Precoding adjusts QP before real encoding to refine bit rate control and compression.
The method uses a single identifier for block size, prediction mode, or color component to select quantization matrices and improve quality.
Multiple encoded views target changing areas of interest, using selective blurring to smooth transitions and improve bandwidth efficiency.
This case uses adaptive context selection and multiple probability updates to balance CABAC stability with changing symbol probabilities.
This case uses inverse reduced secondary transforms and mode-adaptive kernels to improve compression efficiency in high-resolution video.
This case uses raster-ordered rectangular slices and derived entry points to speed multicore image encoding and decoding.
This case combines quad, binary, ternary, and arbitrary block partitioning with adaptive syntax to improve coding efficiency.
A CNN filter uses quantization and block partition maps to improve reconstructed video quality without increasing decoding complexity.
This case derives coded-data entry positions from slice parameters to reduce bitstream overhead and speed multicore image processing.
Transparency-clustered MPI patches carry layer-specific depth quantization laws for clearer rendering from varied viewpoints.
This video decoding case uses grouped intra modes and template costs to improve block prediction without evaluating every mode.
Frame complexity and encoding budgets guide channel-level allocation, increasing encoding density while limiting distortion.
A characteristics message signals application purpose and URI information, adapting neural post-filters without changing core video coding.
A segmented reference-sample buffer simplifies intra block copy processing and supports valid block vectors across CTU sizes.
Quadtree partitioning and selective edge refinement improve prediction accuracy while controlling complexity in screen-content video coding.
This case partitions transform units by coefficient groups to limit context count while preserving accurate significance-map coding.
Machine learning estimates content distortion to select encoding resolution for target bitrate without costly multi-pass analysis.
This case signals changing PDU set importance in coded volumetric video to improve viewport-dependent transmission and decoding efficiency.
This case signals applicable prediction types across base and enhancement layers to balance coding flexibility and compression efficiency.
Residue-domain prediction, selective coefficient sorting, and sign data hiding simplify decoding while reducing transmitted bit overhead.
This case uses frame identifiers, selective reference-block matching, and candidate-vector scaling to improve video prediction efficiency.
Gain and shape indices reconstruct prediction residuals from a codebook, reducing bitrate and decoding time without inverse transforming.
This case uses partial reconstruction and unrefined motion information to reduce latency during template-based video block decoding.
Tile-row thresholds trigger slice-width parsing only when needed, improving bitstream efficiency while preserving decoding precision.
A CT information decoder simplifies chroma modes and tree splits to reduce delay and improve luma-chroma throughput.
Block-specific disparity vectors cut bit rates in multi-view residual coding.
Hierarchy depth information selects where decoding parameters are stored, balancing fine-grained image quality with coding efficiency.
Conditional delay signaling improves video decoding accuracy while limiting bitstream overhead.
This coding case reorders motion vector predictors by cost, reducing redundant candidates while balancing bitrate and processing complexity.
Multiple prediction blocks are compared, while residual flags reduce calculation complexity and bitstream overhead for UHD images.
This image decoding case combines bi-prediction with additional reference blocks to improve accuracy and compression efficiency.
A spatial rate factor allocates bits across video layers to balance quality, bandwidth use, and encoding cost.
This case derives affine candidates from selected neighboring motion vectors, reducing temporal buffer storage for video coding.
This case uses CIIP-aware merge flags and differential motion data to reduce syntax overhead in video decoding.
This coding case restricts MIP coefficients to binary forms, reducing multiplication complexity while preserving useful prediction accuracy.
This case uses scalability identifiers and slice-tile partitioning to support parallel decoding with lower coding complexity.
Unavailable boundary samples are padded from the current processing unit, simplifying adaptive loop filtering across video-unit boundaries.
This video coding case uses the predictor index to signal MVD selectively, reducing motion-vector bits and improving compression efficiency.
A universal video engine fetches format-specific code and frames, reducing controller utilization while maintaining high frame rates.
Piecewise optical flow parameters improve video compression without sacrificing picture quality.
This case adapts luma and chroma mode derivation with flags and pixel lines to improve prediction accuracy and coding efficiency.
Scale-and-shift quantization adapts to block sizes and transform types, improving coding efficiency and visual quality.
This case derives same-size MIP blocks for power-of-two coding blocks, reducing VVC computation and storage without up-sampling.
Common pivot points and constrained optimization reduce gaps while preserving HDR reconstruction quality across SDR and HDR displays.
This case combines predictive and residual coding with rate-distortion switching to compress sparse neural-network weight updates.
This case uses raster-ordered rectangular slices and derived entry positions to improve multicore encoding and decoding efficiency.
This case uses MTS, inverse transforms, and tree-dependent contexts to improve compression for high-quality image data.
The case selects and encodes only the effective quantization matrix, reducing VVC bit overhead and decoder processing demands.
This case derives neighbor prediction modes to select context models and VLC tables, improving intra coding efficiency for HD imagery.
This case applies 15- or 31-tap deblocking selectively to large smooth blocks, reducing artifacts while preserving coding efficiency.