Abstract: Various schemes pertaining to video coding parallelization techniques are described. An apparatus receives video data. The apparatus subsequently calculates a plurality of figures of merits (FOMs), each of the FOM representing how well a particular coding tool may perform in encoding the video data. The apparatus further determines a coding tool that may be suitable for encoding the video data by comparing the FOMs. In determining the coding tool, the apparatus utilizes time-interleaving techniques to parallelly process the video data. The video data may include an array of coding blocks, and the apparatus may receive the video data using a snake-like processing order scanning through the array of coding blocks.

Abstract: Various schemes for managing search memory are described, which are beneficial in achieving enhanced coding gain, low latency, and/or reduced hardware for a video encoder or decoder. In processing a current block of a current picture, an apparatus determines a quantity of a plurality of reference pictures of the current picture. The apparatus subsequently determines, for at least one of the reference pictures, a corresponding search range size based on the quantity. The apparatus then determines, based on the search range size and a location of the current block, a search range of the reference picture, based on which the apparatus encodes or decodes the current block.

Abstract: A video encoder receives raw pixel data to be encoded as a current block of a current picture of a video into a bitstream. The video encoder identifies multiple candidate bi-prediction positions for the current block, including a center position, a first set of offset positions, and a second set of offset positions. The first set of offset positions and the second set of offset positions interleave each other. The encoder computes distortion values for each of the candidate bi-prediction positions based on several possible weighting parameter values. The distortion values for the center position are based on each of the several possible weighting parameter values. The distortion values for the first set of offset positions are based on a first subset of the possible weighting parameter values. The distortion values for the second set of offset positions are based on a second subset of the possible weighting parameter values.

Abstract: A video coding system generating candidates for Merge Mode with Motion Vector Difference (MMVD) with reduced resource usage is provided. The system receives data to be encoded or decoded as a current block of a current picture of a video. The system identifies multiple MMVD candidates for different offset positions based on a merge candidate of the current block. The system generates reference samples for the identified MMVD candidates. The system reconstructs the current block or encodes the current block into a bitstream by using the generated reference samples. The system processes the MMVD candidates in separate groups: a first group of vertical MMVD candidates and a second group of horizontal MMVD candidates. The system generates the reference samples for the identified MMVD candidates by applying a vertical filter to source reference samples of horizontal MMVD candidates and then applying a horizontal filter to outputs of the vertical filter.

Abstract: Video encoding methods and apparatuses for frequency domain mode decision include receiving residual data of a current block, testing multiple coding modes on the residual data, calculating a distortion associated with each of the coding modes in a frequency domain, performing a mode decision to select a best coding mode from the tested coding modes according to the distortion calculated in the frequency domain, and encoding the current block based on the best coding mode.

Abstract: The techniques described herein relate to methods, apparatus, and computer readable media configured to receive compressed video data, wherein the compressed video data is related to a set of frames. A decoder-side predictor refinement technique is used to calculate a new motion vector for a current frame from the set of frames, wherein the new motion vector estimates motion for the current frame based on one or more reference frames. An existing motion vector associated with a different frame from a motion vector buffer is retrieved. The new motion vector is calculated based on the existing motion vector using a decoder-side motion vector prediction technique, such that the existing motion vector is in the motion vector buffer after calculating the new motion vector.

Abstract: Video encoding methods and apparatuses include collecting statistics data, determining a matrix and vector representing a set of linear equations, solving the matrix and vector by a novel Gaussian elimination method to derive optimal parameter adjustments for an affine mode or adaptive loop filter coefficients, and encoding the current block by the affine mode or encoding one or more blocks by applying ALF filtering. Embodiments of the novel Gaussian elimination method reduce the critical path of entry operations in each row elimination step from one reciprocal, two multiplication, and one addition operations to one reciprocal, one multiplication, and one addition operations, or one multiplication and one addition operations.

Abstract: A methods and apparatus for adaptive data dependency between corresponding blocks partitioned from separate partition trees according to an embodiment of the present invention are disclosed. The data dependency between different tree structures is allowed when the tree structures are mutually inclusive. When the splitting is not used in one tree or when splitting is used in one tree and the same splitting or no split is used in the other tree, the data dependency between different tree structures is allowed. When the splitting in different tree structures are different, then the data dependency between different tree structures is disallowed. For each leaf CU in chroma tree, it should include one or more complete luma leaf CUs or it is completely included in one luma leaf CU, when the data dependency between different tree structures is allowed. Otherwise, then the data dependency between different tree structures is disallowed.

Abstract: Method and apparatus of video coding using decoder derived motion information based on bilateral matching or template matching are disclosed. According to one method, merge index for merge candidate group comprising bilateral matching merge candidate and/or template matching merge candidate are signalled using different codewords. According to another method, the first-stage MV or the first-stage MV pair is used as an only initial MV or MV pair or used as a central MV of search window for second-stage search. According to yet another method, after the reference template for a first reference list is found, the current template is modified for template search in the other reference list. According to yet another method, the sub-PU search is disabled for the template search. According to yet another method, block difference calculation is based on reduced bit depth during MV search associated with the decoder-side MV derivation process.

Abstract: A video coding method and apparatus include receiving input data associated with a current block, determining a coding mode for the current block by disabling Geometric Partitioning Mode (GPM) when a size of the current block is greater than or equal to a threshold size, and encoding or decoding the current block according to the determined coding mode. In a high-throughput video encoder performing Rate Distortion Optimization (RDO) by parallel Processing Elements (PEs), all or partial PEs receive search range reference samples in a broadcasting form. The parallel PEs test multiple coding modes on various partitioning for the current block, decide a block partitioning structure for dividing the current block into one or more coding blocks, and decide a coding mode for each of the coding blocks.

Abstract: Video encoding methods and apparatuses include receiving reconstructed video samples, determining an initial clipping setting for ALF coefficients, deriving clipping setting candidates from the initial clipping setting. ALF coefficients for the initial clipping setting and the clipping setting candidates are derived by solving inverse matrices, where partial intermediate results of solving ALF coefficients are shared by two or more clipping settings. A distortion value corresponds to the derived ALF coefficients for each clipping setting is computed, and final clipping indices for final ALF coefficients are determined according to the distortion values. ALF filtering is applied to the reconstructed video samples based on the final ALF coefficients and the final clipping indices.

Abstract: Video processing methods include receiving input data of a current block in a current slice, determine determining whether one or more components of the current block satisfy one or more predefined criteria during partitioning, and applying a mode constraint to the current block only if the one or more components of the current block satisfy the one or more predefined criteria, wherein the mode constraint restricts all blocks within the current block to be processed by a same prediction mode when the current block is split into a plurality of blocks. The methods adaptively split the current block into one or more blocks, and pare one or more prediction mode syntax elements of a first block in the current block according to a constrained mode of the current block. The methods further encode the current block with the mode constraint.

Abstract: Methods and apparatus for loop-filter processing of reconstructed video are disclosed. According to this method, if a loop-filter is disallowed for a target boundary, the target loop filter is disabled at top, bottom, left and right of the target boundary. According to another method, CC-ALF (cross-component adaptive loop filter) shape is changed across a virtual boundary. According to yet another method, CC-ALF and luma ALF handle unavailable samples in a same way.

Abstract: A method for video encoding and decoding using Low-Frequency Non-Separable Transform (LFNST) mode and apparatus thereof are disclosed. The input data correspond to primary transformed data at the encoder and the input data correspond to coded data of the current CU at the decoder. A CU is partitioned into one or more transform blocks (TBs). An LFNST syntax is determined at an encoder side or at a decoder side if one or more conditions are satisfied. The LFNST syntax indicates whether the LFNST mode is applied to the current CU and/or which LFNST kernel is applied when the LFNST mode is applied. The conditions comprise the CBF indications for the target TBs being false. The current CU is encoded or decoded according to the LFNST mode as indicated by the LFNST syntax.

Abstract: According to a method for Adaptive Loop Filter (ALF) processing of reconstructed video, multiple indicators are signaled in slice at an encoder side or parsed at a decoder side, where the multiple indicators are Adaptive Parameter Set (APS) indices associated with temporal ALF filter sets for the ALF processing. A current indicator is determined from the multiple indicators, where the current indicator is used to select a current ALF filter set. Filtered-reconstructed pixels are derived for the current block by applying the current ALF filter to the current block. In another method, if the ALF processing applied at a target sample requires an outside sample on other side of a target virtual boundary from the target sample, the outside sample is replaced by a padded sample.

Abstract: The techniques described herein relate to methods, apparatus, and computer readable media configured to receive compressed video data, wherein the compressed video data is related to a set of frames. A decoder-side predictor refinement technique is used to calculate a new motion vector for a current frame from the set of frames, wherein the new motion vector estimates motion for the current frame based on one or more reference frames. An existing motion vector associated with a different frame from a motion vector buffer is retrieved. The new motion vector is calculated based on the existing motion vector using a decoder-side motion vector prediction technique, such that the existing motion vector is in the motion vector buffer after calculating the new motion vector.

Abstract: Video encoding methods and apparatuses for performing rate-distortion optimization by a hierarchical architecture include receiving input data associated with a current block in a current picture, determining a block partitioning structure to split the current block into coding blocks and determining a corresponding coding mode for each coding block by multiple Processing Element (PE) groups, and entropy encoding the coding blocks in the current block according to the coding modes determined by the PE groups. Each PE group has parallel PEs and is associated with a particular block size. The parallel PEs in each PE group test a number of coding modes on each partition or sub-partition of the current block to derive rate-distortion costs. The block partitioning structure and corresponding coding modes are then decided based on the rate-distortion costs derived by the PE groups.

Abstract: Method and apparatus of video coding are disclosed. According to one method, the left reference boundary samples and the top reference boundary samples are checked jointly. According to another method, selected original left reference boundary samples and selected original top reference boundary samples at specific positions are used for predictor up-sampling. According to yet another method, the horizontal interpolation and the vertical interpolation are in a fixed order regardless of a shape of the current block, size of the current block or both.

Abstract: An intra prediction method comprises receiving input data of a current block in a current picture, deriving multiple Most Probable Modes (MPMs) to be included in a MPM list for the current block, setting remaining intra prediction modes as non-MPMs, and encoding or decoding the current block according to a current intra prediction mode selecting from the MPMs and non-MPMs. The first MPM in the MPM list is Planar mode for blocks coded or to be coded in intra prediction, and one or more other MPMs in the MPM list are derived according to a number of available angular mode of one or more neighboring blocks of the current block.

Abstract: Low-latency video coding methods and apparatuses include receiving input data associated with a current Intra slice composed of Coding Tree Units (CTU), where each CTU includes luma and chroma Coding Tree Blocks (CTBs), partitioning each CTB into non-overlapping pipeline units, and encoding or decoding the CTUs in the current Intra slices by performing processing of chroma pipeline units after beginning processing of luma pipeline units in at least one pipeline stage. Each of the pipeline units is processed by one pipeline stage after another pipeline stage, and different pipeline stages process different pipeline units simultaneously. The pipeline stage in the low-latency video coding methods and apparatuses simultaneously processes one luma pipeline unit and at least one previous chroma pipeline unit within one pipeline unit time interval.