Delta QP Signaling in Video Coding Using Block Depth
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The High-Efficiency Video Coding (HEVC) standard faces challenges in efficiently signaling quantization parameters for video data partitioned into square and non-square blocks, particularly due to increased coding complexity and signaling overhead when using Quad-Tree-Binary-Tree (QTBT) and Multi-Type-Tree (MTT) structures, which complicate the determination and signaling of delta QP and chroma QP offset for non-square blocks.
Innovation Solution
The proposed method determines the depth of a current split node based on its splitting type and compares it with a maximum delta QP signaling depth to derive and signal delta QP, allowing for efficient quantization parameter signaling in video encoding and decoding systems. This involves parsing syntax elements and reconstructing final QPs for Transform Units (TUs) within the current split node, enabling efficient encoding and decoding processes even for non-square blocks.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If QTBT and MTT structures are used to partition video data into square and non-square blocks, then the adaptability to various local characteristics is improved, but the device complexity and signaling overhead increase
Solution Approach 1:
The patent changes the parameter for determining delta QP from block shape type to block area or depth metrics. By using area thresholds or depth comparisons, the system maintains adaptability to different block sizes (including non-square blocks in QTBT/MTT structures) while simplifying the determination process. The encoder/decoder compares the current block's area or depth against predefined thresholds to decide whether to signal delta QP, avoiding complex shape-based logic.
Solution Approach 2:
The patent applies different delta QP signaling strategies based on local block characteristics (area or depth) rather than applying a uniform rule across all blocks. By partitioning the decision space into regions (blocks meeting area/depth thresholds vs. those that don't), the system optimizes signaling for each local context - signaling delta QP for larger/deeper blocks where it provides benefit, and omitting it for smaller/shallower blocks where overhead would dominate.
2Manufacturing precision
If delta QP is signaled for all blocks in QTBT and MTT structures, then the manufacturing precision of quantization parameters is improved, but the loss of information increases due to higher signaling overhead
Solution Approach 1:
Instead of signaling delta QP for all blocks (excessive action) or no blocks (insufficient action), the patent applies partial action by selectively signaling delta QP only for blocks that meet specific area or depth criteria. This partial signaling approach provides sufficient precision for blocks where it matters (larger/deeper blocks with more transform coefficients) while avoiding the overhead of signaling for blocks where the benefit is minimal.
Solution Approach 2:
The patent changes the selection criterion from a universal rule (signal for all blocks or none) to a parameter-based rule (signal based on area or depth thresholds). This allows the system to adapt the precision of quantization parameter signaling to the actual needs of each block, maintaining high precision where required while reducing overall signaling overhead through parameter-driven selection.
3Adaptability or versatility
If the minimum CU size is reduced to 8×8 to enable finer prediction switching, then the adaptability to local characteristics is improved, but the device complexity increases due to more partitioning decisions
Solution Approach 1:
The patent segments the block partitioning process into hierarchical levels (CTU → CU → TU) with the quadtree structure providing coarse segmentation and transform units providing fine-grained segmentation for residual coding. This multi-level segmentation allows the system to use larger CUs for prediction (reducing partitioning complexity) while still achieving fine granularity for transform and quantization through subsequent TU partitioning, effectively decoupling prediction granularity from transform granularity.
Data Source
AI summary
Video data processing methods and apparatuses receive input data associated with a current split node partitioned from a parent node by a splitting type, determine a depth of the current split node according to the splitting type, and compare the depth of the current split node with a maximum delta QP signaling depth. A video decoding system derives a delta QP from one or more syntax elements signaled in a TU associated with the current split node according to the comparing result, reconstructs a final QP for the current split node based on a reference QP and the delta QP, and decodes one or more TUs associated with the current split node using the final QP. The depth is counted in a way considering different splitting types and splitting partitions.