPP-enabled parallel encoding scheme described above in connection with fig. 5B. In particular, the fact that the global MV buffer is shared by all threads of the encoding/decoding process of the video codec prevents WPP threads subsequent to the first WPP thread (i.e., WPP thread 1) from being started because these WPP threads must wait for the HMVP table update of the last CTU (i.e., the rightmost CTU) according to the first WPP thread (i.e., the first CTU row) to complete.

To overcome this problem, it is proposed that the global MV buffer shared by the WPP threads is replaced by a plurality of CTU-row-specific buffers, so that each wavefront of a CTU row has its own buffer for storing an HMVP table corresponding to the CTU row processed by the respective WPP thread when WPP is enabled at the video codec. It should be noted that each CTU row has its own HMVP table, which amounts to resetting the HMVP table before encoding/decoding the first CU of the CTU row. The HMVP table reset is used to clear the HMVP table of all motion vectors resulting from the encoding/decoding of another CTU row. In one embodiment, the reset operation is to set the size of the available motion vector predictors in the HMVP table to zero. In yet another embodiment, a reset operation may be used to set the reference index of all entries in the HMVP table to an invalid value, such as-1. By doing so, regardless of which of the three modes, AMVP, merge, and skip, the construction of the MVP candidate list for the current CTU within a particular wavefront depends on the HMVP table associated with the WPP thread processing that particular wavefront. Apart from the two CTU delays described above, there is no interdependence between the different wavefronts and the construction of the motion vector candidate lists associated with the different wavefronts can be done in parallel like the WPP process depicted in fig. 5B. In other words, the HMVP table is reset to empty at the beginning of processing a particular wavefront, without affecting the encoding of another wavefront of the CTU by another WPP thread. In some cases, the HMVP table may be reset to empty prior to encoding each individual CTU. In this case, the motion vectors in the HMVP table are limited to a particular CTU, and there may be a higher likelihood of selecting the motion vectors in the HMVP table as the motion vectors of the current CU within the particular CTU.

Fig. 6 is a block diagram illustrating an example modified insertion order of merged motion vector prediction candidates (also referred to as merge candidates) in a merged motion vector prediction candidate list (also referred to as a merge list or a merge candidate list) according to some embodiments of the present disclosure. Table 602 shows an exemplary unmodified order in which various merge candidates are inserted into the merge list, and table 604 shows an exemplary modified order in which various merge candidates are inserted. For convenience, the insertion of the merge candidates into the merge list will be described as being performed by a prediction processing unit (e.g., prediction processing unit 81 of fig. 3). Each merge candidate includes motion information from the associated CU, such as (1) a motion vector, (2) a signal ("Inter _ Dir") indicating whether the reference list L0 or the reference list L1 is used, and (3) a reference index to select a reference from L0 or L1.

During exemplary unmodified merge-list construction as shown in table 602, the candidate list has a fixed length (e.g., 6). The prediction processing unit adds the merged MVP candidate to the candidate list in the following order:

1. adding spatial merge candidates

First, the prediction processing unit adds at most a predefined number (e.g., four) of spatial merging candidates from the spatial neighboring CUs a1, B1, B0, a0, and B2 to the merge list in a particular order of the spatial neighboring CUs a1, B1, B0, a0, and B2. For the positions of A1-B2, refer to FIG. 5A. If no merge candidates from the previous four CUs (A1, B1, B0, and A0) are available, the prediction processing unit only adds a merge candidate from B2. For example, if a CU is intra-coded (and thus has no motion information at all) or if the CU belongs to another slice or partition, the merging candidates of the CU are not available.

In some embodiments, after the prediction processing unit adds the merge candidate at a1, the addition of the remaining spatial merge candidates to the candidate list is subject to redundancy checking. The prediction processing unit performs a redundancy check to ensure that merge candidates having the same motion information are excluded from being added to the candidate list, so that encoding efficiency can be improved. To reduce computational complexity, not all possible merged MVP candidate pairs are checked (this would result in up to 4 | ═ 24 computations); instead, the merge candidates at B1 and a0 will be checked against the merge candidate at a1, and the merge MVP candidates at B0 and B2 will be checked against the merge MVP candidate at B1.

2. Adding temporal merge candidates

Next, the prediction processing unit adds at most a predefined number (e.g., one) of temporal merging candidates from the temporal co-located picture frames at positions T0 and T1 to the merge list in a particular order of positions T0 and T1. For positions for T0 and T1, refer to fig. 5A. When calculating the temporal merging candidate, the scaled motion vector is derived from the motion vector of the co-located picture (e.g., the co-located picture was previously encoded in the reference list L0 or L1). The scaled motion vectors for the temporal merging candidates are scaled from the selected motion vectors of the co-located PUs of the co-located pictures using POC distances.

3. Adding HMVP merge candidates

Next, the prediction processing unit adds up to a predefined number (e.g., 5) of HMVP merge candidates to the merge list. The HMVP merge candidates are derived as explained in fig. 5A and the related description. A special condition associated with adding an HMVP merge candidate is that the HMVP merge candidate cannot be added to the last position of the merge list. For example, if there are four spatial/temporal merge candidates that have been added to the merge list, and there are two HMVP merge candidates available, only one of the two HMVP candidates may be added to the merge list.

4. Adding average merge candidates

Next, the prediction processing unit adds up to a predefined number (e.g., one) of average merge candidates to the merge list. The average merge candidate is generated by averaging (e.g., averaging motion vectors) the selected set of merge candidates (e.g., pairs) that are already stored in the merge list. An average motion vector is calculated separately for each reference list of the selected merge candidate pair. If the two motion vectors of the selected merge candidate pair are available in one reference list but the reference pictures are different, the motion vector associated with the larger merge index is scaled to the reference picture of the merge candidate with the smaller merge index to calculate an average merge candidate. If only one motion vector is available, it is used as an "average" merge candidate. If no motion vector is available (e.g., no candidate has been previously inserted into the merge list), then no average merge candidate is inserted into the merge list.

5. Adding zero motion merge candidates

If there are less than the maximum allowed number of candidates (e.g., six) in the merge list after the prediction processing unit finishes adding the average merge candidate, the prediction processing unit adds one or more zero-motion merge candidates to the merge list so that the merge list reaches its maximum capacity. The zero motion merge candidate has a zero spatial displacement and a reference picture index that starts at 0 and increases by 1 for each new zero motion candidate added to the merge list (e.g., the third zero motion candidate would have a reference picture index of "3"). No redundancy check is performed on these zero motion merge candidates.

Alternatively, the prediction processing unit may add another type of merge candidate (such as non-adjacent spatial merge candidates and/or subblock merge candidates) to the merge list. For non-neighboring spatial merge candidates, spatial merge candidates from positions similar to the positions used in the spatial merge candidates (e.g., A0-A1, B0-B2) are used, but are derived from the current CU, the positions being from an external reference region containing the current CU. Sub-block merging candidates are derived by partitioning the current CU into sub-blocks, identifying spatial and temporal motion vectors associated with each sub-block, and averaging the identified spatial/temporal motion vectors.

There are several disadvantages associated with the insertion order shown in table 602. Each merge candidate in the merge list is assigned a merge candidate index, indicating the position of the respective merge candidate on the list. The smaller the merge index (e.g., indicating that the corresponding merge candidate is near the top of the merge list), the less memory (smaller number of bits) is required for the merge index. Therefore, it is preferable to store merge candidates for generating more accurate motion estimation near the top of the merge list (e.g., assign them smaller merge indices) in order to improve coding efficiency. It has been observed that the first several average merge candidates may yield better prediction efficiency than the non-neighboring spatial merge candidates and the HMVP merge candidate. In addition, there is a special condition associated with inserting an HMVP merge candidate into the merge list-the HMVP merge candidate cannot be inserted into the merge list as the last merge candidate. This special condition reduces coding efficiency because in some cases there is no average merge candidate available for the last position on the merge list (e.g., only one spatial/temporal merge candidate has been added) and no available HMVP can be inserted into the merge list.

The modified merge candidate insertion order shown in table 604 improves coding efficiency by allowing insertion of one or more average merge candidates before non-adjacent spatial merge candidates and HMVP merge candidates.

In some embodiments, the prediction processing unit calculates the set of average merge candidates after completing adding the spatial and temporal merge candidates to the merge list. The prediction processing unit then adds a subset of the average merge candidates to the merge list. For example, the subset of average merge candidates may be the first two average merge candidates calculated using a particular method (e.g., pairwise averaging). Then, the prediction processing unit adds the HMVP merge candidate, the remaining average merge candidates, and the zero motion merge candidate, respectively (assuming that there is a sufficient position in the merge list).

In some embodiments, the prediction processing unit discards the remaining average merge candidates. Thus, after averaging the subset of merge candidates, the prediction processing unit adds an HMVP merge candidate and a zero motion merge candidate.

In some embodiments, the prediction processing unit performs a group-based redundancy check when building the merge list. The prediction processing unit performs a full redundancy check, a partial redundancy check, or a no redundancy check for each candidate group. A full redundancy check means that a redundancy check is performed for each merge candidate with respect to all other candidates within the same group. Partial checking means that redundancy checking is performed on some merging candidates relative to some other merging candidates within the same group. The redundancy check-free means that the redundancy check is not performed on any merging candidates within the same group.

In some embodiments, the prediction processing unit classifies the merging candidates as follows:

the prediction processing unit performs a partial redundancy check on group 1 and group 4, and does not perform a redundancy check on group 2 and group 3.

Alternatively, the prediction processing unit performs a partial redundancy check on group 1 and group 4, does not perform a redundancy check on group 2, and performs a full redundancy check on group 3.

for example, the non-adjacent spatial merge candidates may be divided into a first subset, a second subset, … … nth subset based on their relative positions with respect to the used reference frame.

In some embodiments, the prediction processing unit performs a partial redundancy check on group 1, group 4-2, group 4-3 … …, group 4-N, does not perform a redundancy check on group 2, and performs a full redundancy check on group 3.

the prediction processing unit performs a partial redundancy check on group 1-1, group 1-2, … … group 1-N, does not perform a redundancy check on group 2, and performs a full redundancy check on group 3. It should be noted that the partial redundancy check described above is for illustrative purposes, and that other possible redundancy check scenarios will be apparent to those skilled in the art.

Fig. 7 is a flow diagram illustrating an exemplary process 700 for a video decoder implementing a technique for constructing a merge motion vector prediction candidate list (also referred to as a merge list) according to some embodiments of the present disclosure. For convenience, process 700 will be described as being performed by a prediction processing unit (e.g., prediction processing unit 81 of fig. 3).

In some embodiments, the merge list may be set to have a maximum length (e.g., store up to 6 merge candidates), giving it an array-like behavior. Alternatively, the merge list is configured to store all the merge candidates calculated by the prediction processing unit so as to be given a list-like characteristic.

The prediction processing unit may use one of many methods to compute the merge candidates: generating spatial merge candidates using spatially neighboring blocks, generating temporal merge candidates using temporally collocated blocks, averaging motion vectors of existing merge candidates to generate average merge candidates, storing past motion information to generate history-based motion vector prediction (HMVP) candidates, and so on. Each merge candidate includes motion information indicating: (1) a motion vector pointing from a block associated with a merge candidate to another block in the video signal, (2) a reference list (e.g., reference list L0 or L1), (3) a flag indicating which reference list is used (e.g., "inter dir" parameter), and (4) a reference index indicating which reference frame of the reference list is used.

Each merge candidate in the merge list is associated with a merge index. The higher the merge candidate is in the merge list, the smaller its associated merge index will be. For example, the first candidate to add to the merge list is represented by a merge index of "0". As a result, merge candidates that are added earlier in the merge list (and thus have smaller merge indices) will require less bit size to store their associated merge indices. For example, "0" requires only one bit to store ("0"), but "2" requires two bits to store ("10"), and 4 requires four bits to store "100"). Since the encoder stores the merge index of the merge candidate to be used for a specific decoding process in the video signal, it is preferable that the merge candidate that is more likely to generate an accurate prediction is stored in the merge list early.

As a first step for constructing a merge list, the prediction processing unit receives decoding information for a current coding block encoded in merge mode from a bitstream representing the encoded video signal (710). As explained with reference to fig. 5A, if motion information for decoding a coding block is shared from another coding block (e.g., a spatially neighboring coding block or a temporally co-located coding block) instead of being directly calculated (e.g., the coding block is encoded using an inter mode), the coding block is encoded in a merge mode.

Next, the prediction processing unit generates a merge list (also referred to as a merge candidate list) for the current coding block (720). The prediction processing unit starts with a null merge list and calculates a plurality of sets of merge candidates to be added to the merge list. Each merge candidate includes respective motion information (e.g., motion vectors, reference lists, reference frame indices, etc.) and may be used by a motion compensation unit (e.g., motion compensation unit 82) to generate a prediction unit for the current coding block.

To add the first set of merging candidates to the merge list, the prediction processing unit first checks whether at least one spatially neighboring block of the currently encoded block satisfies a first condition. If the first condition is satisfied, the prediction processing unit inserts a first set of spatial merge candidates into the merge list (730). Five possible spatial merge candidates are described with reference to fig. 5A and 6. The prediction processing unit checks five spatial merge candidates, such as a1- > B1- > B0- > a0- > B2, in a predefined order. Other orders of examining spatially adjacent blocks are also possible.

In some embodiments, the first condition includes that spatially neighboring blocks (e.g., a0-a1, B0-B2) are inter-coded and contain motion information.

In some embodiments, the number of spatial merge candidates that can be added to the merge list is limited, and the first condition includes that the maximum number of spatial merge candidates has not been reached in the merge list. For example, in some cases, the prediction processing unit may only add up to four spatial merge candidates to the merge list and check all possible neighboring blocks in a predefined order (a0-a1, B0-B2).

In some embodiments, the first condition includes that no previous spatial merge candidates have been added to the merge list. For example, the first condition may be that the merge candidate at B2 may only be added to the merge if no merge candidates at A0-A1 and B0-B1 have already been added.

After completing the addition of the spatial merge candidate (if any spatial merge candidate to be added exists), the prediction processing unit checks whether at least one temporally co-located block of the current block satisfies a second condition. If the second condition is satisfied, the prediction processing unit adds a second set of temporal merge candidates to the merge list, each temporal merge candidate including motion information at the temporally co-located block (740). Two possible block positions (e.g., T0 and T1) for generating temporal merging candidates are described with reference to fig. 5A and 6. The prediction processing unit checks the two temporal merging candidates in a predefined order (T0- > T1 or T1- > T0).

In some embodiments, the second condition includes that the temporal co-located blocks (e.g., T0 and T1) are inter-coded and contain motion information.

In some embodiments, the number of temporal merge candidates that can be added to the merge list is limited, and the second condition includes that the maximum number of temporal merge candidates has not been reached in the merge list. For example, in some cases, the prediction processing unit may only add at most one temporal merge candidate to the merge list and check T0 and T1 in a predefined order. Once the first temporal merging candidate has been added, the prediction processing unit abandons the addition of the second temporal merging candidate even if the second temporal merging candidate contains valid motion information. Alternatively, the prediction processing unit may add all temporal merging candidates.

After adding the spatial and temporal merging candidates to the merge list (if there are any spatial and temporal merging candidates to be added), the prediction processing unit determines whether an average merge candidate can be calculated and added to the merge list (750). The prediction processing unit first checks whether at least two merging candidates exist in the merging list (e.g., spatial or temporal merging candidates). If so, the prediction processing unit generates an average merge candidate based on the existing at least two merge candidates.

In some embodiments, the prediction processing unit performs pairwise averaging to generate average merge candidates. To this end, the prediction processing unit divides the existing merge candidates into pairs based on their positions in the merge list, and averages the motion vectors associated with each pair of merge candidates. The prediction processing unit may select a reference frame associated with any one of the pair of merging candidates as a reference frame for the average merging candidate.

In some embodiments, the prediction processing unit generates the average merge candidate by: the motion vectors associated with all existing merge candidates are averaged and a reference frame is selected from the merge candidates as the reference frame for the averaged merge candidate.

In some embodiments, there are multiple average merge candidates available. The prediction processing unit may append all average merge candidates to the merge list or append only a subset of the average merge candidates to the merge list. After the prediction processing unit appends the HMVP merge candidate, the remaining average merge candidates may be appended to the merge list.

After the prediction processing unit adds the average merge candidate to the merge list (if any average merge candidate to be added exists), the prediction processing unit determines whether there are less than the maximum allowable number of merge candidates in the merge list. If so, the prediction processing unit adds a fourth set of HMVP merge candidates to the merge list (760). The HMVP merge candidates are calculated as described with reference to fig. 5A.

In some embodiments, a special condition is associated with the fourth set of HMVP merge candidates-an HMVP merge candidate may not be the last merge candidate in the merge list.

In some embodiments, there is a maximum allowed number for the set of HMVP merge candidates. For example, the prediction processing unit may only add up to 5 HMVP merge candidates.

In some embodiments, after the prediction processing unit adds the fourth set of HMVP merge candidates to the merge list, the prediction processing unit further adds one or more average merge candidates to the merge list. For example, the prediction processing unit may add the first two average merge candidates to the merge list before the HMVP merge candidate and the remaining average merge candidates after the HMVP merge candidate.

In some embodiments, the prediction processing unit performs redundancy check before adding the generated merge candidate (spatial merge candidate, temporal merge candidate, average merge candidate, etc.) to the merge list. Since the redundancy check is already performed when the HMVP merge candidate is first generated, the redundancy check is not performed on the HMVP merge candidate.

After generating merge candidates and adding them to the merge list, the prediction processing unit receives a merge index from the bitstream (770). The merge index indicates which merge candidate in the merge list is to be used for prediction block generation.

Finally, the prediction processing unit generates a prediction block for the current coding block by using the identified merge candidate (e.g., identified by the merge index) and the residual block (780). For example, the prediction processing unit may use the motion information to find a portion of another frame and add the frame portion to the residual block.

In one or more examples, the functions described may be implemented in hardware, software, firmware, or any combination thereof. If implemented in software, the functions may be stored on or transmitted over as one or more instructions or code on a computer-readable medium and executed by a hardware-based processing unit. The computer readable medium may include a computer readable storage medium, which corresponds to a tangible medium, such as a data storage medium, or a communication medium including any medium that facilitates transfer of a computer program from one place to another (e.g., according to a communication protocol). In this manner, the computer-readable medium may generally correspond to (1) a non-transitory tangible computer-readable storage medium, or (2) a communication medium, such as a signal or carrier wave. A data storage medium may be any available medium that can be accessed by one or more computers or one or more processors to retrieve instructions, code and/or data structures for implementing the implementations described herein. The computer program product may include a computer-readable medium.

The terminology used in the description of the embodiments herein is for the purpose of describing particular embodiments only and is not intended to limit the scope of the claims. As used in the description of the embodiments and the appended claims, the singular forms "a," "an," and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise. It will also be understood that the term "and/or" as used herein refers to and encompasses any and all possible combinations of one or more of the associated listed items. It will be further understood that the terms "comprises" and/or "comprising …," when used in this specification, specify the presence of stated features, elements, and/or components, but do not preclude the presence or addition of one or more other features, elements, components, and/or groups thereof.

It will also be understood that, although the terms first, second, etc. may be used herein to describe various elements, these elements should not be limited by these terms. These terms are only used to distinguish one element from another. For example, a first electrode may be referred to as a second electrode, and similarly, a second electrode may be referred to as a first electrode, without departing from the scope of embodiments. The first electrode and the second electrode are both electrodes, but they are not the same electrode.

The description of the present application has been presented for purposes of illustration and description, and is not intended to be exhaustive or limited to the invention in the form disclosed. Many modifications, variations and alternative embodiments will become apparent to those of ordinary skill in the art having the benefit of the teachings presented in the foregoing descriptions and the associated drawings. The embodiment was chosen and described in order to best explain the principles of the invention, the practical application, and to enable others of ordinary skill in the art to understand the invention for various embodiments and with the best mode contemplated for use with the general principles and with various modifications as are suited to the particular use contemplated. Therefore, it is to be understood that the scope of the claims is not to be limited to the specific examples of the embodiments disclosed and that modifications and other embodiments are intended to be included within the scope of the appended claims.

Claims

1. A method performed at a computing device having one or more processors and memory storing a plurality of programs to be executed by the one or more processors for decoding a video signal, the method comprising:

receiving decoding information for a current coding block coded in a merge mode from a bitstream representing an encoded video signal;

generating a merge list for a current coding block, wherein respective merge candidates in the merge list include respective motion information for constructing a prediction block, and wherein generating the merge list comprises:

in response to determining that at least one spatially neighboring block of a currently encoded block satisfies a first condition, inserting a first set of spatial merge candidates into the merge list using respective motion information of the at least one spatially neighboring block;

responsive to determining that at least one temporally collocated block of a current coding block satisfies a second condition, appending a second set of temporal merging candidates to the merge list using respective motion information of the at least one temporally collocated block;

in response to determining that there are at least two merge candidates in the merge list, appending a third set of average merge candidates to the merge list based on the at least two merge candidates;

responsive to determining that there are less than a maximum allowed number of merge candidates in the merge list, appending a fourth set of history-based motion vector prediction (HMVP) merge candidates to the merge list; and

receiving a merge index from the bitstream that identifies a merge candidate in the merge list; and

a prediction block for the current coding block is generated using the identified merge candidates.

2. The method of claim 1, further comprising:

appending one or more further average merge candidates to the merge list after the fourth set of HMVP merge candidates.

3. The method of claim 2, further comprising:

appending one or more zero motion merge candidates to the merge list after the one or more further average merge candidates such that the merge list reaches the maximum allowed number of merge candidates.

4. The method of claim 1, further comprising:

performing redundancy check on the first set of spatial merging candidates, the second set of temporal merging candidates, the third set of average merging candidates, or the fourth set of HMVP merging candidates, respectively, to remove redundant merging candidates.

5. The method of claim 1, wherein the average merge candidate is generated by performing a pairwise averaging of a first set of spatial merge candidates and a second set of temporal merge candidates.

6. The method of claim 1, wherein the maximum allowed number of merge candidates is 6.

7. A computing device, comprising:

one or more processors;

a memory coupled to the one or more processors; and

a plurality of programs stored in the memory, which when executed by the one or more processors, cause the computing device to perform operations comprising:

8. The computing device of claim 7, wherein the operations further comprise:

9. The computing device of claim 8, wherein the operations further comprise:

10. The computing device of claim 7, wherein the operations further comprise:

11. The computing device of claim 7, wherein the average merge candidate is generated by performing a pairwise average of a first set of spatial merge candidates and a second set of temporal merge candidates.

12. The computing device of claim 7, wherein the maximum allowed number of merge candidates is 6.

13. A non-transitory computer readable storage medium storing a plurality of programs for execution by a computing device having one or more processors, wherein the plurality of programs, when executed by the one or more processors, cause the computing device to perform operations comprising:

14. The non-transitory computer-readable storage medium of claim 13, wherein the operations further comprise:

15. The non-transitory computer-readable storage medium of claim 14, wherein the operations further comprise:

16. The non-transitory computer-readable storage medium of claim 13, wherein the operations further comprise:

17. The non-transitory computer-readable storage medium of claim 13, wherein the average merge candidate is generated by performing a pairwise average of a first set of spatial merge candidates and a second set of temporal merge candidates.

18. The non-transitory computer-readable storage medium of claim 13, wherein the maximum allowed number of merge candidates is 6.