DIFFERENCE CALCULATION BASED ON PARTIAL POSITION

Abstract

Difference calculation based on partial position is described. In a representative aspect, a method of video processing comprising: calculating, during a conversion between a current block of video and a bitstream representation of the current block, differences between two reference blocks associated with the current block or differences between two reference sub-blocks associated with a sub-block within the current block based on representative positions of the reference blocks or representative positions of the reference sub-blocks; and performing the conversion based on the differences.

Description

This application is a continuation of International Application No. PCT/CN2019/119634, filed on Nov. 20, 2019, which claims the priority to and benefits of International Patent Applications No. PCT/CN2018/116371, filed on Nov. 20, 2018, No. PCT/CN2019/070062, filed on Jan. 2, 2019, and No. PCT/CN2019/072060, filed on Jan. 16, 2019. All the aforementioned patent applications are hereby incorporated by reference in their entireties.

TECHNICAL FIELD

This patent document relates to video coding techniques, devices and systems.

BACKGROUND

Motion compensation (MC) is a technique in video processing to predict a frame in a video, given the previous and/or future frames by accounting for motion of the camera and/or objects in the video. Motion compensation can be used in the encoding/decoding of video data for video compression.

SUMMARY

This document discloses methods, systems, and devices related to the use of motion compensation in video coding and decoding.

In one example aspect, a method for video processing is discloses. The method comprises: calculating, during a conversion between a current block of video and a bitstream representation of the current block, differences between two reference blocks associated with the current block or differences between two reference sub-blocks associated with a sub-block within the current block based on representative positions of the reference blocks or representative positions of the reference sub-blocks; and performing the conversion based on the differences.

In one example aspect, a method for video processing is discloses. The method comprises: making a decision, based on a determination that a current block of a video is coded using a specific coding mode, regarding a selective enablement of a decoder side motion vector derivation (DMVD) tool for the current block, wherein the DMVD tool derives a refinement of motion information signaled in a bitstream representation of the video; and performing, based on the decision, a conversion between the current block and the bitstream representation.

In one example aspect, a video processing method is disclosed. The method includes generating, using a multi-step refinement process, multiple refinement values of motion vector information based on decoded motion information from a bitstream representation of a current video block, and reconstructing the current video block or decoding other video blocks based on multiple refinement values.

In another example aspect, another video processing method is disclosed. The method includes performing, for conversion between a current block and a bitstream representation of the current block, a multi-step refinement process for a sub-block of the current block and a temporal gradient modification process between two prediction blocks of the sub-block, wherein, using the multi-step refinement process, multiple refinement values of motion vector information based on decoded motion information from a bitstream representation of the current video block and performing the conversion between the current block and the bitstream representation based on refinement values.

In yet another example aspect, another video processing method is disclosed. The method includes determining, using a multi-step decoder-side motion vector refinement process a current video block, a final motion vector and performing conversion between the current block and the bitstream representation using the final motion vector.

In yet another aspect, another method of video processing is disclosed. The method includes applying, during conversion between a current video block and a bitstream representation of the current video block; multiple different motion vector refinement processes to different sub-blocks of the current video block and performing conversion between the current block and the bitstream representation using a final motion vector for the current video block generated from the multiple different motion vector refinement processes.

In yet another aspect, another method of video processing is disclosed. The method includes performing a conversion between a current video block and a bitstream representation of the current video block using a rule that limits a maximum number of sub-blocks that a coding unit or a prediction unit in case that the current video block is coded using a sub-block based coding tool,

wherein the sub-block based coding tool includes one or more of affine coding, advanced temporal motion vector predictor, bi-directional optical flow or a decoder-side motion vector refinement coding tool.

In yet another example aspect, another method of video processing is disclosed. The method includes performing a conversion between a current video block and a bitstream representation of the current video block using a rule that specifies to use different partitioning for chroma components of the current video block than a luma component of the current video block in case that the current video block is coded using a sub-block based coding tool, wherein the sub-block based coding tool includes one or more of affine coding, advanced temporal motion vector predictor, bi-directional optical flow or a decoder-side motion vector refinement coding tool.

In yet another example aspect, another method of video processing is disclosed. The method includes determining, in an early termination stage of a bi-directional optical flow (BIO) technique or a decoder-side motion vector refinement (DMVR) technique, differences between reference video blocks associated with a current video block, and performing further processing of the current video block based on the differences.

In yet another representative aspect, the various techniques described herein may be embodied as a computer program product stored on a non-transitory computer readable media. The computer program product includes program code for carrying out the methods described herein.

In yet another representative aspect, a video decoder apparatus may implement a method as described herein.

The details of one or more implementations are set forth in the accompanying attachments, the drawings, and the description below. Other features will be apparent from the description and drawings, and from the claims.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 shows an example of a derivation process for merge candidates list construction.

FIG. 2 shows example positions of spatial merge candidates.

FIG. 3 shows examples of Candidate pairs considered for redundancy check of spatial merge candidates.

FIG. 4 shows example Positions for the second PU of N×2N and 2N×N partitions.

FIG. 5 is an Illustration of motion vector scaling for temporal merge candidate.

FIG. 6 shows examples of Candidate positions for temporal merge candidate, C0 and C1.

FIG. 7 shows an example of combined bi-predictive merge candidate

FIG. 8 shows an example of a derivation process for motion vector prediction candidates.

FIG. 9 is an example illustration of motion vector scaling for spatial motion vector candidate.

FIG. 10 illustrates an example of advanced temporal motion vector predictor (ATMVP) for a Coding Unit (CU).

FIG. 11 shows an Example of one CU with four sub-blocks (A-D) and its neighbouring blocks (a-d).

FIG. 12 is an example Illustration of sub-blocks where OBMC applies.

FIG. 13 shows an example of Neighbouring samples used for deriving IC parameters.

FIG. 14 shows an example of a simplified affine motion model.

FIG. 15 shows an example of affine MVF per sub-block.

FIG. 16 shows an example of a Motion Vector Predictor (MV) for AF_INTER mode.

FIG. 17A-17B shows examples of candidates for AF_MERGE mode.

FIG. 18 shows example process for bilateral matching.

FIG. 19 shows example process of template matching.

FIG. 20 illustrates an implementation of unilateral motion estimation (ME) in frame rate upconversion (FRUC).

FIG. 21 illustrates an embodiment of an Ultimate Motion Vector Expression (UMVE) search process.

FIG. 22 shows examples of UMVE search points.

FIG. 23 shows an example of distance index and distance offset mapping.

FIG. 24 shows an example of an optical flow trajectory.

FIG. 25A-25B show examples of Bi-directional Optical flow (BIO) w/o block extension: a) access positions outside of the block; b) padding used in order to avoid extra memory access and calculation.

FIG. 26 illustrates an example of using Decoder-side motion vector refinement (DMVR) based on bilateral template matching.

FIG. 27 shows an example of interweaved prediction.

FIG. 28 shows an example of iterative motion vector refinement for BIO.

FIG. 29 is a block diagram of a hardware platform for implementing the video coding or decoding techniques described in the present document.

FIG. 30 shows an example of a hardware platform for implementing methods and techniques described in the present document.

FIG. 31 is a flowchart of an example method of video processing.

FIG. 32 is a flowchart of an example method of video processing.

FIG. 33 is a flowchart of an example method of video processing.

DETAILED DESCRIPTION

The present document provides several techniques that can be embodied into digital video encoders and decoders. Section headings are used in the present document for clarity of understanding and do not limit scope of the techniques and embodiments disclosed in each section only to that section.

1. Summary

The present document is related to video coding technologies. Specifically, it is related to motion compensation in video coding. The disclosed techniques may be applied to the existing video coding standard like HEVC, or the standard (Versatile Video Coding) to be finalized. It may be also applicable to future video coding standards or video codec.

In the present document, the term “video processing” may refer to video encoding, video decoding, video compression or video decompression. For example, video compression algorithms may be applied during conversion from pixel representation of a video to a corresponding bitstream representation or vice versa.

2. Introduction

Video coding standards have evolved primarily through the development of the well-known ITU-T and ISO/IEC standards. The ITU-T produced H.261 and H.263, ISO/IEC produced MPEG-1 and MPEG-4 Visual, and the two organizations jointly produced the H.262/MPEG-2 Video and H.264/MPEG-4 Advanced Video Coding (AVC) and H.265/HEVC standards. Since H.262, the video coding standards are based on the hybrid video coding structure wherein temporal prediction plus transform coding are utilized. To explore the future video coding technologies beyond HEVC, Joint Video Exploration Team (JVET) was founded by VCEG and MPEG jointly in 2015. Since then, many new methods have been adopted by JVET and put into the reference software named Joint Exploration Model (JEM). In April 2018, the Joint Video Expert Team (JVET) between VCEG (Q6/16) and ISO/IEC JTC1 SC29/WG11 (MPEG) was created to work on the VVC standard targeting at 50% bitrate reduction compared to HEVC.

2.1 Inter Prediction in HEVC/H.265

Each inter-predicted PU has motion parameters for one or two reference picture lists. Motion parameters include a motion vector and a reference picture index. Usage of one of the two reference picture lists may also be signalled using inter_pred_idc. Motion vectors may be explicitly coded as deltas relative to predictors.

When a CU is coded with skip mode, one PU is associated with the CU, and there are no significant residual coefficients, no coded motion vector delta or reference picture index. A merge mode is specified whereby the motion parameters for the current PU are obtained from neighbouring PUs, including spatial and temporal candidates. The merge mode can be applied to any inter-predicted PU, not only for skip mode. The alternative to merge mode is the explicit transmission of motion parameters, where motion vector (to be more precise, motion vector difference compared to a motion vector predictor), corresponding reference picture index for each reference picture list and reference picture list usage are signalled explicitly per each PU. Such a mode is named Advanced motion vector prediction (AMVP) in this disclosure.

When signalling indicates that one of the two reference picture lists is to be used, the PU is produced from one block of samples. This is referred to as ‘uni-prediction’. Uni-prediction is available both for P-slices and B-slices.

When signalling indicates that both of the reference picture lists are to be used, the PU is produced from two blocks of samples. This is referred to as ‘bi-prediction’. Bi-prediction is available for B-slices only.

The following text provides the details on the inter prediction modes specified in HEVC. The description will start with the merge mode.

2.1.1 Merge Mode

2.1.1.1 Derivation of Candidates for Merge Mode

When a PU is predicted using merge mode, an index pointing to an entry in the merge candidates list is parsed from the bitstream and used to retrieve the motion information. The construction of this list is specified in the HEVC standard and can be summarized according to the following sequence of steps:

Step 1: Initial candidates derivation

• • Step 1.1: Spatial candidates derivation
  • Step 1.2: Redundancy check for spatial candidates
  • Step 1.3: Temporal candidates derivation

Step 2: Additional candidates insertion

• • Step 2.1: Creation of bi-predictive candidates
  • Step 2.2: Insertion of zero motion candidates

These steps are also schematically depicted in FIG. 1. For spatial merge candidate derivation, a maximum of four merge candidates are selected among candidates that are located in five different positions. For temporal merge candidate derivation, a maximum of one merge candidate is selected among two candidates. Since constant number of candidates for each PU is assumed at decoder, additional candidates are generated when the number of candidates obtained from step 1 does not reach the maximum number of merge candidate (MaxNumMergeCand) which is signalled in slice header. Since the number of candidates is constant, index of best merge candidate is encoded using truncated unary binarization (TU). If the size of CU is equal to 8, all the PUs of the current CU share a single merge candidate list, which is identical to the merge candidate list of the 2N×2N prediction unit.

In the following, the operations associated with the aforementioned steps are detailed.

2.1.1.2 Spatial Candidates Derivation

In the derivation of spatial merge candidates, a maximum of four merge candidates are selected among candidates located in the positions depicted in FIG. 2. The order of derivation is A₁, B₁, B₀, A₀ and B₂. Position B₂ is considered only when any PU of position A₁, B₁, B₀, A₀ is not available (e.g. because it belongs to another slice or tile) or is intra coded. After candidate at position A₁ is added, the addition of the remaining candidates is subject to a redundancy check which ensures that candidates with same motion information are excluded from the list so that coding efficiency is improved. To reduce computational complexity, not all possible candidate pairs are considered in the mentioned redundancy check. Instead only the pairs linked with an arrow in FIG. 3 are considered and a candidate is only added to the list if the corresponding candidate used for redundancy check has not the same motion information. Another source of duplicate motion information is the “second PU” associated with partitions different from 2N×2N. As an example, FIG. 4 depicts the second PU for the case of N×2N and 2N×N, respectively. When the current PU is partitioned as N×2N, candidate at position A₁ is not considered for list construction. In fact, by adding this candidate will lead to two prediction units having the same motion information, which is redundant to just have one PU in a coding unit. Similarly, position B₁ is not considered when the current PU is partitioned as 2N×N.

2.1.1.3 Temporal Candidates Derivation

In this step, only one candidate is added to the list. Particularly, in the derivation of this temporal merge candidate, a scaled motion vector is derived based on co-located PU belonging to the picture which has the smallest POC difference with current picture within the given reference picture list. The reference picture list to be used for derivation of the co-located PU is explicitly signalled in the slice header. The scaled motion vector for temporal merge candidate is obtained as illustrated by the dotted line in FIG. 5 which is scaled from the motion vector of the co-located PU using the POC distances, tb and td, where tb is defined to be the POC difference between the reference picture of the current picture and the current picture and td is defined to be the POC difference between the reference picture of the co-located picture and the co-located picture. The reference picture index of temporal merge candidate is set equal to zero. A practical realization of the scaling process is described in the HEVC specification. For a B-slice, two motion vectors, one is for reference picture list 0 and the other is for reference picture list 1, are obtained and combined to make the bi-predictive merge candidate.

In the co-located PU (Y) belonging to the reference frame, the position for the temporal candidate is selected between candidates C₀ and C₁, as depicted in FIG. 6. If PU at position C₀ is not available, is intra coded, or is outside of the current CTU row, position C₁ is used. Otherwise, position C₀ is used in the derivation of the temporal merge candidate.

2.1.1.4 Additional Candidates Insertion

Besides spatial and temporal merge candidates, there are two additional types of merge candidates: combined bi-predictive merge candidate and zero merge candidate. Combined bi-predictive merge candidates are generated by utilizing spatial and temporal merge candidates. Combined bi-predictive merge candidate is used for B-Slice only. The combined bi-predictive candidates are generated by combining the first reference picture list motion parameters of an initial candidate with the second reference picture list motion parameters of another. If these two tuples provide different motion hypotheses, they will form a new bi-predictive candidate. As an example, FIG. 7 depicts the case when two candidates in the original list (on the left), which have mvL0 and refIdxL0 or mvL1 and refIdxL1, are used to create a combined bi-predictive merge candidate added to the final list (on the right). There are numerous rules regarding the combinations which are considered to generate these additional merge candidates.

Zero motion candidates are inserted to fill the remaining entries in the merge candidates list and therefore hit the MaxNumMergeCand capacity. These candidates have zero spatial displacement and a reference picture index which starts from zero and increases every time a new zero motion candidate is added to the list. The number of reference frames used by these candidates is one and two for uni and bi-directional prediction, respectively. Finally, no redundancy check is performed on these candidates.

2.1.1.5 Motion Estimation Regions for Parallel Processing

To speed up the encoding process, motion estimation can be performed in parallel whereby the motion vectors for all prediction units inside a given region are derived simultaneously. The derivation of merge candidates from spatial neighbourhood may interfere with parallel processing as one prediction unit cannot derive the motion parameters from an adjacent PU until its associated motion estimation is completed. To mitigate the trade-off between coding efficiency and processing latency, HEVC defines the motion estimation region (MER) whose size is signalled in the picture parameter set using the “log 2_parallel_merge_level_minus2” syntax element. When a MER is defined, merge candidates falling in the same region are marked as unavailable and therefore not considered in the list construction.

2.1.2 AMVP

AMVP exploits spatio-temporal correlation of motion vector with neighbouring PUs, which is used for explicit transmission of motion parameters. For each reference picture list, a motion vector candidate list is constructed by firstly checking availability of left, above temporally neighbouring PU positions, removing redundant candidates and adding zero vector to make the candidate list to be constant length. Then, the encoder can select the best predictor from the candidate list and transmit the corresponding index indicating the chosen candidate. Similarly with merge index signalling, the index of the best motion vector candidate is encoded using truncated unary. The maximum value to be encoded in this case is 2 (see FIG. 8). In the following sections, details about derivation process of motion vector prediction candidate are provided.

FIG. 8 summarizes derivation process for motion vector prediction candidate.

2.1.2.1 Derivation of AMVP Candidates

In motion vector prediction, two types of motion vector candidates are considered: spatial motion vector candidate and temporal motion vector candidate. For spatial motion vector candidate derivation, two motion vector candidates are eventually derived based on motion vectors of each PU located in five different positions as depicted in FIG. 2.

For temporal motion vector candidate derivation, one motion vector candidate is selected from two candidates, which are derived based on two different co-located positions. After the first list of spatio-temporal candidates is made, duplicated motion vector candidates in the list are removed. If the number of potential candidates is larger than two, motion vector candidates whose reference picture index within the associated reference picture list is larger than 1 are removed from the list. If the number of spatio-temporal motion vector candidates is smaller than two, additional zero motion vector candidates is added to the list.

2.1.2.2 Spatial Motion Vector Candidates

In the derivation of spatial motion vector candidates, a maximum of two candidates are considered among five potential candidates, which are derived from PUs located in positions as depicted in FIG. 2, those positions being the same as those of motion merge. The order of derivation for the left side of the current PU is defined as A₀, A₁, and scaled A₀, scaled A₁. The order of derivation for the above side of the current PU is defined as B₀, B₁, B₂, scaled B₀, scaled B₁, scaled B₂. For each side there are therefore four cases that can be used as motion vector candidate, with two cases not required to use spatial scaling, and two cases where spatial scaling is used. The four different cases are summarized as follows.

• • No spatial scaling
    • (1) Same reference picture list, and same reference picture index (same POC)
    • (2) Different reference picture list, but same reference picture (same POC)
  • Spatial scaling
    • (3) Same reference picture list, but different reference picture (different POC)
    • (4) Different reference picture list, and different reference picture (different POC)

The no-spatial-scaling cases are checked first followed by the spatial scaling. Spatial scaling is considered when the POC is different between the reference picture of the neighbouring PU and that of the current PU regardless of reference picture list. If all PUs of left candidates are not available or are intra coded, scaling for the above motion vector is allowed to help parallel derivation of left and above MV candidates. Otherwise, spatial scaling is not allowed for the above motion vector.

In a spatial scaling process, the motion vector of the neighbouring PU is scaled in a similar manner as for temporal scaling, as depicted as FIG. 9. The main difference is that the reference picture list and index of current PU is given as input; the actual scaling process is the same as that of temporal scaling.

2.1.2.3 Temporal Motion Vector Candidates

Apart for the reference picture index derivation, all processes for the derivation of temporal merge candidates are the same as for the derivation of spatial motion vector candidates (see FIG. 6). The reference picture index is signalled to the decoder.

2.2 New Inter Prediction Methods in JEM

2.2.1 Sub-CU Based Motion Vector Prediction

In the JEM with QTBT, each CU can have at most one set of motion parameters for each prediction direction. Two sub-CU level motion vector prediction methods are considered in the encoder by splitting a large CU into sub-CUs and deriving motion information for all the sub-CUs of the large CU. Alternative temporal motion vector prediction (ATMVP) method allows each CU to fetch multiple sets of motion information from multiple blocks smaller than the current CU in the collocated reference picture. In spatial-temporal motion vector prediction (STMVP) method motion vectors of the sub-CUs are derived recursively by using the temporal motion vector predictor and spatial neighbouring motion vector.

To preserve more accurate motion field for sub-CU motion prediction, the motion compression for the reference frames is currently disabled.

2.2.1.1 Alternative Temporal Motion Vector Prediction

In the alternative temporal motion vector prediction (ATMVP) method, the motion vectors temporal motion vector prediction (TMVP) is modified by fetching multiple sets of motion information (including motion vectors and reference indices) from blocks smaller than the current CU. As shown in the FIG. 10, the sub-CUs are square N×N blocks (N is set to 4 by default).

ATMVP predicts the motion vectors of the sub-CUs within a CU in two steps. The first step is to identify the corresponding block in a reference picture with a so-called temporal vector. The reference picture is called the motion source picture. The second step is to split the current CU into sub-CUs and obtain the motion vectors as well as the reference indices of each sub-CU from the block corresponding to each sub-CU, as shown in FIG. 10.

In the first step, a reference picture and the corresponding block is determined by the motion information of the spatial neighbouring blocks of the current CU. To avoid the repetitive scanning process of neighbouring blocks, the first merge candidate in the merge candidate list of the current CU is used. The first available motion vector as well as its associated reference index are set to be the temporal vector and the index to the motion source picture. This way, in ATMVP, the corresponding block may be more accurately identified, compared with TMVP, wherein the corresponding block (sometimes called collocated block) is always in a bottom-right or center position relative to the current CU.

In the second step, a corresponding block of the sub-CU is identified by the temporal vector in the motion source picture, by adding to the coordinate of the current CU the temporal vector. For each sub-CU, the motion information of its corresponding block (the smallest motion grid that covers the center sample) is used to derive the motion information for the sub-CU. After the motion information of a corresponding N×N block is identified, it is converted to the motion vectors and reference indices of the current sub-CU, in the same way as TMVP of HEVC, wherein motion scaling and other procedures apply. For example, the decoder checks whether the low-delay condition (i.e. the POCs of all reference pictures of the current picture are smaller than the POC of the current picture) is fulfilled and possibly uses motion vector MV_x (the motion vector corresponding to reference picture list X) to predict motion vector MV_y (with X being equal to 0 or 1 and Y being equal to 1-X) for each sub-CU.

2.2.1.2 Spatial-Temporal Motion Vector Prediction

In this method, the motion vectors of the sub-CUs are derived recursively, following raster scan order. FIG. 11 illustrates this concept. Let us consider an 8×8 CU which contains four 4×4 sub-CUs A, B, C, and D. The neighbouring 4×4 blocks in the current frame are labelled as a, b, c, and d.

The motion derivation for sub-CU A starts by identifying its two spatial neighbours. The first neighbour is the N×N block above sub-CU A (block c). If this block c is not available or is intra coded the other N×N blocks above sub-CU A are checked (from left to right, starting at block c). The second neighbour is a block to the left of the sub-CU A (block b). If block b is not available or is intra coded other blocks to the left of sub-CU A are checked (from top to bottom, staring at block b). The motion information obtained from the neighbouring blocks for each list is scaled to the first reference frame for a given list. Next, temporal motion vector predictor (TMVP) of sub-block A is derived by following the same procedure of TMVP derivation as specified in HEVC. The motion information of the collocated block at location D is fetched and scaled accordingly. Finally, after retrieving and scaling the motion information, all available motion vectors (up to 3) are averaged separately for each reference list. The averaged motion vector is assigned as the motion vector of the current sub-CU.

FIG. 11 shows an example of one CU with four sub-blocks (A-D) and its neighbouring blocks (a-d).

2.2.1.3 Sub-CU Motion Prediction Mode Signalling

The sub-CU modes are enabled as additional merge candidates and there is no additional syntax element required to signal the modes. Two additional merge candidates are added to merge candidates list of each CU to represent the ATMVP mode and STMVP mode. Up to seven merge candidates are used, if the sequence parameter set indicates that ATMVP and STMVP are enabled. The encoding logic of the additional merge candidates is the same as for the merge candidates in the HM, which means, for each CU in P or B slice, two more RD checks is needed for the two additional merge candidates.

In the JEM, all bins of merge index is context coded by CABAC. While in HEVC, only the first bin is context coded and the remaining bins are context by-pass coded.

2.2.2 Adaptive Motion Vector Difference Resolution

In HEVC, motion vector differences (MVDs) (between the motion vector and predicted motion vector of a PU) are signalled in units of quarter luma samples when use_integer_mv_flag is equal to 0 in the slice header. In the JEM, a locally adaptive motion vector resolution (LAMVR) is introduced. In the JEM, MVD can be coded in units of quarter luma samples, integer luma samples or four luma samples. The MVD resolution is controlled at the coding unit (CU) level, and MVD resolution flags are conditionally signalled for each CU that has at least one non-zero MVD components.

For a CU that has at least one non-zero MVD components, a first flag is signalled to indicate whether quarter luma sample MV precision is used in the CU. When the first flag (equal to 1) indicates that quarter luma sample MV precision is not used, another flag is signalled to indicate whether integer luma sample MV precision or four luma sample MV precision is used.

When the first MVD resolution flag of a CU is zero, or not coded for a CU (meaning all MVDs in the CU are zero), the quarter luma sample MV resolution is used for the CU. When a CU uses integer-luma sample MV precision or four-luma-sample MV precision, the MVPs in the AMVP candidate list for the CU are rounded to the corresponding precision.

In the encoder, CU-level RD checks are used to determine which MVD resolution is to be used for a CU. That is, the CU-level RD check is performed three times for each MVD resolution. To accelerate encoder speed, the following encoding schemes are applied in the JEM.

• • During RD check of a CU with normal quarter luma sample MVD resolution, the motion information of the current CU (integer luma sample accuracy) is stored. The stored motion information (after rounding) is used as the starting point for further small range motion vector refinement during the RD check for the same CU with integer luma sample and 4 luma sample MVD resolution so that the time-consuming motion estimation process is not duplicated three times.
  • RD check of a CU with 4 luma sample MVD resolution is conditionally invoked. For a CU, when RD cost integer luma sample MVD resolution is much larger than that of quarter luma sample MVD resolution, the RD check of 4 luma sample MVD resolution for the CU is skipped.

2.2.3 Higher motion vector storage accuracy

In HEVC, motion vector accuracy is one-quarter pel (one-quarter luma sample and one-eighth chroma sample for 4:2:0 video). In the JEM, the accuracy for the internal motion vector storage and the merge candidate increases to 1/16 pel. The higher motion vector accuracy ( 1/16 pel) is used in motion compensation inter prediction for the CU coded with skip/merge mode. For the CU coded with normal AMVP mode, either the integer-pel or quarter-pel motion is used.

SHVC upsampling interpolation filters, which have same filter length and normalization factor as HEVC motion compensation interpolation filters, are used as motion compensation interpolation filters for the additional fractional pel positions. The chroma component motion vector accuracy is 1/32 sample in the JEM, the additional interpolation filters of 1/32 pel fractional positions are derived by using the average of the filters of the two neighbouring 1/16 pel fractional positions.

2.2.4 Overlapped Block Motion Compensation

Overlapped Block Motion Compensation (OBMC) has previously been used in H.263. In the JEM, unlike in H.263, OBMC can be switched on and off using syntax at the CU level. When OBMC is used in the JEM, the OBMC is performed for all motion compensation (MC) block boundaries except the right and bottom boundaries of a CU. Moreover, it is applied for both the luma and chroma components. In the JEM, a MC block is corresponding to a coding block. When a CU is coded with sub-CU mode (includes sub-CU merge, affine and FRUC mode), each sub-block of the CU is a MC block. To process CU boundaries in a uniform fashion, OBMC is performed at sub-block level for all MC block boundaries, where sub-block size is set equal to 4×4, as illustrated in FIG. 12.

When OBMC applies to the current sub-block, besides current motion vectors, motion vectors of four connected neighbouring sub-blocks, if available and are not identical to the current motion vector, are also used to derive prediction block for the current sub-block. These multiple prediction blocks based on multiple motion vectors are combined to generate the final prediction signal of the current sub-block.

Prediction block based on motion vectors of a neighbouring sub-block is denoted as P_N, with N indicating an index for the neighbouring above, below, left and right sub-blocks and prediction block based on motion vectors of the current sub-block is denoted as P_C. When P_N is based on the motion information of a neighbouring sub-block that contains the same motion information to the current sub-block, the OBMC is not performed from P_N. Otherwise, every sample of P_N is added to the same sample in Pc, i.e., four rows/columns of P_N are added to P_C. The weighting factors {¼, ⅛, 1/16, 1/32} are used for P_N and the weighting factors {¾, ⅞, 15/16, 31/32} are used for P_C. The exception are small MC blocks, (i.e., when height or width of the coding block is equal to 4 or a CU is coded with sub-CU mode), for which only two rows/columns of P_N are added to P_C. In this case weighting factors {¼, ⅛} are used for P_N and weighting factors {¾, ⅞} are used for P_C. For P_N generated based on motion vectors of vertically (horizontally) neighbouring sub-block, samples in the same row (column) of P_N are added to P_C with a same weighting factor.

In the JEM, for a CU with size less than or equal to 256 luma samples, a CU level flag is signalled to indicate whether OBMC is applied or not for the current CU. For the CUs with size larger than 256 luma samples or not coded with AMVP mode, OBMC is applied by default. At the encoder, when OBMC is applied for a CU, its impact is taken into account during the motion estimation stage. The prediction signal formed by OBMC using motion information of the top neighbouring block and the left neighbouring block is used to compensate the top and left boundaries of the original signal of the current CU, and then the normal motion estimation process is applied.

2.2.5 Local Illumination Compensation

Local Illumination Compensation (LIC) is based on a linear model for illumination changes, using a scaling factor a and an offset b. And it is enabled or disabled adaptively for each inter-mode coded coding unit (CU).

When LIC applies for a CU, a least square error method is employed to derive the parameters a and b by using the neighbouring samples of the current CU and their corresponding reference samples. More specifically, as illustrated in FIG. 13, the subsampled (2:1 subsampling) neighbouring samples of the CU and the corresponding samples (identified by motion information of the current CU or sub-CU) in the reference picture are used. The IC parameters are derived and applied for each prediction direction separately.

When a CU is coded with merge mode, the LIC flag is copied from neighbouring blocks, in a way similar to motion information copy in merge mode; otherwise, an LIC flag is signalled for the CU to indicate whether LIC applies or not.

When LIC is enabled for a picture, additional CU level RD check is needed to determine whether LIC is applied or not for a CU. When LIC is enabled for a CU, mean-removed sum of absolute difference (MR-SAD) and mean-removed sum of absolute Hadamard-transformed difference (MR-SATD) are used, instead of SAD and SATD, for integer pel motion search and fractional pel motion search, respectively.

To reduce the encoding complexity, the following encoding scheme is applied in the JEM.

• • LIC is disabled for the entire picture when there is no obvious illumination change between a current picture and its reference pictures. To identify this situation, histograms of a current picture and every reference picture of the current picture are calculated at the encoder. If the histogram difference between the current picture and every reference picture of the current picture is smaller than a given threshold, LIC is disabled for the current picture; otherwise, LIC is enabled for the current picture.

2.2.6 Affine Motion Compensation Prediction

In HEVC, only translation motion model is applied for motion compensation prediction (MCP). While in the real world, there are many kinds of motion, e.g. zoom in/out, rotation, perspective motions and the other irregular motions. In the JEM, a simplified affine transform motion compensation prediction is applied. As shown FIG. 14, the affine motion field of the block is described by two control point motion vectors.

The motion vector field (MVF) of a block is described by the following equation:

$\begin{array}{cc}\{\begin{array}{c}{v}_x=\frac{\left(v_{1x}-v_{0x}\right)}{w}x-\frac{\left(v_{1y}-v_{0y}\right)}{w}y+v_{0x}\\ v_y=\frac{\left(v_{1y}-v_{0y}\right)}{w}x+\frac{\left(v_{1x}-v_{0x}\right)}{w}y+v_{0y}\end{array}\hspace{1em}& \left(1\right)\end{array}$

Where (v_0x, v_0y) is motion vector of the top-left corner control point, and (v_1x, v_1y) is motion vector of the top-right corner control point.

In order to further simplify the motion compensation prediction, sub-block based affine transform prediction is applied. The sub-block size M×N is derived as in Equation 2, where MvPre is the motion vector fraction accuracy ( 1/16 in JEM), (v_2x, v_2y) is motion vector of the bottom-left control point, calculated according to Equation 1.

$\begin{array}{cc}\{\begin{array}{c}M=\mathrm{clip}3\left(4,w,\frac{w\times \mathrm{MvPre}}{\max \left(\mathrm{abs}\left(v_{1x}-v_{0x}\right),\mathrm{abs}\left(v_{1y}-v_{0y}\right)\right)}\right)\\ N=\mathrm{clip}3\left(4,h,\frac{h\times \mathrm{MvPre}}{\max \left(\mathrm{abs}\left(v_{2x}-v_{0x}\right),\mathrm{abs}\left(v_{2y}-v_{0y}\right)\right)}\right)\end{array}& \left(2\right)\end{array}$

After derived by Equation 2, M and N should be adjusted downward if necessary to make it a divisor of w and h, respectively.

To derive motion vector of each M×N sub-block, the motion vector of the center sample of each sub-block, as shown in FIG. 15, is calculated according to Equation 1, and rounded to 1/16 fraction accuracy. Then the motion compensation interpolation filters mentioned in previous section [00111] are applied to generate the prediction of each sub-block with derived motion vector.

After MCP, the high accuracy motion vector of each sub-block is rounded and saved as the same accuracy as the normal motion vector.

In the JEM, there are two affine motion modes: AF_INTER mode and AF_MERGE mode. For CUs with both width and height larger than 8, AF_INTER mode can be applied. An affine flag in CU level is signalled in the bitstream to indicate whether AF_INTER mode is used. In this mode, a candidate list with motion vector pair {(v₀, v₁)|v₀={v_A, v_B, v_C}, v₁={v_D, v_E}} is constructed using the neighbour blocks. As shown in FIG. 16, v₀ is selected from the motion vectors of the block A, B or C. The motion vector from the neighbour block is scaled according to the reference list and the relationship among the POC of the reference for the neighbour block, the POC of the reference for the current CU and the POC of the current CU. And the approach to select v₁ from the neighbour block D and E is similar. If the number of candidate list is smaller than 2, the list is padded by the motion vector pair composed by duplicating each of the AMVP candidates. When the candidate list is larger than 2, the candidates are firstly sorted according to the consistency of the neighbouring motion vectors (similarity of the two motion vectors in a pair candidate) and only the first two candidates are kept. An RD cost check is used to determine which motion vector pair candidate is selected as the control point motion vector prediction (CPMVP) of the current CU. And an index indicating the position of the CPMVP in the candidate list is signalled in the bitstream. After the CPMVP of the current affine CU is determined, affine motion estimation is applied and the control point motion vector (CPMV) is found. Then the difference of the CPMV and the CPMVP is signalled in the bitstream.

When a CU is applied in AF_MERGE mode, it gets the first block coded with affine mode from the valid neighbour reconstructed blocks. And the selection order for the candidate block is from left, above, above right, left bottom to above left as shown in FIG. 17A. If the neighbour left bottom block A is coded in affine mode as shown in FIG. 17B, the motion vectors v₂, v₃ and v₄ of the top left corner, above right corner and left bottom corner of the CU which contains the block A are derived. And the motion vector v₀ of the top left corner on the current CU is calculated according to v₂, v₃ and v₄. Secondly, the motion vector v₁ of the above right of the current CU is calculated.

After the CPMV of the current CU v₀ and v₁ are derived, according to the simplified affine motion model Equation 1, the MVF of the current CU is generated. In order to identify whether the current CU is coded with AF_MERGE mode, an affine flag is signalled in the bitstream when there is at least one neighbour block is coded in affine mode.

2.2.7 Pattern Matched Motion Vector Derivation

Pattern matched motion vector derivation (PMMVD) mode is a special merge mode based on Frame-Rate Up Conversion (FRUC) techniques. With this mode, motion information of a block is not signalled but derived at decoder side.

A FRUC flag is signalled for a CU when its merge flag is true. When the FRUC flag is false, a merge index is signalled and the regular merge mode is used. When the FRUC flag is true, an additional FRUC mode flag is signalled to indicate which method (bilateral matching or template matching) is to be used to derive motion information for the block.

At encoder side, the decision on whether using FRUC merge mode for a CU is based on RD cost selection as done for normal merge candidate. That is the two matching modes (bilateral matching and template matching) are both checked for a CU by using RD cost selection. The one leading to the minimal cost is further compared to other CU modes. If a FRUC matching mode is the most efficient one, FRUC flag is set to true for the CU and the related matching mode is used.

Motion derivation process in FRUC merge mode has two steps. A CU-level motion search is first performed, then followed by a Sub-CU level motion refinement. At CU level, an initial motion vector is derived for the whole CU based on bilateral matching or template matching. First, a list of MV candidates is generated and the candidate which leads to the minimum matching cost is selected as the starting point for further CU level refinement. Then a local search based on bilateral matching or template matching around the starting point is performed and the MV results in the minimum matching cost is taken as the MV for the whole CU. Subsequently, the motion information is further refined at sub-CU level with the derived CU motion vectors as the starting points.

For example, the following derivation process is performed for a W×H CU motion information derivation. At the first stage, MV for the whole W×H CU is derived. At the second stage, the CU is further split into M×M sub-CUs. The value of M is calculated as in (3), D is a predefined splitting depth which is set to 3 by default in the JEM. Then the MV for each sub-CU is derived.

$\begin{array}{cc}M=\max \left\{4,\min \left\{\frac{M}{2^D},\frac{N}{2^D}\right\}\right\}& \left(3\right)\end{array}$

As shown in the FIG. 18, the bilateral matching is used to derive motion information of the current CU by finding the closest match between two blocks along the motion trajectory of the current CU in two different reference pictures. Under the assumption of continuous motion trajectory, the motion vectors MV0 and MV1 pointing to the two reference blocks shall be proportional to the temporal distances, i.e., TD0 and TD1, between the current picture and the two reference pictures. As a special case, when the current picture is temporally between the two reference pictures and the temporal distance from the current picture to the two reference pictures is the same, the bilateral matching becomes mirror based bi-directional MV.

As shown in FIG. 19, template matching is used to derive motion information of the current CU by finding the closest match between a template (top and/or left neighbouring blocks of the current CU) in the current picture and a block (same size to the template) in a reference picture. Except the aforementioned FRUC merge mode, the template matching is also applied to AMVP mode. In the JEM, as done in HEVC, AMVP has two candidates. With template matching method, a new candidate is derived. If the newly derived candidate by template matching is different to the first existing AMVP candidate, it is inserted at the very beginning of the AMVP candidate list and then the list size is set to two (meaning remove the second existing AMVP candidate). When applied to AMVP mode, only CU level search is applied.

2.2.7.1 CU Level MV Candidate Set

The MV candidate set at CU level consists of:

• • (i) Original AMVP candidates if the current CU is in AMVP mode
  • (ii) all merge candidates,
  • (iii) several MVs in the interpolated MV field, which is introduced in previous section[00155].
  • (iv) top and left neighbouring motion vectors

When using bilateral matching, each valid MV of a merge candidate is used as an input to generate a MV pair with the assumption of bilateral matching. For example, one valid MV of a merge candidate is (MVa, refa) at reference list A. Then the reference picture refb of its paired bilateral MV is found in the other reference list B so that refa and refb are temporally at different sides of the current picture. If such a refb is not available in reference list B, refb is determined as a reference which is different from refa and its temporal distance to the current picture is the minimal one in list B. After refb is determined, MVb is derived by scaling MVa based on the temporal distance between the current picture and refa, refb.

Four MVs from the interpolated MV field are also added to the CU level candidate list. More specifically, the interpolated MVs at the position (0, 0), (W/2, 0), (0, H/2) and (W/2, H/2) of the current CU are added.

When FRUC is applied in AMVP mode, the original AMVP candidates are also added to CU level MV candidate set.

At the CU level, up to 15 MVs for AMVP CUs and up to 13 MVs for merge CUs are added to the candidate list.

2.2.7.2 Sub-CU Level MV Candidate Set

The MV candidate set at sub-CU level consists of:

• • (i) an MV determined from a CU-level search,
  • (ii) top, left, top-left and top-right neighbouring MVs,
  • (iii) scaled versions of collocated MVs from reference pictures,
  • (iv) up to 4 ATMVP candidates,
  • (v) up to 4 STMVP candidates

The scaled MVs from reference pictures are derived as follows. All the reference pictures in both lists are traversed. The MVs at a collocated position of the sub-CU in a reference picture are scaled to the reference of the starting CU-level MV.

ATMVP and STMVP candidates are limited to the four first ones.

At the sub-CU level, up to 17 MVs are added to the candidate list.

2.2.7.3 Generation of Interpolated MV Field

Before coding a frame, interpolated motion field is generated for the whole picture based on unilateral ME. Then the motion field may be used later as CU level or sub-CU level MV candidates.

First, the motion field of each reference pictures in both reference lists is traversed at 4×4 block level. For each 4×4 block, if the motion associated to the block passing through a 4×4 block in the current picture (As shown in FIG. 20) and the block has not been assigned any interpolated motion, the motion of the reference block is scaled to the current picture according to the temporal distance TD0 and TD1 (the same way as that of MV scaling of TMVP in HEVC) and the scaled motion is assigned to the block in the current frame. If no scaled MV is assigned to a 4×4 block, the block's motion is marked as unavailable in the interpolated motion field.

2.2.7.4 Interpolation and Matching Cost

When a motion vector points to a fractional sample position, motion compensated interpolation is needed. To reduce complexity, bi-linear interpolation instead of regular 8-tap HEVC interpolation is used for both bilateral matching and template matching.

The calculation of matching cost is a bit different at different steps. When selecting the candidate from the candidate set at the CU level, the matching cost is the absolute sum difference (SAD) of bilateral matching or template matching. After the starting MV is determined, the matching cost C of bilateral matching at sub-CU level search is calculated as follows:

C=SAD+w·(|MV_x−MV_x^s|+|MV_y−MV_y^s|)

where w is a weighting factor which is empirically set to 4, MV and MV^s indicate the current MV and the starting MV, respectively. SAD is still used as the matching cost of template matching at sub-CU level search.

In FRUC mode, MV is derived by using luma samples only. The derived motion will be used for both luma and chroma for MC inter prediction. After MV is decided, final MC is performed using 8-taps interpolation filter for luma and 4-taps interpolation filter for chroma.

2.2.7.5 MV Refinement

MV refinement is a pattern based MV search with the criterion of bilateral matching cost or template matching cost. In the JEM, two search patterns are supported—an unrestricted center-biased diamond search (UCBDS) and an adaptive cross search for MV refinement at the CU level and sub-CU level, respectively. For both CU and sub-CU level MV refinement, the MV is directly searched at quarter luma sample MV accuracy, and this is followed by one-eighth luma sample MV refinement. The search range of MV refinement for the CU and sub-CU step are set equal to 8 luma samples.

2.2.7.6 Selection of Prediction Direction in Template Matching FRUC Merge Mode

In the bilateral matching merge mode, bi-prediction is always applied since the motion information of a CU is derived based on the closest match between two blocks along the motion trajectory of the current CU in two different reference pictures. There is no such limitation for the template matching merge mode. In the template matching merge mode, the encoder can choose among uni-prediction from list0, uni-prediction from list1 or bi-prediction for a CU. The selection is based on a template matching cost as follows:

• • If costBi<=factor*min(cost0, cost1)
    • bi-prediction is used;
  • Otherwise, if cost0<=cost1
    • uni-prediction from list0 is used;
  • Otherwise,
    • uni-prediction from list1 is used;

where cost0 is the SAD of list0 template matching, cost1 is the SAD of list1 template matching and costBi is the SAD of bi-prediction template matching. The value of factor is equal to 1.25, which means that the selection process is biased toward bi-prediction.

The inter prediction direction selection is only applied to the CU-level template matching process.

2.2.8 Generalized Bi-Prediction

In conventional bi-prediction, the predictors from L0 and L1 are averaged to generate the final predictor using the equal weight 0.5. The predictor generation formula is shown as in Equ. (4)

P_{TraditionalBiPred}=(P_L0+P_L1+RoundingOffset)>>shiftNum,  (4)

In Equ. (4), P_{TraditionalBiPred} is the final predictor for the conventional bi-prediction, P_L0 and P_L1 are predictors from L0 and L1, respectively, and RoundingOffset and shiftNum are used to normalize the final predictor.

Generalized Bi-prediction (GBI) in is proposed to allow applying different weights to predictors from L0 and L1. The predictor generation is shown in Equ. (5).

P_GBi=((1−w₁)*P_L0+w₁*P_L1+RoundingOffset_GBi)>>shiftNum_GBi,  (5)

In Equ. (5), P_GBi is the final predictor of GBi. (1−w₁) and w₁ are the selected GBI weights applied to the predictors of L0 and L1, respectively. RoundingOffset_GBi and shiftNum_GBi are used to normalize the final predictor in GBi.

The supported weights of w₁ is {−¼, ⅜, ½, ⅝, 5/4}. One equal-weight set and four unequal-weight sets are supported. For the equal-weight case, the process to generate the final predictor is exactly the same as that in the conventional bi-prediction mode. For the true bi-prediction cases in random access (RA) condition, the number of candidate weight sets is reduced to three.

For advanced motion vector prediction (AMVP) mode, the weight selection in GBI is explicitly signaled at CU-level if this CU is coded by bi-prediction. For merge mode, the weight selection is inherited from the merge candidate. In this proposal, GBI supports DMVR to generate the weighted average of template as well as the final predictor for BMS-1.0.

2.2.9 Multi-Hypothesis Inter Prediction

In the multi-hypothesis inter prediction mode, one or more additional prediction signals are signaled, in addition to the conventional uni/bi prediction signal. The resulting overall prediction signal is obtained by sample-wise weighted superposition. With the uni/bi prediction signal p_uni/bi and the first additional inter prediction signal/hypothesis h₃, the resulting prediction signal p₃ is obtained as follows:

p₃=(1−α)p_uni/bi+αh₃

The changes to the prediction unit syntax structure are shown below:

prediction_unit( x0, y0, nPbW, nPbH ) { Descriptor
...
if( ! cu_skip_flag[ x0 ][ y0 ] ) {
i = 0
readMore = 1
while( i < MaxNumAdditionalHypotheses && readMore ) {
additional_hypothesis_flag[ x0 ][ y0 ][ i ] ae(v)
if( additional_hypothesis_flag[ x0 ][ y0 ][ i ] ) {
ref_idx_add_hyp[ x0 ][ y0 ][ i ] ae(v)
mvd_coding( x0, y0, 2+i )
mvp_add_hyp_flag[ x0 ][ y0 ][ i ] ae(v)
add_hyp_weight_idx[ x0 ][ y0 ][ i ] ae(v)
}
readMore = additional_hypothesis_flag[ x0 ][ y0 ][ i ]
i++
}
}
}

The weighting factor α is specified by the syntax element add_hyp_weight_idx, according to the following mapping:

add_hyp_weight_idx α
0                  1/4
1                  −1/8

Note that for the additional prediction signals, the concept of prediction list0/list1 is abolished, and instead one combined list is used. This combined list is generated by alternatingly inserting reference frames from list0 and list1 with increasing reference index, omitting reference frames which have already been inserted, such that double entries are avoided.

Analogously to above, more than one additional prediction signals can be used. The resulting overall prediction signal is accumulated iteratively with each additional prediction signal.

p_n+1=(1−α_n+1)p_n+α_n+1h_n+1

The resulting overall prediction signal is obtained as the last p_n (i.e., the p_n having the largest index n).

Note that also for inter prediction blocks using MERGE mode (but not SKIP mode), additional inter prediction signals can be specified. Further note, that in case of MERGE, not only the uni/bi prediction parameters, but also the additional prediction parameters of the selected merging candidate can be used for the current block.

Multi-hypothesis intra and inter prediction mode is also known as Combined Inter and Intra Prediction (CIIP) mode.

2.2.10 Multi-Hypothesis Prediction for Uni-Prediction of AMVP Mode

In some example, when the multi-hypothesis prediction is applied to improve uni-prediction of AMVP mode, one flag is signaled to enable or disable multi-hypothesis prediction for inter_dir equal to 1 or 2, where 1, 2, and 3 represent list 0, list 1, and bi-prediction, respectively. Moreover, one more merge index is signaled when the flag is true. In this way, multi-hypothesis prediction turns uni-prediction into bi-prediction, where one motion is acquired using the original syntax elements in AMVP mode while the other is acquired using the merge scheme. The final prediction uses 1:1 weights to combine these two predictions as in bi-prediction. The merge candidate list is first derived from merge mode with sub-CU candidates (e.g., affine, alternative temporal motion vector prediction (ATMVP)) excluded. Next, it is separated into two individual lists, one for list 0 (L0) containing all L0 motions from the candidates, and the other for list 1 (L1) containing all L1 motions. After removing redundancy and filling vacancy, two merge lists are generated for L0 and L1 respectively. There are two constraints when applying multi-hypothesis prediction for improving AMVP mode. First, it is enabled for those CUs with the luma coding block (CB) area larger than or equal to 64. Second, it is only applied to L1 when in low delay B pictures.

2.2.11 Multi-Hypothesis Prediction for Skip/Merge Mode

In examples, when the multi-hypothesis prediction is applied to skip or merge mode, whether to enable multi-hypothesis prediction is explicitly signaled. An extra merge indexed prediction is selected in addition to the original one. Therefore, each candidate of multi-hypothesis prediction implies a pair of merge candidates, containing one for the 1^st merge indexed prediction and the other for the 2^nd merge indexed prediction. However, in each pair, the merge candidate for the 2^nd merge indexed prediction is implicitly derived as the succeeding merge candidate (i.e., the already signaled merge index plus one) without signaling any additional merge index. After removing redundancy by excluding those pairs, containing similar merge candidates and filling vacancy, the candidate list for multi-hypothesis prediction is formed. Then, motions from a pair of two merge candidates are acquired to generate the final prediction, where 5:3 weights are applied to the 1^st and 2^nd merge indexed predictions, respectively. Moreover, a merge or skip CU with multi-hypothesis prediction enabled can save the motion information of the additional hypotheses for reference of the following neighboring CUs in addition to the motion information of the existing hypotheses. Note that sub-CU candidates (e.g., affine, ATMVP) are excluded from the candidate list, and for low delay B pictures, multi-hypothesis prediction is not applied to skip mode. Moreover, when multi-hypothesis prediction is applied to merge or skip mode, for those CUs with CU width or CU height less than 16, or those CUs with both CU width and CU height equal to 16, bi-linear interpolation filter is used in motion compensation for multiple hypotheses. Therefore, the worst-case bandwidth (required access samples per sample) for each merge or skip CU with multi-hypothesis prediction enabled is calculated in Table 1 and each number is less than half of the worst-case bandwidth for each 4×4 CU with multi-hypothesis prediction disabled.

2.2.12 Ultimate Motion Vector Expression

In examples, ultimate motion vector expression (UMVE) is presented. UMVE is used for either skip or merge modes with a proposed motion vector expression method. Merge mode with Motion Vector Difference (MMVD) mode is also known as Ultimate Motion Vector Expression (UMVE) mode.

UMVE re-uses merge candidate as same as using in VVC. Among the merge candidates, a candidate can be selected, and is further expanded by the proposed motion vector expression method.

UMVE provides a new motion vector expression with simplified signaling. The expression method includes starting point, motion magnitude, and motion direction.

FIG. 21 shows an example of a UMVE Search Process

FIG. 22 shows an example of UMVE Search Points.

This proposed technique uses a merge candidate list as it is. But only candidates which are default merge type (MRG_TYPE_DEFAULT_N) are considered for UMVE's expansion.

Base candidate index defines the starting point. Base candidate index indicates the best candidate among candidates in the list as follows.

TABLE 1
Base candidate IDX
	Base candidate
	IDX	0	1	2	3
	Nth MVP	1st MVP	2nd MVP	3rd MVP	4th MVP

If the number of base candidate is equal to 1, Base candidate IDX is not signaled.

Distance index is motion magnitude information. Distance index indicates the pre-defined distance from the starting point information. Pre-defined distance is as follows:

TABLE 2
Distance IDX
Distance IDX	0	1	2	3	4	5	6	7
Pixel	1/4-	1/2-	1-pel	2-pel	4-pel	8-pel	16-pel	32-pel
distance	pel	pel

Direction index represents the direction of the MVD relative to the starting point. The direction index can represent of the four directions as shown below.

TABLE 3
Direction IDX
Direction IDX	00	01	10	11
x-axis	+	−	N/A	N/A
y-axis	N/A	N/A	+	−

UMVE flag is signaled right after sending a skip flag and merge flag. If skip and merge flag is true, UMVE flag is parsed. If UMVE flag is equal to 1, UMVE syntaxes are parsed. But, if not 1, AFFINE flag is parsed. If AFFINE flag is equal to 1, that is AFFINE mode, But, if not 1, skip/merge index is parsed for VTM's skip/merge mode.

Additional line buffer due to UMVE candidates is not needed. Because a skip/merge candidate of software is directly used as a base candidate. Using input UMVE index, the supplement of MV is decided right before motion compensation. There is no need to hold long line buffer for this.

2.2.13 Affine Merge Mode with Prediction Offsets

In examples, UMVE is extended to affine merge mode, we will call this UMVE affine mode thereafter. The proposed method selects the first available affine merge candidate as a base predictor. Then it applies a motion vector offset to each control point's motion vector value from the base predictor. If there's no affine merge candidate available, this proposed method will not be used.

The selected base predictor's inter prediction direction, and the reference index of each direction is used without change.

In the current implementation, the current block's affine model is assumed to be a 4-parameter model, only 2 control points need to be derived. Thus, only the first 2 control points of the base predictor will be used as control point predictors.

For each control point, a zero_MVD flag is used to indicate whether the control point of current block has the same MV value as the corresponding control point predictor. If zero_MVD flag is true, there's no other signaling needed for the control point. Otherwise, a distance index and an offset direction index is signaled for the control point.

A distance offset table with size of 5 is used as shown in the table below. Distance index is signaled to indicate which distance offset to use. The mapping of distance index and distance offset values is shown in FIG. 23.

TABLE 1
Distance offset table
	Distance IDX	0	1	2	3	4
	Distance-offset	1/2-pel	1-pel	2-pel	4-pel	8-pel

The direction index can represent four directions as shown below, where only x or y direction may have an MV difference, but not in both directions.

Offset Direction IDX	00	01	10	11
x-dir-factor	+1	−1	0	0
y-dir-factor	0	0	+1	−1

If the inter prediction is uni-directional, the signaled distance offset is applied on the offset direction for each control point predictor. Results will be the MV value of each control point.

For example, when base predictor is uni-directional, and the motion vector values of a control point is MVP (v_px, v_py). When distance offset and direction index are signaled, the motion vectors of current block's corresponding control points will be calculated as below.

MV(v_x,v_y)=MVP(v_px,v_py)+MV(x-dir-factor*distance-offset,y-dir-factor*distance-offset);

If the inter prediction is bi-directional, the signaled distance offset is applied on the signaled offset direction for control point predictor's L0 motion vector; and the same distance offset with opposite direction is applied for control point predictor's L1 motion vector. Results will be the MV values of each control point, on each inter prediction direction.

For example, when base predictor is uni-directional, and the motion vector values of a control point on L0 is MVP_L0 (v_0px, v_0py), and the motion vector of that control point on L1 is MVP_L1 (v_1px, v_1py). When distance offset and direction index are signaled, the motion vectors of current block's corresponding control points will be calculated as below.

MV_L0(v_0x,v_0y)=MVP_L0(v_0px,v_0py)+MV(x-dir-factor*distance-offset,y-dir-factor*distance-offset);

MV_L1(v_0x,v_0y)=MVP_L1(v_1px,v_1py)+MV(−x-dir-factor*distance-offset,−y-dir-factor*distance-offset);

2.2.14 Bi-Directional Optical Flow

In BIO, motion compensation is first performed to generate the first predictions (in each prediction direction) of the current block. The first predictions are used to derive the spatial gradient, the temporal gradient and the optical flow of each subblock/pixel within the block, which are then used to generate the second prediction, i.e., the final prediction of the subblock/pixel. The details are described as follows.

Bi-directional Optical flow (BIO) is sample-wise motion refinement which is performed on top of block-wise motion compensation for bi-prediction. The sample-level motion refinement doesn't use signalling.

FIG. 24 shows an example of an optical flow trajectory

Let I^(k) be the luma value from reference k (k=0, 1) after block motion compensation, and ∂I^(k)/∂x, ∂I^(k)/∂y are horizontal and vertical components of the I^(k) gradient, respectively. Assuming the optical flow is valid, the motion vector field (v^x, v_y) is given by an equation as follows:

∂I^(k)/∂t+v_x∂I^(k)/∂x+v_y∂I^(k)/∂y=0.

Combining this optical flow equation with Hermite interpolation for the motion trajectory of each sample results in a unique third-order polynomial that matches both the function values I^(k) and derivatives ∂I^(k)/∂x, ∂I^(k)/∂y at the ends. The value of this polynomial at t=0 is the BIO prediction:

pred_BIO=1/2·(I⁽⁰⁾+I⁽¹⁾+v_x/2·(τ₁∂I⁽¹⁾/∂x−τ₀∂I⁽⁰⁾/∂x)+v_y/2·(τ₁∂I⁽¹⁾/∂y−τ₀∂I⁽⁰⁾/∂y)).  (4)

Here, τ₀ and τ₁ denote the distances to the reference frames as shown on a FIG. 24. Distances τ₀ and τ₁ are calculated based on POC for Ref0 and Ref1: τ₀=POC(current)−POC(Ref0), τ₁=POC(Ref1)−POC(current). If both predictions come from the same time direction (either both from the past or both from the future) then the signs are different (i.e., τ₀·τ₁<0). In this case, BIO is applied only if the prediction is not from the same time moment (i.e., τ₀≠τ₁), both referenced regions have non-zero motion (MVx₀, MVy₀, MVx₁, MVy₁≠0) and the block motion vectors are proportional to the time distance (MVx₀/MVx₁=MVy₀/MVy₁=−τ₀/τ₁).

The motion vector field (v_x, v_y) is determined by minimizing the difference Δ between values in points A and B (intersection of motion trajectory and reference frame planes on FIG. 24). Model uses only first linear term of a local Taylor expansion for Δ:

Δ=(I⁽⁰⁾−I⁽¹⁾₀+v_x(τ₁∂I⁽¹⁾/∂x+τ₀∂I⁽⁰⁾/∂x)+v_y(τ₁∂I⁽¹⁾/∂y+τ₀∂I⁽⁰⁾/∂y))  (5)

All values in Equation 5 depend on the sample location (i′, j′), which was omitted from the notation so far. Assuming the motion is consistent in the local surrounding area, we minimize Δ inside the (2M+1)×(2M+1) square window Ω centered on the currently predicted point (i, j), where M is equal to 2:

$\begin{array}{cc}\left(v_x,v_y\right)=\underset{v_x,v_y}{\arg \min }\sum _{\left[i^{\prime },j\right]\in \Omega }{\Delta }^2\left[i^{\prime },j^{\prime }\right]& \left(6\right)\end{array}$

For this optimization problem, the JEM uses a simplified approach making first a minimization in the vertical direction and then in the horizontal direction. This results in

$\begin{array}{cc}{v}_x=\left(s_1+r\right)>m?\mathrm{clip}3\left(-\mathrm{thBIO},\mathrm{thBIO},-\frac{s_3}{\left(s_1+r\right)}\right)\text{:}0& \left(7\right)\\ v_y=\left(s_5+r\right)>m?\mathrm{clip}3\left(-\mathrm{thBIO},\mathrm{thBIO},-\frac{s_6-v_x{s}_2/2}{\left(s_5+r\right)}\right)\text{:}0\mathrm{where},& \left(8\right)\\ s_1=\sum _{\left[i^{\prime },j\right]\in \Omega }{\left({\tau }_1\partial I^{\left(1\right)}/{\partial }_x+{\tau }_0\partial I^{\left(0\right)}/{\partial }_x\right)}^2;& \left(9\right)\\ s_3=\sum _{\left[i^{\prime },j\right]\in \Omega }\left(I^{\left(1\right)}-I^{\left(0\right)}\right)\left({\tau }_1\partial I^{\left(1\right)}/{\partial }_x+{\tau }_0\partial I^{\left(0\right)}/{\partial }_x\right);& \\ s_2=\sum _{\left[i^{\prime },j\right]\in \Omega }\left({\tau }_1\partial I^{\left(1\right)}/{\partial }_x+{\tau }_0\partial I^{\left(0\right)}/{\partial }_x\right)\left({\tau }_1\partial I^{\left(1\right)}/\partial y+{\tau }_0\partial I^{\left(0\right)}/\partial y\right);& \\ s_5=\sum _{\left[i^{\prime },j\right]\in \Omega }{\left({\tau }_1\partial I^{\left(1\right)}/\partial y+{\tau }_0\partial I^{\left(0\right)}/\partial y\right)}^2;& \\ s_6=\sum _{\left[i^{\prime },j\right]\in \Omega }\left(I^{\left(1\right)}-I^{\left(0\right)}\right)\left({\tau }_1\partial I^{\left(1\right)}/\partial y+{\tau }_0\partial I^{\left(0\right)}/\partial y\right)& \end{array}$

In order to avoid division by zero or a very small value, regularization parameters r and m are introduced in Equations 10 and 11.

r=500·4^d−8  (10)

m=700·4^d−8  (11)

Here d is bit depth of the video samples.

In order to keep the memory access for BIO the same as for regular bi-predictive motion compensation, all prediction and gradients values, I^(k), ∂I^(k)/∂x, ∂I^(k)/∂y, are calculated only for positions inside the current block. In Equation 9, (2M+1)×(2M+1) square window Ω centered in currently predicted point on a boundary of predicted block needs to accesses positions outside of the block (as shown in FIG. 25A). In the JEM, values of I^(k), ∂I^(k)/∂x, ∂I^(k)/∂y outside of the block are set to be equal to the nearest available value inside the block. For example, this can be implemented as padding, as shown in FIG. 25B.

FIG. 26 shows BIO w/o block extension: a) access positions outside of the block; b) padding is used in order to avoid extra memory access and calculation.

With BIO, it's possible that the motion field can be refined for each sample. To reduce the computational complexity, a block-based design of BIO is used in the JEM. The motion refinement is calculated based on 4×4 block. In the block-based BIO, the values of s_n in Equation 9 of all samples in a 4×4 block are aggregated, and then the aggregated values of s_n in are used to derived BIO motion vectors offset for the 4×4 block. More specifically, the following formula is used for block-based BIO derivation:

$\begin{array}{cc}{s}_{1,b_k}=\sum _{\left(x,y\right)\in b_k}\sum _{\left[i^{\prime },j\right]\in \Omega \left(x,y\right)}{\left({\tau }_1\partial I^{\left(1\right)}/\partial x+{\tau }_0\partial I^{\left(0\right)}/{\partial }_x\right)}^2;s_{3,b_k}=\sum _{\left(x,y\right)\in b_k}\sum _{\left[i^{\prime },j\right]\in \Omega }\left(I^{\left(1\right)}-I^{\left(0\right)}\right)\left({\tau }_1\partial I^{\left(1\right)}/\partial x+{\tau }_0\partial I^{\left(0\right)}/{\partial }_X\right\};s_{2,b_k}=\sum _{\left(x,y\right)\in b_k}\sum _{\left[i^{\prime },j\right]\in \Omega }\left({\tau }_1\partial I^{\left(1\right)}/\partial x+{\tau }_0\partial I^{\left(0\right)}/\partial x\right)\left({\tau }_1\partial I^{\left(1\right)}/\partial y+{\tau }_0\partial I^{\left(0\right)}/\partial y\right);s_{5,b_k}=\sum _{\left(x,y\right)\in b_k}\sum _{\left[i^{\prime },j\right]\in \Omega }{\left({\tau }_1\partial I^{\left(1\right)}/\partial y+{\tau }_0\partial I^{\left(0\right)}/\partial y\right)}^2;s_{6,b_k}=\sum _{\left(x,y\right)\in b_k}\sum _{\left[i^{\prime },j\right]\in \Omega }\left(I^{\left(1\right)}-I^{\left(0\right)}\right)\left({\tau }_1\partial I^{\left(1\right)}/\partial y+{\tau }_0\partial I^{\left(0\right)}/\partial y\right)& \left(12\right)\end{array}$

where b_k denotes the set of samples belonging to the k-th 4×4 block of the predicted block. s_n in Equations 7 and 8 are replaced by ((s_n,bk)>>4) to derive the associated motion vector offsets.

In some cases, MV regiment of BIO might be unreliable due to noise or irregular motion. Therefore, in BIO, the magnitude of MV regiment is clipped to a threshold value thBIO. The threshold value is determined based on whether the reference pictures of the current picture are all from one direction. If all the reference pictures of the current picture are from one direction, the value of the threshold is set to 12×2^14−d; otherwise, it is set to 12×2^13−d.

Gradients for BIO are calculated at the same time with motion compensation interpolation using operations consistent with HEVC motion compensation process (2D separable FIR). The input for this 2D separable FIR is the same reference frame sample as for motion compensation process and fractional position (fracX, fracY) according to the fractional part of block motion vector. In case of horizontal gradient ∂I/∂x signal first interpolated vertically using BIOfilterS corresponding to the fractional position fracY with de-scaling shift d−8, then gradient filter BIOfilterG is applied in horizontal direction corresponding to the fractional position fracX with de-scaling shift by 18−d. In case of vertical gradient ∂I/∂y first gradient filter is applied vertically using BIOfilterG corresponding to the fractional position fracY with de-scaling shift d−8, then signal displacement is performed using BIOfilterS in horizontal direction corresponding to the fractional position fracX with de-scaling shift by 18−d. The length of interpolation filter for gradients calculation BIOfilterG and signal displacement BIOfilterF is shorter (6-tap) in order to maintain reasonable complexity. Table 2 shows the filters used for gradients calculation for different fractional positions of block motion vector in BIO. Table 3 shows the interpolation filters used for prediction signal generation in BIO.

TABLE 2
Filters for gradients calculation in BIO
Fractional pel position Interpolation filter for gradient(BIOfilterG)
0                       {8, −39, −3, 46, −17, 5}
1/16                    {8, −32, −13, 50, −18, 5}
1/8                     {7, −27, −20, 54, −19, 5}
3/16                    {6, −21, −29, 57, −18, 5}
1/4                     {4, −17, −36, 60, −15, 4}
5/16                    {3, −9, −44, 61, −15, 4}
3/8                     {1, −4, −48, 61, −13, 3}
7/16                    {0, 1, −54, 60, −9, 2}
1/2                     {−1, 4, −57, 57, −4, 1}

TABLE 3
Interpolation filters for prediction signal generation in BIO
Fractional pel position Interpolation filter for prediction signal(BIOfilterS)
0                       {0, 0, 64, 0, 0, 0}
1/16                    {1, −3, 64, 4, −2, 0}
1/8                     {1, −6, 62, 9, −3, 1}
3/16                    {2, −8, 60, 14, −5, 1}
1/4                     {2, −9, 57, 19, −7, 2}
5/16                    {3, −10, 53, 24, −8, 2}
3/8                     {3, −11, 50, 29, −9, 2}
7/16                    {3, −11, 44, 35, −10, 3}
1/2                     {3, −10, 35, 44, −11, 3}

In the JEM, BIO is applied to all bi-predicted blocks when the two predictions are from different reference pictures. When LIC is enabled for a CU, BIO is disabled.

In the JEM, OBMC is applied for a block after normal MC process. To reduce the computational complexity, BIO is not applied during the OBMC process. This means that BIO is only applied in the MC process for a block when using its own MV and is not applied in the MC process when the MV of a neighboring block is used during the OBMC process.

It is proposed that before calculating the temporal gradient in BIO, a reference block (or a prediction block) may be modified firstly, and the calculation of temporal gradient is based on the modified reference block. In one example, mean is removed for all reference blocks. In one example, mean is defined as the average of selected samples in the reference block. In one example, all pixels in a reference block X or a sub-block of the reference block X are used to calculate MeanX. In one example, only partial pixels in a reference block X or a sub-block of the reference block are used to calculate MeanX. For example, only pixels in every second row/column are used.

2.2.15 Decoder-Side Motion Vector Refinement

In bi-prediction operation, for the prediction of one block region, two prediction blocks, formed using a motion vector (MV) of list0 and a MV of list1, respectively, are combined to form a single prediction signal. In the decoder-side motion vector refinement (DMVR) method, the two motion vectors of the bi-prediction are further refined by a bilateral template matching process. The bilateral template matching applied in the decoder to perform a distortion-based search between a bilateral template and the reconstruction samples in the reference pictures in order to obtain a refined MV without transmission of additional motion information.

In DMVR, a bilateral template is generated as the weighted combination (i.e. average) of the two prediction blocks, from the initial MV0 of list0 and MV1 of list1, respectively, as shown in FIG. 26. The template matching operation consists of calculating cost measures between the generated template and the sample region (around the initial prediction block) in the reference picture. For each of the two reference pictures, the MV that yields the minimum template cost is considered as the updated MV of that list to replace the original one. In the JEM, nine MV candidates are searched for each list. The nine MV candidates include the original MV and 8 surrounding MVs with one luma sample offset to the original MV in either the horizontal or vertical direction, or both. Finally, the two new MVs, i.e., MV0′ and MV1′ as shown in FIG. 26, are used for generating the final bi-prediction results. A sum of absolute differences (SAD) is used as the cost measure. Please note that when calculating the cost of a prediction block generated by one surrounding MV, the rounded MV (to integer pel) is actually used to obtain the prediction block instead of the real MV.

DMVR is applied for the merge mode of bi-prediction with one MV from a reference picture in the past and another from a reference picture in the future, without the transmission of additional syntax elements. In the JEM, when LIC, affine motion, FRUC, or sub-CU merge candidate is enabled for a CU, DMVR is not applied.

FIG. 26 shows an example of a DMVR based on bilateral template matching

2.3 Related Method

For motion refinement and coding in video coding, a MV update method and a two-step inter prediction method is proposed. The derived MV between reference block 0 and reference block 1 in BIO are scaled and added to the original motion vector of list 0 and list 1. Meanwhile, the updated MV is used to perform motion compensation and a second inter prediction is generated as the final prediction.

Meanwhile, the temporal gradient is modified by removing the mean difference between reference block 0 and reference block 1.

In methods for sub-block based prediction in video coding, for methods for sub-block based prediction in video coding, for several sub-blocks with different, only one set of MVs is generate for chroma component.

3. Relationship to Other Technologies

Sub-block based prediction method is proposed. First, we propose to divide the current block into sub-blocks in different ways depending on the color component and the color format (such as 4:2:0 or 4:2:2). Second, we propose that the MV of a sub-block of one color component can be derived from the MV(s) of one or more sub-blocks of another color component, which has (have) already been derived. Third, we propose to unify the constrain for merge affine mode and non-merge affine mode.

For example, if an 8×8 CU is split into four 4×4 sub-blocks and each of the sub-block has its own motion vectors, then, we calculate average of the four motion vectors and use it (scaled by 2) as the motion vectors of the chroma component in YCbCr 4:2:0 case. In this way, motion compensation of the Cb/Cr component is performed for a 4×4 block instead of four 2×2 blocks and the memory bandwidth can be saved.

Interweaved prediction is proposed for sub-block motion compensation. With interweaved prediction, a block is divided into sub-blocks with more than one dividing patterns. A dividing pattern is defined as the way to divide a block into sub-blocks, including the size of sub-blocks and the position of sub-blocks. For each dividing pattern, a corresponding prediction block may be generated by deriving motion information of each sub-block based on the dividing pattern. Therefore, even for one prediction direction, multiple prediction blocks may be generated by multiple dividing patterns. Alternatively, for each prediction direction, only a dividing pattern may be applied.

Suppose there are X dividing patterns, and X prediction blocks of the current block, denoted as P₀, P₁, . . . , P_X−1 are generated by sub-block based prediction with the X dividing patterns. The final prediction of the current block, denoted as P, can be generated as:

$\begin{array}{cc}P\left(x,y\right)=\frac{{\sum }_{i=0}^{X-1}{w}_i\left(x,y\right)\times P_i\left(x,y\right)}{{\sum }_{i=0}^{X-1}{w}_i\left(x,y\right)}& \left(15\right)\end{array}$

where (x, y) is the coordinate of a pixel in the block and w_i(x, y) is the weighting value of P_i. Without losing generalization, it is supposed that Σ_i=0^X−1w_i(x, y)=(1<<N) wherein N is a non-negative value. FIG. 27 shows an example of interweaved prediction with two dividing patterns.

4. Examples of Problems Solved My Embodiments that Use the Present Techniques

A two-step inter prediction method is proposed, however, such method can be performed multiple times to get more accurate motion information such that higher coding gains may be expected.

In DMVR, there is no two-step inter prediction method.

5. Example Embodiments Presented in an Itemized Format

To solve the problems mentioned above, it is proposed to refine motion information (e.g., motion vectors) more than once for those coding tools which need to perform decoder-side motion refinement process (e.g., DMVR), and/or for those coding tools which relies on some intermediate motion information different from the final motion information used for motion compensation (e.g., BIO).

It is also proposed that even for coding tools which don't apply motion information refinement at the decoder side, motion information of a block/a sub-block within a coded block may be refined once or multiple times and the refined motion information may be used for motion vector prediction of blocks to be coded afterwards, and/or filtering process.

Hereinafter, DMVD is used to represent DMVR or BIO or other decoder-sider motion vector refinement method or pixel refinement method.

Denote SATD as sum of absolute transformed differences, MRSATD as mean removed sum of absolute transformed differences, and SSE as sum of squares error, and MRSSE as mean removed sum of squares error.

The detailed items below should be considered as examples to explain general concepts. These inventions should not be interpreted in a narrow way. Furthermore, these items described coding/decoding techniques that can be combined in any manner.

• • 1. It is proposed that the motion vector refinement process may be performed multiple times, e.g., performed N times wherein N is a non-negative integer number. Suppose the signaled MVs are (MVLX0_x, MVLX0_y) and the i-th refined MVs are (MVLXi_x, MVLXi_y), wherein LX=L0 or L1 and i=1, . . . , N.
    • a) In one example, the (i−1)th refined motion vectors (i.e., motion vectors after the (i−1)th iteration, and when (i−1) equals to 0, the signaled motion vectors are used) may be used to generate the i-th motion compensated reference blocks of the PU/CU/block/sub-block. Alternatively, furthermore, the i-th motion compensated reference blocks may be further used to derive the i-th temporal gradients, spatial gradients and refined motion vectors. An example is shown in FIG. 28.
    • b) Different interpolation filters from those used for inter-coded blocks which are not coded with proposed method may be used for motion compensation in different motion vector refinement steps to reduce the memory bandwidth. For example, short-tap filters are used in 1st˜(N−1)th steps.
    • c) Intermediate refined MVs from the i-th iteration may be firstly modified before being used to generate reference blocks. In one example, fractional MVs are rounded to integer MVs and then are used to generate reference blocks in some steps, for example, in 1st˜(N−1)th steps.
    • d) In some or all iterations, one block may be first split into several sub-blocks and each sub-block is treated in the same way as a normal coding block with size equal to the sub-block size.
      • i. In one example, a block is firstly split into multiple sub-blocks, and each block's motion information may be refined multiple times.
      • ii. Alternatively, a block is firstly split into multiple sub-blocks, and only for partial of the sub-blocks, their motion information may be refined multiple times.
      • iii. Alternatively, a block is firstly split into multiple sub-blocks, different sub-block's motion information may be refined with different numbers of iterations (e.g., for some sub-block, no refinement is applied, and for some, motion information may be refined with multiple times.)
      • iv. Alternatively, the motion information of the whole block is refined N−1 times and afterwards, based on the (N−1)th refined motion information, the block is split to multiple sub-blocks, and for each sub-block, its motion information may be further refined.
    • e) In different steps, the refined MVs may be derived at different sub-block size.
    • f) In one example, the refined motion vectors in the Nth step is used to perform the motion compensation and then the method described in previous section[00213] is used to generate the final prediction of the CU/sub-block.
    • g) In one example, predictions are generated for a block/sub-block in each step (or some steps) and they are weighted averaged to generate the final predictions of the block/sub-block.
    • h) In one example, MVs derived in each step may be further constrained.
      • i. For example, |MVLXi_x−MVLX0_x|<=Tx and |MVLXi_y−MVLX0_y|<=Ty, for all 1<=i<=N.
      • ii. For example, Max{MVLXi_x−MVLXj_x}<=Tx and Max{MVLXi_y−MVLXj_y}<=Ty, for all 1<=i, j<=N.
      • iii. The thresholds Tx and Ty can be equal or not. They can be predefined numbers or signaled from the encoder to the decoder in VPS/SPS/PPS/slice header/tile group header/tile/CTU/CU.
    • i) The motion vector refinement process may be terminated after the Kth step for a block/sub-block, if the refined MVs after the Kth step and the input MVs in the Kth step are similar.
      • i. For example, if the absolute difference between the vertical or/and horizontal component of the refined MV and the input MV (in any prediction direction) is not larger than T quarter-pel distance, wherein T=¼, ⅓, ½, 1, 2, 3, 4, . . . etc., the motion vector refinement process is terminated.
      • ii. For example, if sum of the absolute difference between the vertical and horizontal component of the refined MV and the input MV (in any prediction direction) is not larger than T quarter-pel distance, the motion vector refinement process is terminated.
    • j) The iterative number N may be adaptive
      • i. For example, N depends on the current block size.
        • 1. For example, N is larger for a larger block and vice versa.
      • ii. For example, N depends on the coding mode of the current block.
      • iii. For example, N depends on MVD (Motion Vector Difference) of the current block.
        • 1. For example, N is larger when |MVD| is larger.
      • iv. For example, N depends on QP
        • 1. For example, N is larger when QP is larger.
      • v. N may be signaled from the encoder to the decoder in VPS/SPS/PPS/picture header/slice header/tile group header/tile/CTU/CU.
  • 2. With the refined motion vectors in bullet 1, the two-step inter-prediction process may be extended to multiple-step inter-prediction, wherein the finally refined motion vectors after N iterations is used to perform the final motion compensation and generate the final prediction of a block/sub-block.
    • a) Alternatively, predictions are generated for a block/sub-block in each step (or some steps) and the final predictions of the block/sub-block may be generated by those predictions. In one example, they are weighted averaged to generate the final predictions of the block/sub-block.
  • 3. It is proposed that the temporal gradient modification process can be performed for each M1×N1 sub-block though the BIO process may be performed for each M2×N2 sub-block.
    • a) In one example, refined motion vectors are derived for each 4×4 block while the temporal gradient modification process is performed for each 8×8 block. That is, M1=N1=8, M2=N2=4.
    • b) In one example, refined motion vectors are derived for each 8×8 block while the temporal gradient modification process is performed for each 4×4 block.
    • c) In one example, refined motion vectors are derived for each 4×4 block while the temporal gradient modification process is performed for each 4×4 block.
    • d) In one example, refined motion vectors are derived for each 8×8 block while the temporal gradient modification process is performed for each 8×8 block.
    • e) M1, N1, M2, N2 may be pre-defined or depend on the block size/coded modes/signaled in VPS/SPS/PPS/picture header/tile groups/etc. al.
  • 4. It is proposed to only use partial pixels of a block/sub-block for calculating the temporal/spatial gradients, which may be used for deriving the motion vector of the block/sub-block.
    • a) In one example, temporal and spatial gradients are calculated for every N rows or/and columns. For example, N=2.
    • b) In one example, temporal and spatial gradients are calculated for the top-left/top-right/bottom-left/bottom-right quarter of the CU/sub-block.
    • c) In one example, temporal and spatial gradients are calculated for every N rows or/and columns the top-left/top-right/bottom-left/bottom-right quarter of the CU/sub-block.
    • d) Such methods may be enabled for the two-step inter-prediction or multiple-step inter-prediction in bullet 2, wherein temporal/spatial gradients may only be used to derive refined motion vectors of a block/sub-block, and is not directly used to refine prediction of the block/sub-block.
  • 5. It is proposed that the motion vector refinement process in DMVR may be performed multiple times.
    • a) In one example, the (i−1)th refined motion vectors (i.e., motion vectors after the (i−1)th iteration, and when (i−1) equals to 0, the signaled motion vectors are used) can be used as the start searching point in the i-th motion vector refinement process, i=1, . . . , N, wherein N is a non-negative integer number.
    • b) Different interpolation filters from those used for inter-coded blocks which are not coded with proposed method may be used in different motion vector refinement steps to reduce the memory bandwidth. For example, short-tap filters are used in 1st˜(N−1)th steps.
    • c) In one example, fractional MVs are rounded to integer MVs and are then used as the start searching point in some steps, for example, in 1st˜(N−1)th step.
  • 6. It is proposed that the refined motion vectors derived in BIO or DMVR or other decoder side motion refinement technologies may be only used for the final motion compensation of some components.
    • a) In one example, the refined motion vectors are only used for the final motion compensation of Cb or/and Cr component.
    • b) In one example, the refined motion vectors are only used for the final motion compensation of luma component.
    • c) In one example, in BIO, the refined motion vectors are used to perform motion compensation and generate the final prediction of chroma components, and the method described in [00185] is used to generate the final prediction of luma component.
      • i. For example, the motion vector is refined only once and is used for motion compensation of the chroma component, and the method described in previous section (section 2.2.14)[00185] is used to generate the final prediction of luma component.
    • d) In one example, in BIO, the method described in [00185](section 2.2.14) is used to generate the final prediction of both luma and chroma components.
    • e) In one example, in BIO and DMVR, the refined motion vectors are used to perform motion compensation and generate the final prediction of both luma and chroma components.
  • 7. Methods for sub-block based prediction in Video Coding can be used for motion compensation of chroma component to reduce memory bandwidth. For example, four neighboring 4×4 blocks are grouped together, and only one set of motion vector is derived for the chroma component (in YCbCr 4:2:0 case) and is used to perform motion compensation of a 4×4 chroma block.
  • 8. It is proposed that BIO or/and DMVR and/or or other decoder side motion refinement technologies may be performed at sub-block level.
    • a) Alternatively, furthermore, Interweaved Prediction in Video Coding can be used to derive different motion vectors for different dividing patterns, and the final prediction is generated based on the prediction value of all dividing patterns.
  • 9. The proposed methods may be applied under certain conditions, such as based on block sizes, encoded mode information, motion information, slice/picture/tile types, etc. al.
    • a) In one example, when a block size contains less than M*H samples, e.g., 16 or 32 or 64 luma samples, the above methods are not allowed.
    • b) In one example, when a block size contains more than M*H samples, e.g., 16 or 32 or 64 luma samples, the above methods are not allowed.
    • c) Alternatively, when minimum size of a block's width or/and height is smaller than or no larger than X, the above methods are not allowed. In one example, X is set to 8.
    • d) Alternatively, when a block's width >th1 or >=th1 and/or a block's height >th2 or >=th2, the above methods are not allowed. In one example, X is set to 64.
      • i. For example, the above methods are disabled for M×M (e.g., 128×128) block.
      • ii. For example, the above methods are disabled for N×M/M×N block, e.g., wherein N>=64, M=128.
      • iii. For example, the above methods are disabled for N×M/M×N block, e.g., wherein N>=4, M=128.
    • e) Alternatively, when a block's width <th1 or <=th1 and/or a block's height <th2 or <=th2, the above methods are not allowed. In one example, th1 and/or th2 is set to 8.
    • f) In one example, in BIO, the above methods are disabled for blocks coded in AMVP mode.
    • g) In one example, in BIO or DMVR, the above methods are disabled for blocks coded in skip mode.
  • 10. For sub-block based methods (e.g., Affine, ATMVP, BIO, DMVR, etc.), maximum number of sub-blocks may be fixed for all kinds of CU/PU sizes. Suppose there are will be K×L sub-blocks and one block size is denoted by M×N.
    • a) In one example, the width of a sub-block is set to max(TH_w, M/K).
    • b) In one example, the height of a sub-block is set to max(TH_h, N/L).
    • c) TH_w and/or TH_h may be pre-defined (e.g., 4) or signaled in SPS/PPS/picture/slice/tile group/tile level/group of CTUs/CTU row/CTU/CU/PU.
    • d) TH_w and/or TH_h may be dependent on whether current block is bi-prediction or uni-prediction. In one example, TH_w and/or TH_h may be set to 4 for uni-prediction or 8 for bi-prediction.
  • 11. For sub-block based methods (e.g., Affine, ATMVP, BIO, DMVR, etc. al), whether and how to split the block into sub-blocks may be different for different color components.
    • a) In one example, whether and how to split a chroma block depend on the width and height of the chroma block, independently of the whether and how to split its corresponding luma block.
    • b) In one example, whether and how to split a chroma block depend on the width and height of the chroma block, independently of the whether and how to split its corresponding luma block.
    • c) In one example, whether and how to split a chroma block depend on the whether and how to split its corresponding luma block.
  • 12. The above methods including proposed methods and BIO, DMVR or other decoder side motion refinement technologies, or sub-block based methods (e.g., affine, ATMVP etc.) may be applied in a sub-block level.
    • a) In one example, the iterative motion vector refinement for BIO and DMVR in bullet 1 and bullet 2 may be invoked for each sub-block.
    • b) In one example, when a block with either width or height or both width and height are both larger than (or equal to) a threshold L, the block may be split into multiple sub-blocks. Each sub-block is treated in the same way as a normal coding block with size equal to the sub-block size.
      • i. In one example, L is 64, a 64×128/128×64 block is split into two 64×64 sub-blocks, and a 128×128 block is split into four 64×64 sub-blocks. However, N×128/128×N block, wherein N<64, is not split into sub-blocks.
      • ii. In one example, L is 64, a 64×128/128×64 block is split into two 64×64 sub-blocks, and a 128×128 block is split into four 64×64 sub-blocks. Meanwhile, N×128/128×N block, wherein N<64, is split into two N×64/64×N sub-blocks.
      • iii. In one example, when width (or height) is larger than L, it is split vertically (or horizontally), and the width or/and height of the sub-block is no larger than L.
      • iv. In one example, L may be different for vertical direction and horizontal direction. For example, if width of block is larger than LW, the block may be split vertically; if height of a block is larger than LH, the block may be split horizontally.
      • v. In one example, LW may be width of the VPDU (virtual pipeline data unit) and LH may be height of the VPDU.
    • c) In one example, when size (i.e., width*height) of block is larger than a threshold L1, it may be split into multiple sub-blocks. Each sub-block is treated in the same way as a normal coding block with size equal to the sub-block size.
      • i. In one example, the block is split into sub-blocks with same size that is no larger than L1.
      • ii. In one example, if width (or height) of the block is no larger than a threshold L2, it is not split vertically (or horizontally).
      • iii. In one example, L1 is size of the VPDU.
      • iv. In one example, L1 is 1024, and L2 is 32. For example, a 16×128 block is split into two 16×64 sub-blocks.
      • v. In one example, L2=sqrt(L1).
      • vi. In one example, if block size (width and height denoted by W and H, respectively) is larger than L1, width (denoted by subW) and height (denoted by subH) of a sub-block is derived as follows:
        • If W>=L2 and H>=L2

subW=W/L2;

subH=H/L2;

• • • • • Else if W>L2 and H<L2

subH=H;

subW=W*H/L1;

• • • • • Else if W<L2 and H>L2

subW=W;

subH=W*H/L1;

• • • d) In one example, two-level splitting of one block may be applied wherein different rules may be applied to decide how to do splitting.
      • i. In one example, a block may be first split into sub-blocks using method in bullet 12.b, and these sub-blocks may be further split using method in bullet 12.c.
      • ii. In one example, a block may be first split into sub-blocks using method in bullet 12.c, and these sub-blocks may be further split using method in bullet 12.b.
    • e) The threshold L may be pre-defined or signaled in SPS/PPS/picture/slice/tile group/tile level.
    • f) Alternatively, the thresholds may depend on certain coded information, such as block size, picture type, temporal layer index, etc. al.
    • g) In one example, deblocking may be performed at boundary of these sub-blocks.
  • 13. It is proposed that DMVD and may be disabled in multi-hypothesis intra and inter prediction.
    • a) Alternatively, DMVD may be enabled in multi-hypothesis intra and inter prediction.
  • 14. It is proposed that DMVD may be disabled in MMVD (merge mode with MVD) or UMVE mode.
  • a) Alternatively, DMVR may be enabled in MMVD (merge mode with MVD) or UMVE mode.
  • 15. It is proposed that DMVD and may be disabled in triangle prediction.
    • a) Alternatively, DMVR may be enabled in triangle prediction.
  • 16. In one example, whether to and how to apply motion refinement methods such as DMVR or/and BIO and/or other decoder side motion refinement technologies depends on the reference picture.
    • a) In one example, motion refinement methods are not applied if the reference picture is the current coding picture;
    • b) In one example, multi-time motion refinement methods claimed in previous bullets are not applied if the reference picture is the current coding picture;
    • c) whether to and how to apply motion refinement methods such as DMVR or/and BIO and/or other decoder side motion refinement technologies depends on the positions of sub-blocks relative to the block covering the sub-block, and/or relative to the coding tree unit (CTU), and/or relative to the top-left position of the tile/picture.
  • 17. It is proposed that in the early termination stage of BIO or/and DMVR or other coding tools rely on difference calculation, the difference (e.g., SAD/SATD/SSE/MRSAD/MRSATD/MRSSE etc.) between the two reference blocks or/and sub-blocks may be calculated only for some representative positions.
    • a) In one example, only difference of even rows is calculated for the block or/and sub-block.
    • b) In one example, only difference of four corner samples of one block/sub-block is calculated for the block or/and sub-block.
    • c) In one example, method in improvements of decoder side motion vector derivation in video coding may be used to select the representative positions.
    • d) In one example, the difference (e.g., SAD/SATD/SSE/MRSAD/MRSATD/MRSSE etc.) between the two reference blocks may be calculated only for some representative sub-blocks.
    • e) In one example, the difference (e.g., SAD/SATD/SSE/MRSAD/MRSATD/MRSSE etc.) calculated for representative positions or sub-blocks are summed up to get the difference for the whole block/sub-block.
  • 18. In one example, the difference between the two reference blocks is calculated directly (instead of being calculated as sum of differences between the reference sub-blocks) and is used to decide whether BIO or/and DMVR or other coding tools rely on difference calculation is enabled or disabled for the entire block.
    • a) In one example, methods described in bullet 14 may be used to calculate the difference between the two reference blocks.
  • 19. Embodiment
    • This section presents an embodiment for how to split a block into sub-blocks in sub-block level DMVD.
    • Embodiment #1
      • a) Step 1: if a block is of size 128×128, it is split into 4 64×64 sub-blocks. If a block is of size N×128 or 128×N (N<128), it is split into 2 N×64 or 64×N sub-blocks. For other blocks, they are not split.
      • b) Step 2: for block that is not of size 128×128, or N×128 or 128×N (N<128), and for sub-block generated in step 1, if its size (i.e., width*height) is larger than 256, it is further split into sub-blocks of size 256 using method described in 12.c, with L1=256 and L2=16.
    • Embodiment #2
      • a) Step 1: if a block is of size 128×128, it is split into 4 64×64 sub-blocks. If a block is of size N×128 or 128×N, it is split into 2 N×64 or 64×N sub-blocks (N<128). For other blocks, they are not split.
      • b) Step 2: for block that is not of size 128×128, or N×128 or 128×N (N<128), and for sub-block generated in step 1, if its size (i.e., width*height) is larger than 1024, it is further split into sub-blocks of size 1024 using method described in 12.c, with L1=1024 and L2=32.
    • Embodiment #3
      • a) Step 1: if a block is of size 128×128, it is split into 4 64×64 sub-blocks. If a block is of size N×128 or 128×N, it is split into 2 N×64 or 64×N sub-blocks (N<128). For other blocks, they are not split.
    • Embodiment #4
      • a) Step 1: if a block is of size 256×256, it is split into 4 128×128 sub-blocks. If a block is of size N×256 or 256×N, it is split into 2 N×128 or 128×N sub-blocks (N<256). For other blocks, they are not split.
      • b) Step 2: for block that is not of size 256×256, or N×256 or 256×N (N<256), and for sub-block generated in step 1, if its size (i.e., width*height) is larger than 1024, it is further split into sub-blocks of size 1024 using method described in 12.c, with L1=1024 and L2=32.
    • Embodiment #5
      • a) Step 1: if width or height of a block is larger than 64, it is split into sub-blocks using method described in 12.b, with LW=LH=64.
      • b) Step 2: for block whose width and height are no larger than 64, and for sub-block generated in step 1, if its size (i.e., width*height) is larger than 1024, it is further split into sub-blocks of size 1024 using method described in 12.c, with L1=1024 and L2=32.

FIG. 29 is a block diagram illustrating an example of the architecture for a computer system or other control device 2600 that can be utilized to implement various portions of the presently disclosed technology. In FIG. 29, the computer system 2600 includes one or more processors 2605 and memory 2610 connected via an interconnect 2625. The interconnect 2625 may represent any one or more separate physical buses, point to point connections, or both, connected by appropriate bridges, adapters, or controllers. The interconnect 2625, therefore, may include, for example, a system bus, a Peripheral Component Interconnect (PCI) bus, a HyperTransport or industry standard architecture (ISA) bus, a small computer system interface (SCSI) bus, a universal serial bus (USB), IIC (I2C) bus, or an Institute of Electrical and Electronics Engineers (IEEE) standard 674 bus, sometimes referred to as “Firewire.”

The processor(s) 2605 may include central processing units (CPUs) to control the overall operation of, for example, the host computer. In certain embodiments, the processor(s) 2605 accomplish this by executing software or firmware stored in memory 2610. The processor(s) 2605 may be, or may include, one or more programmable general-purpose or special-purpose microprocessors, digital signal processors (DSPs), programmable controllers, application specific integrated circuits (ASICs), programmable logic devices (PLDs), or the like, or a combination of such devices.

The memory 2610 can be or include the main memory of the computer system. The memory 2610 represents any suitable form of random access memory (RAM), read-only memory (ROM), flash memory, or the like, or a combination of such devices. In use, the memory 2610 may contain, among other things, a set of machine instructions which, when executed by processor 2605, causes the processor 2605 to perform operations to implement embodiments of the presently disclosed technology.

Also connected to the processor(s) 2605 through the interconnect 2625 is a (optional) network adapter 2615. The network adapter 2615 provides the computer system 2600 with the ability to communicate with remote devices, such as the storage clients, and/or other storage servers, and may be, for example, an Ethernet adapter or Fiber Channel adapter.

FIG. 30 shows a block diagram of an example embodiment of a device 2700 that can be utilized to implement various portions of the presently disclosed technology. The mobile device 2700 can be a laptop, a smartphone, a tablet, a camcorder, or other types of devices that are capable of processing videos. The mobile device 2700 includes a processor or controller 2701 to process data, and memory 2702 in communication with the processor 2701 to store and/or buffer data. For example, the processor 2701 can include a central processing unit (CPU) or a microcontroller unit (MCU). In some implementations, the processor 2701 can include a field-programmable gate-array (FPGA). In some implementations, the mobile device 2700 includes or is in communication with a graphics processing unit (GPU), video processing unit (VPU) and/or wireless communications unit for various visual and/or communications data processing functions of the smartphone device. For example, the memory 2702 can include and store processor-executable code, which when executed by the processor 2701, configures the mobile device 2700 to perform various operations, e.g., such as receiving information, commands, and/or data, processing information and data, and transmitting or providing processed information/data to another device, such as an actuator or external display. To support various functions of the mobile device 2700, the memory 2702 can store information and data, such as instructions, software, values, images, and other data processed or referenced by the processor 2701. For example, various types of Random Access Memory (RAM) devices, Read Only Memory (ROM) devices, Flash Memory devices, and other suitable storage media can be used to implement storage functions of the memory 2702. In some implementations, the mobile device 2700 includes an input/output (I/O) unit 2703 to interface the processor 2701 and/or memory 2702 to other modules, units or devices. For example, the I/O unit 2703 can interface the processor 2701 and memory 2702 with to utilize various types of wireless interfaces compatible with typical data communication standards, e.g., such as between the one or more computers in the cloud and the user device. In some implementations, the mobile device 2700 can interface with other devices using a wired connection via the I/O unit 2703. The mobile device 2700 can also interface with other external interfaces, such as data storage, and/or visual or audio display devices 2704, to retrieve and transfer data and information that can be processed by the processor, stored in the memory, or exhibited on an output unit of a display device 2704 or an external device. For example, the display device 2704 can display a video frame modified based on the MVPs in accordance with the disclosed technology.

FIG. 31 is a flowchart for a method 3100 of video processing. The method 3100 includes generating (3102), using a multi-step refinement process, multiple refinement values of motion vector information based on decoded motion information from a bitstream representation of a current video block, and reconstructing (3104) the current video block or decoding other video blocks based on multiple refinement values.

Another video processing method includes performing, for conversion between a current block and a bitstream representation of the current block, a multi-step refinement process for a first sub-block of the current block and a temporal gradient modification process for a second sub-block of the current block, wherein, using the multi-step refinement process, multiple refinement values of motion vector information signaled in a bitstream representation of the current video block and performing the conversion between the current block and the bitstream representation using a selected one of the multiple refinement values.

Another video processing method includes determining, using a multi-step decoder-side motion vector refinement process a current video block, a final motion vector and performing conversion between the current block and the bitstream representation using the final motion vector.

Another video processing method includes applying, during conversion between a current video block and a bitstream representation of the current video block; multiple different motion vector refinement processes to different sub-blocks of the current video block and performing conversion between the current block and the bitstream representation using a final motion vector for the current video block generated from the multiple different motion vector refinement processes.

In the disclosed embodiments, the bitstream representation of a current block of video may include bits of a bitstream (compressed representation of a video) that may be non-contiguous and may depend on header information, as is known in the art of video compression. Furthermore, a current block may include samples representative of one or more of luma and chroma components, or rotational variations thereof (e.g, YCrCb or YUV, and so on).

The listing of clauses below describes some embodiments and techniques as follows.

1. A method of video processing, comprising: generating, using a multi-step refinement process, multiple refinement values of motion vector information based on decoded motion information from a bitstream representation of a current video block, and reconstructing the current video block or decoding other video blocks based on multiple refinement values. For example, the refinement operation may include an averaging operation. In yet another example, refined values of one step of refinement process are used to reconstruct the current video block. In yet another example, refined values of one step of refinement process are used for decoding other video blocks.

2. The method of clause 1, wherein the conversion generates the current block from the bitstream representation

3. The method of clause 1, wherein the conversion generates the bitstream representation from the current block.

4. The method of any of clauses 1 to 3, wherein the multi-step refinement process includes using refinement values of (i−1)th step for generating refinement values of ith step, wherein i=1 to N, where N is a total number of refinement steps performed during the multi-step refinement process and wherein N is greater than 1.

5. The method of any of clauses 1 to 4, wherein the performing conversion includes generating, in a step of the multi-step refinement process, a motion compensated reference block for the current block using refinement values of the motion vector information for the step.

6. The method of clause 5, wherein the generating the motion compensated reference block uses different filters for some steps of the multi-step refinement process.

7. The method of any of clauses 1 to 6, wherein the performing the conversion includes generating, at each step, a reference block using a refined motion vector generated for that step.

8. The method of clause 7, wherein the reference block is generated by first rounding the refined motion vector to an integer value.

9. The method of clause 1, wherein a step of the multi-step refinement process includes splitting the current block into multiple sub-blocks and performing an additional multi-step refinement process for at least some of the multiple sub-blocks.

10. The method of clause 9, wherein the additional multi-step refinement process is performed for all of the multiple-sub-blocks.

11. The method of clause 1, wherein the splitting the current block is performed after implementing a number of steps of the multi-step refinement process based on a characteristic of the current block.

12. The method of any of clauses 9 to 12 wherein the splitting the current block includes splitting the current block in a step-dependent manner such that, in at least two steps, sub-blocks of different sizes are used for the additional multi-step refinement process.

13. The method of any of clauses 1 to 12, wherein the selected one of the multiple refinement values is computed by a weighted average of the multiple refinement values.

14. The method of clause 1, wherein a number of steps performed prior to terminating the multi-step refinement process is decided based on changes in refinement values in successive steps exceeding a measure.

15. The method of clause 14, wherein the measure includes an absolute difference between a vertical or a horizontal component of a refinement value at a step and the motion information signaled in the bitstream.

16. The method of clause 14, wherein the number of steps performed prior to terminating the multi-step refinement process is a function of a characteristic of the current block.

17. The method of clause 16, wherein the characteristic of the current block includes a size of the current block or a coding mode of the current block or a value of motion vector difference of the current block or a quantization parameter.

18. The method of clause 14, wherein the number of steps performed prior to terminating the multi-step refinement process is signaled in the bitstream at a video level, a sequence level, a picture level, a slice level, a tile level, coding tree unit level or a coding unit level.

19. The method of any of clauses 1 to 18, wherein the selected one of the multiple refinement values is a refinement value calculated at a final step of the multi-step refinement process.

20. The method of any of clauses 1 to 19, wherein, during each step of the multi-step refinement process, an intermediate motion vector is generated from a previous step and used for calculating a refined estimate for a next step after the previous step.

21. A method of video processing, comprising:

• • performing, for conversion between a current block and a bitstream representation of the current block, a multi-step refinement process for a sub-block of the current block and a temporal gradient modification process between two prediction blocks of the sub-block, wherein, using the multi-step refinement process, multiple refinement values of motion vector information based on decoded motion information from a bitstream representation of the current video block; and
  • performing the conversion between the current block and the bitstream representation based on refinement values.

22. The method of clause 21, wherein second sub-block comprises M1×N1 pixels and the first sub-block comprises M2×N2 pixels where M1, N1, M2 and N2 are integers.

23. The method of clause 22, wherein M1=N1=8, M2=N2=4.

24. The method of any of clauses 21 to 23, wherein M1, N1, M2, N2 are pre-defined or depend on a size of the current block or a coding mode of the current block or signaled at a video level, sequence level, picture level, a slice level, a picture header level or a tile level.

25. The method of any of clauses 1 to 24, wherein each step of the multi-step refinement process uses partial pixels of the current block or a sub-block of the current block.

26. The method of clause 25, wherein the multi-step refinement process uses pixels from every Nth row or column.

27. A method of video processing, comprising: determining, using a multi-step decoder-side motion vector refinement process a current video block, a final motion vector; and performing conversion between the current block and the bitstream representation using the final motion vector.

28. The method of clause 27, wherein the multi-step decoder-side motion vector refinement process for the current macroblock is performed on a refinement values at an ith step of the multi-step refinement process recited in Clause 1, where i is an integer.

29. The method of any of clauses 27 to 28, wherein interpolation filters used for the multi-step refinement process are different from interpolation filters used for conversion of another block.

30. The method of any of clauses 1 to 26, wherein the selected one of the multiple refinement values is used for motion compensation of a subset of luma and Cr, Cb chroma components.

31. The method of clause 30, wherein the subset corresponds to the luma component.

32. The method of clause 30, wherein the subset corresponds to the chroma components.

33. The method of clause 32, wherein the motion component of the chroma components comprises a low bandwidth motion compensation process in which a subsampled version of the chroma components is used for motion compensation.

34. A method of video processing, comprising: applying, during conversion between a current video block and a bitstream representation of the current video block; multiple different motion vector refinement processes to different sub-blocks of the current video block; and performing conversion between the current block and the bitstream representation using a final motion vector for the current video block generated from the multiple different motion vector refinement processes.

35. The method of clause 34, wherein the conversion generates the current block from the bitstream representation

36. The method of clause 34, wherein the conversion generates the bitstream representation from the current block.

37. The method of any of clauses 34 to 36, wherein the multiple different motion vector refinement processes include a bi-directional optical flow process or a multi-step refinement process.

38. The method of any of clauses 34 to 37, wherein the multiple different motion vector refinement processes are selectively applied to the sub-blocks based on size or coding mode of the sub-blocks or positions of the sub-blocks within the current block or a type of reference picture used for coding the current block.

39. The method of clause 38, further including refraining from applying the multiple different motion vector refinement processes in case that the reference picture is a current coding picture of the current block.

40. The method of any of clause 1 to 39, wherein the current block corresponds to a prediction unit or a coding unit or a block or a sub-block.

41. A method of video processing, comprising:

• • performing a conversion between a current video block and a bitstream representation of the current video block using a rule that limits a maximum number of sub-blocks that a coding unit or a prediction unit in case that the current video block is coded using a sub-block based coding tool, wherein the sub-block based coding tool includes one or more of affine coding, advanced temporal motion vector predictor, bi-directional optical flow or a decoder-side motion vector refinement coding tool.

42. The method of clause 41, wherein the rule further specifies a width or a height of the sub-blocks.

43. The method of clause 42, wherein the current video block includes M×N pixels and wherein the sub-blocks have size K×L pixels, where K, L, M, N are integers.

44. The method of clause 43, wherein the width of each sub-block is maximum of THw and M/K, wherein THw is an integer number.

45. The method of any of clauses 43-44, wherein the width of each sub-block is maximum of THh and N/L, wherein THn is an integer number.

46. A method of video processing, comprising:

• • performing a conversion between a current video block and a bitstream representation of the current video block using a rule that specifies to use different partitioning for chroma components of the current video block than a luma component of the current video block in case that the current video block is coded using a sub-block based coding tool, wherein the sub-block based coding tool includes one or more of affine coding, advanced temporal motion vector predictor, bi-directional optical flow or a decoder-side motion vector refinement coding tool.

47. The method of clause 46, wherein the rule specifies partitioning for chroma components based on a size and a width of the current video block or the chroma component of the current video block.

48. The method of any of clauses 41-47, wherein the conversion comprises generating pixel values of the current video block from the bitstream representation.

49. The method of any of clauses 41-47, wherein the conversion comprises generating the bitstream representation from pixel values of the current video block.

50. The method of clause 9, wherein splitting the current block into multiple sub-blocks includes determining a size of the multiple sub-blocks based on a size of the current video block.

51. A method of video processing, comprising determining, in an early termination stage of a bi-directional optical flow (BIO) technique or a decoder-side motion vector refinement (DMVR) technique, differences between reference video blocks associated with a current video block. Further processing of the current video block based on the differences can be performed.

52. The method of clause 51, wherein determining the differences is based on even rows of the reference video blocks.

53. The method of clause 51, wherein determining the differences is based on corner samples of the reference video blocks.

54. The method of clause 51, wherein determining the differences is based on sub-blocks of the reference video blocks.

55. The method of clause 51, wherein the reference video blocks include a first reference video block and a second reference video block, the differences based on a summation of the differences between the first reference video block and the second reference video block.

56. The method of clauses 51-55, wherein the differences include one or more of: sum of absolute differences (SAD), sum of absolute transformed differences (SATD), sum of squares error (SSE), mean removed sum of absolute differences (MRSAD), mean removed sum of absolute transformed differences (MRSATD), or mean removed sum of squares error (MRSSE).

57. A video processing apparatus comprising a processor configured to implement a method recited in any one or more of clauses 1 to 56.

58. A computer program product stored on a non-transitory computer readable media, the computer program product including program code for carrying out the method in any one of clauses 1 to 56.

FIG. 32 is a flowchart for a method 3200 of video processing. The method 3200 includes calculating (3202), during a conversion between a current block of video and a bitstream representation of the current block, differences between two reference blocks associated with the current block or differences between two reference sub-blocks associated with a sub-block within the current block based on representative positions of the reference blocks or representative positions of the reference sub-blocks, and performing (3204) the conversion based on the differences.

In some example, calculating the differences comprises calculating differences of interlaced positions of the two reference blocks and/or two reference sub-blocks.

In some example, calculating the differences comprising calculating differences of even rows of the two reference blocks and/or two reference sub-blocks.

In some example, calculating the differences comprising calculating difference of four corner samples of the two reference blocks and/or two reference sub-blocks.

In some example, calculating the differences comprising calculating differences between two reference blocks based on representative sub-blocks within the reference blocks.

In some example, the representative positions are selected by using predetermined strategy.

In some example, the performing the conversion based on the differences comprises: summing up the differences calculated for the representative positions of the reference sub-blocks to obtain the difference for the sub-block.

In some example, the performing the conversion based on the differences comprises: summing up the differences calculated for the representative positions of the reference blocks to obtain the difference for the current block.

In some example, calculating the differences in an early termination stage of a motion vector refinement processing or a prediction refinement processing relying on difference calculation.

In some example, the performing the conversion based on the differences comprises: summing up the differences calculated for the representative positions to obtain the difference for the current block; determining whether a motion vector refinement processing or a prediction refinement processing relying on difference calculation is enabled or disabled for the current block based on the difference of the current block.

In some example, the prediction refinement processing includes bi-directional optical flow (BIO) technique, and/or the motion vector refinement processing includes a decoder-side motion vector refinement (DMVR) technique or a frame-rate up conversion (FRUC) technique.

In some example, the differences include one or more of: sum of absolute differences (SAD), sum of absolute transformed differences (SATD), sum of squares error (SSE), mean removed sum of absolute differences (MRSAD), mean removed sum of absolute transformed differences (MRSATD), or mean removed sum of squares error (MRSSE).

FIG. 33 is a flowchart for a method 3300 of video processing. The method 3300 includes: making (3302) a decision, based on a determination that a current block of a video is coded using a specific coding mode, regarding a selective enablement of a decoder side motion vector derivation (DMVD) tool for the current block, wherein the DMVD tool derives a refinement of motion information signaled in a bitstream representation of the video; and performing (3304), based on the decision, a conversion between the current block and the bitstream representation.

In some example, the DMVD tool is disabled upon a determination that prediction signal of the current block is generated at least based on an intra prediction signal and an inter prediction signal.

In some example, the DMVD tool is enabled upon a determination that prediction signal of the current block is generated at least based on an intra prediction signal and an inter prediction signal.

In some example, the current block is coded in a Combined Inter and Intra Prediction (CIIP) mode.

In some example, the DMVD tool is disabled upon a determination that the current block is coded with a Merge mode and motion vector differences.

In some example, the DMVD tool is enabled upon a determination that the current block is coded with a Merge mode and Motion Vector Differences.

In some example, the current block is coded in a Merge mode with Motion Vector Difference (MMVD) mode.

In some example, the DMVD tool is disabled upon a determination that the current block is coded with multiple sub-regions and at least one of them is non-rectangular.

In some example, the DMVD tool is enabled upon a determination that the current block is coded with multiple sub-regions and at least one of them is non-rectangular.

In some example, the current block is coded with the triangular prediction mode.

In some example, the DMVD tool comprises a decoder side motion vector refinement (DMVR) tool.

In some example, the DMVD tool comprises a bi-directional optical flow (BDOF) tool.

In some example, the DMVD tool comprises a frame-rate up conversion (FRUC) tool or other decoder-sider motion vector refinement method or sample refinement method.

In some example, the conversion generates the current block from the bitstream representation.

In some example, the conversion generates the bitstream representation from the current block.

The disclosed and other embodiments, modules and the functional operations described in this document can be implemented in digital electronic circuitry, or in computer software, firmware, or hardware, including the structures disclosed in this document and their structural equivalents, or in combinations of one or more of them. The disclosed and other embodiments can be implemented as one or more computer program products, i.e., one or more modules of computer program instructions encoded on a computer readable medium for execution by, or to control the operation of, data processing apparatus. The computer readable medium can be a machine-readable storage device, a machine-readable storage substrate, a memory device, a composition of matter effecting a machine-readable propagated signal, or a combination of one or more them. The term “data processing apparatus” encompasses all apparatus, devices, and machines for processing data, including by way of example a programmable processor, a computer, or multiple processors or computers. The apparatus can include, in addition to hardware, code that creates an execution environment for the computer program in question, e.g., code that constitutes processor firmware, a protocol stack, a database management system, an operating system, or a combination of one or more of them. A propagated signal is an artificially generated signal, e.g., a machine-generated electrical, optical, or electromagnetic signal, that is generated to encode information for transmission to suitable receiver apparatus.

A computer program (also known as a program, software, software application, script, or code) can be written in any form of programming language, including compiled or interpreted languages, and it can be deployed in any form, including as a stand-alone program or as a module, component, subroutine, or other unit suitable for use in a computing environment. A computer program does not necessarily correspond to a file in a file system. A program can be stored in a portion of a file that holds other programs or data (e.g., one or more scripts stored in a markup language document), in a single file dedicated to the program in question, or in multiple coordinated files (e.g., files that store one or more modules, sub programs, or portions of code). A computer program can be deployed to be executed on one computer or on multiple computers that are located at one site or distributed across multiple sites and interconnected by a communication network.

The processes and logic flows described in this document can be performed by one or more programmable processors executing one or more computer programs to perform functions by operating on input data and generating output. The processes and logic flows can also be performed by, and apparatus can also be implemented as, special purpose logic circuitry, e.g., an FPGA (field programmable gate array) or an ASIC (application specific integrated circuit).

Processors suitable for the execution of a computer program include, by way of example, both general and special purpose microprocessors, and any one or more processors of any kind of digital computer. Generally, a processor will receive instructions and data from a read only memory or a random-access memory or both. The essential elements of a computer are a processor for performing instructions and one or more memory devices for storing instructions and data. Generally, a computer will also include, or be operatively coupled to receive data from or transfer data to, or both, one or more mass storage devices for storing data, e.g., magnetic, magneto optical disks, or optical disks. However, a computer need not have such devices. Computer readable media suitable for storing computer program instructions and data include all forms of non-volatile memory, media and memory devices, including by way of example semiconductor memory devices, e.g., EPROM, EEPROM, and flash memory devices; magnetic disks, e.g., internal hard disks or removable disks; magneto optical disks; and CD ROM and DVD-ROM disks. The processor and the memory can be supplemented by, or incorporated in, special purpose logic circuitry.

While this patent document contains many specifics, these should not be construed as limitations on the scope of any invention or of what may be claimed, but rather as descriptions of features that may be specific to particular embodiments of particular inventions. Certain features that are described in this patent document in the context of separate embodiments can also be implemented in combination in a single embodiment. Conversely, various features that are described in the context of a single embodiment can also be implemented in multiple embodiments separately or in any suitable subcombination. Moreover, although features may be described above as acting in certain combinations and even initially claimed as such, one or more features from a claimed combination can in some cases be excised from the combination, and the claimed combination may be directed to a subcombination or variation of a subcombination.

Similarly, while operations are depicted in the drawings in a particular order, this should not be understood as requiring that such operations be performed in the particular order shown or in sequential order, or that all illustrated operations be performed, to achieve desirable results. Moreover, the separation of various system components in the embodiments described in this patent document should not be understood as requiring such separation in all embodiments.

Only a few implementations and examples are described and other implementations, enhancements and variations can be made based on what is described and illustrated in this patent document.

Claims

1. A method of video processing, comprising: the other of the first tool and the second tool comprises:

determining an enablement of a first tool, for a conversion between a current block and a bitstream based on a rule, wherein the rule indicates that the first tool and a second tool are not both used by the current block; and

performing, based on the determining, a conversion between the current block and the bitstream;

wherein one of the first tool and the second tool is a decoder side motion vector derivation tool to derive a refinement of motion information related to the current block, and

a combined inter and intra prediction tool in which a prediction signal of a block is generated at least based on an intra prediction signal and an inter prediction signal.

2. The method of claim 1, wherein the decoder side motion vector derivation tool is disabled upon a determination that the combined inter and intra prediction tool is used for the current block.

3. The method of claim 1, wherein the other of the first tool and the second tool further comprises:

a merge mode with motion vector differences which comprises motion vector expression comprising a first parameter representing a motion vector difference and a second parameter indicating a base candidate, wherein the first parameter comprises a motion magnitude and a motion direction.

4. The method of claim 3, wherein the decoder side motion vector derivation tool is disabled upon a determination that the current block is coded with the merge mode with motion vector differences.

5. The method of claim 1, wherein the other of the first tool and the second tool further comprises:

a geometric partitioning tool comprising a partition scheme which divides a block in to two partitions, at least one of which is non-square and non-rectangular.

6. The method of claim 5, wherein the geometric partitioning tool is a triangular prediction tool.

7. The method of claim 1, wherein the decoder side motion vector derivation tool comprises a decoder side motion vector refinement tool.

8. The method of claim 1, wherein the decoder side motion vector derivation tool comprises a bi-directional optical flow tool.

9. The method of claim 1, wherein the first tool is the decoder side motion vector derivation tool, and the conversion comprises:

dividing the current block into at least one subblock;

calculating differences between two reference subblocks associated with a subblock within the current block based on representative positions of the reference subblocks;

deriving the refinement of the motion information based on the differences.

10. The method of claim 9, wherein the differences are calculated in an early termination stage of the first tool.

11. The method of claim 9, wherein calculating differences comprises:

calculating differences of specific rows of two reference subblocks.

12. The method of claim 11, wherein the specific rows are composed of one row of every N rows, wherein N is greater than 1.

13. The method of claim 11, wherein N is equal to 2.

14. The method of claim 9, wherein the representative positions are selected by using predetermined strategy.

15. The method of claim 9, wherein deriving the refinement of the motion information based on the differences comprises:

summing up the differences calculated for the representative positions of the reference sub-blocks to obtain the difference for the sub-block;

deriving the refinement of the motion information using the difference for the subblock.

16. The method of claim 15, wherein the difference for the subblock include sum of absolute differences (SAD) of the representative positions.

17. An apparatus for processing video data comprising a processor and a non-transitory memory with instructions thereon, wherein the instructions upon execution by the processor, cause the processor to: the other of the first tool and the second tool comprises:

determine an enablement of a first tool, for a conversion between a current block and a bitstream based on a rule, wherein the rule indicates that the first tool and a second tool are not both used by the current block; and

perform, based on the determining, a conversion between the current block and the bitstream;

wherein one of the first tool and the second tool is a decoder side motion vector derivation tool to derive a refinement of motion information related to the current block, and

a combined inter and intra prediction tool in which a prediction signal of a block is generated at least based on an intra prediction signal and an inter prediction signal.

18. The apparatus of claim 17, wherein the other of the first tool and the second tool further comprises:

a merge mode with motion vector differences which comprises motion vector expression comprising a first parameter representing a motion vector difference and a second parameter indicating a base candidate, wherein the first parameter comprises a motion magnitude and a motion direction.

19. A non-transitory computer-readable storage medium storing instructions that cause a processor to: the other of the first tool and the second tool comprises:

determine an enablement of a first tool, for a conversion between a current block and a bitstream based on a rule, wherein the rule indicates that the first tool and a second tool are not both used by the current block; and

perform, based on the determining, a conversion between the current block and the bitstream;

wherein one of the first tool and the second tool is a decoder side motion vector derivation tool to derive a refinement of motion information related to the current block, and

a combined inter and intra prediction tool in which a prediction signal of a block is generated at least based on an intra prediction signal and an inter prediction signal.

20. A non-transitory computer-readable recording medium storing a bitstream which is generated by a method performed by a video processing apparatus, wherein the method comprises: the other of the first tool and the second tool comprises:

determining an enablement of a first tool, for a conversion between a current block and a bitstream based on a rule, wherein the rule indicates that the first tool and a second tool are not both used by the current block; and

generating the bitstream from the current block based on the determining;

wherein one of the first tool and the second tool is a decoder side motion vector derivation tool to derive a refinement of motion information related to the current block, and

a combined inter and intra prediction tool in which a prediction signal of a block is generated at least based on an intra prediction signal and an inter prediction signal.