Video Encoder Deblocking Filter Length at TU Boundaries
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video coding technologies face challenges in improving encoding efficiency, image quality, reducing processing amount, and circuitry scale, while maintaining processing speed.
Innovation Solution
An encoder and decoder implementation that simplifies the deblocking filtering process by determining maximum filter lengths based on sub-block boundaries at intervals of eight samples, reducing the need to check for TU boundaries at four-sample intervals, and accounting for CU edges to enhance accuracy and reduce circuitry complexity.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If deblocking filtering is performed by checking TU boundaries at four-sample intervals, then filtering accuracy is improved, but processing complexity and circuitry scale increase
Solution Approach 1:
The patent divides the CU into sub-blocks with sub-block boundaries at eight-sample intervals, and further divides into TU blocks. The deblocking filtering is performed by checking boundaries at both eight-sample intervals (sub-block boundaries) and four-sample intervals (TU boundaries), but the segmentation allows selective checking rather than exhaustive checking at all positions, reducing overall complexity while maintaining accuracy where needed.
Solution Approach 2:
The patent applies different filtering strategies at different locations: at sub-block boundaries (eight-sample intervals), it performs determination processes based on TU boundary checks; at TU boundaries (four-sample intervals), it applies filtering when specific conditions are met. This local differentiation ensures accurate filtering at critical boundaries while reducing processing at less critical positions, resolving the contradiction between accuracy and complexity.
2Measurement precision
If exhaustive TU boundary checks are performed at all sample locations, then filtering precision is improved, but processing time increases
Solution Approach 1:
The patent performs preliminary determination processes at sub-block boundaries (eight-sample intervals) to identify potential TU boundaries before performing exhaustive checks. The flowchart shows that the encoder/decoder first checks if a location at four-sample displacement is a TU boundary, and only performs additional determination processes when needed. This preliminary filtering reduces the number of exhaustive checks required, decreasing processing time while maintaining detection precision.
3Reliability
If maximum filter length is increased to cover all possible TU boundaries, then filtering completeness is improved, but processing complexity increases
Solution Approach 1:
The patent dynamically determines the maximum filter length based on the specific boundary conditions detected during processing. Rather than using a fixed maximum length that covers all possible cases, the filter length is adapted according to the actual TU boundary locations and sub-block boundary positions. This dynamic adjustment ensures filtering completeness when needed while reducing complexity when boundaries are fewer or more regularly spaced.
Data Source
AI summary
For a location displaced by four samples in a vertical direction or a horizontal direction from a current location, the encoder performs a first determination of determining only whether the location displaced by four samples is a TU boundary, where the current location is a sample location of a current sub-block boundary on which the determination process is to be performed. In the first determination, when it is determined that the location displaced by four samples is a TU boundary, the encoder sets a maximum filter length to a first value, and in the case otherwise, the encoder performs a second determination of determining whether a location displaced by eight samples in the vertical direction or the horizontal direction from the current location is a TU boundary.