Intra Block Copy Buffer Constraints for Valid Block Vectors
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current video coding technologies face challenges in efficiently managing buffer memory and block vectors for intra block copy mode, leading to complex processing, invalid block vectors, inefficient coding, and unclear handling of CTU sizes smaller than 128x128, particularly in screen content coding.
Innovation Solution
Implementing a regular buffer for intra block copy mode to store reference samples, using defined block vectors and bitstream conformance constraints to ensure valid pixel displacements, and applying in-loop reshaping to convert signals into a reshaped domain for unified processing.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If intra block copy mode is implemented without regular buffer management, then coding flexibility is maintained, but buffer memory management becomes complex and generates invalid block vectors
Solution Approach 1:
The reference buffer is segmented into distinct regions (current CTU buffer and previous CTU buffer) with clear boundaries and management rules. Each segment has specific responsibilities for storing reference samples, and the segmentation allows independent management of buffer validity and content, reducing overall complexity while ensuring block vector validity.
Solution Approach 2:
The buffer is pre-initialized with valid reference samples before intra block copy operations begin. By preparing the buffer in advance with proper validity flags and reference data, the system avoids generating invalid block vectors during decoding, eliminating the need for complex runtime validation logic.
2Measurement precision
If complex buffer management is used to track reference samples, then reference availability is accurately tracked, but encoder/decoder processing becomes complex
Solution Approach 1:
The buffer management system is self-servicing through automatic validity flag updates and reference sample tracking. As blocks are decoded and reconstructed, the buffer automatically updates its own state, marking samples as available or unavailable based on decoding progress, eliminating the need for complex external tracking mechanisms.
Solution Approach 2:
The system uses simple binary validity flags (available/unavailable) to track reference sample status, changing these parameters automatically as decoding progresses. This parameter-based approach provides precise tracking of reference availability while keeping the processing logic simple and efficient.
3Adaptability or versatility
If CTU sizes larger than 128x128 are supported, then coding flexibility is improved, but handling of smaller CTU sizes becomes unclear
Solution Approach 1:
The buffer management system is designed to be universal and adaptable to any CTU size. The same buffer structure and management rules apply regardless of whether the CTU is 64x64, 128x128, or larger, providing consistent behavior across all sizes and eliminating implementation ambiguity for smaller CTU sizes.
Solution Approach 2:
The buffer dimensions and management parameters are dynamically adjusted based on the actual CTU size being processed. This dynamic adaptation allows the system to handle various CTU sizes uniformly, with the buffer automatically scaling to accommodate different resolutions without requiring size-specific handling logic.
4Measurement precision
If more reference samples are stored in the buffer, then prediction accuracy is improved, but memory usage increases
Solution Approach 1:
The buffer implements a sliding window approach where reference samples are continuously updated: as new blocks are decoded, their samples are added to the buffer, and older samples that are no longer needed for prediction are automatically discarded. This maintains prediction accuracy by keeping relevant reference data while managing memory usage efficiently.
Solution Approach 2:
The buffer stores reference samples selectively based on their utility for prediction. Rather than storing all possible samples or using a fixed large buffer, the system stores only the necessary reference samples required for accurate intra block copy prediction, achieving good prediction accuracy with minimal memory usage.
Data Source
AI summary
A method of processing visual media includes performing a conversion between a current video block of a current picture of a visual media data and a bitstream representation of the visual media data using a buffer comprising reference samples from the current picture for derivation of a prediction block of the current video block. The conversion is based according to rule which specifies that, for the bitstream representation to conform the rule, a reference sample in the buffer is to satisfy a bitstream conformance constraint.