Adaptive In-Loop Filtering for Bit-Depth-Aware Video Coding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video coding technologies lack an efficient mechanism for encoding and decoding chroma and luma components of an image frame while maintaining image quality, particularly in adaptive loop filtering processes.
Innovation Solution
Implementing a bit depth based clipping operation in adaptive loop filtering for chroma and luma samples, where each sample is filtered based on surrounding samples using a respective adaptive loop filter with a defined clip boundary value, and applying a clipping operation based on a clip value index and filter coefficient.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If adaptive loop filtering is applied to chroma and luma components, then image quality is improved, but computational complexity increases
Solution Approach 1:
The patent applies bit depth based clipping operations where the clipping boundary is dynamically adjusted according to the bit depth of the video data. For example, 8-bit data uses a clipping boundary of 128 while 10-bit data uses 512. This parameter adaptation allows the filter to maintain optimal performance across different bit depths without requiring separate filter designs, thereby improving image quality while controlling computational complexity through a unified approach.
Solution Approach 2:
The patent implements selective filtering where the adaptive loop filter is applied differently to luma and chroma components based on their respective importance to perceived image quality. The filter strength and clipping parameters are adjusted per component type, applying more aggressive filtering to luma where quality is more critical, while using milder filtering for chroma to reduce computational load where quality tolerance is higher.
2Measurement precision
If bit depth based clipping operation is implemented, then filtering accuracy is improved, but processing time increases
Solution Approach 1:
The patent pre-calculates and stores optimal clipping boundaries for different bit depths (e.g., 128 for 8-bit, 512 for 10-bit, 2048 for 12-bit) in lookup tables or as hardcoded constants. During the adaptive loop filtering process, the system simply retrieves the appropriate clipping boundary based on the bit depth parameter rather than performing complex real-time calculations. This preliminary preparation significantly reduces processing time while maintaining high filtering accuracy through bit depth adaptive clipping.
3Adaptability or versatility
If clip boundary value is increased, then dynamic range is improved, but risk of overflow increases
Solution Approach 1:
The patent implements a feedback mechanism where the clipping operation monitors the bit depth of the input video data and dynamically adjusts the clipping boundary value accordingly. The system uses the relationship ClipBoundary = 2^(bit_depth - 1) to automatically select appropriate boundaries: 128 for 8-bit data, 512 for 10-bit data, and 2048 for 12-bit data. This feedback-driven adaptation ensures the clipping boundary always matches the data range, maximizing dynamic range utilization while preventing overflow by never setting the boundary beyond what the bit depth can represent.
Data Source
AI summary
This application is directed to coding video data that includes a plurality of image samples of a video frame. Each image sample corresponds to one of a luma sample and a chroma sample. Each image sample is filtered using an adaptive in-loop filter having a filter length and a set of filter coefficients. A set of related image samples are identified in the filter length of each image sample. For each related image sample, a respective clip value index and a corresponding filter coefficient are identified. A difference of each related image sample and the respective image sample is clipped based on the respective clip value index that corresponds to a respective clipping boundary value equal to 2 to a power of a respective clipping number. The respective image sample is modified with the clipped difference of each of the related image samples based on the respective filter coefficient.