PSAO Filtering Parameter Range Extension
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current video compression systems, such as HEVC and VVC, face limitations in the adaptation of Pre-Sample Adaptive Offset (PSAO) filtering due to restricted ranges of PSAO parameter values, which can lead to suboptimal filtering performance for video picture data after deblocking.
Innovation Solution
The proposed solution extends the ranges of PSAO parameter values, allowing for positive and negative offset coefficients and thresholds, enabling better adaptation to video picture data by defining these parameters within wider ranges and encoding them efficiently to maintain bandwidth.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the ranges of PSAO parameter values are restricted, then the encoding bandwidth is maintained, but the filtering performance is suboptimal
Solution Approach 1:
The patent extends the range of PSAO parameter values from the conventional limited range to include both positive and negative values. Specifically, the offset coefficient range is expanded from {0, 1, 2, 3} to {-Dmax, ..., Dmax} where Dmax ≥ 3, and the threshold range is expanded from {0, ..., Tmax} to {-Tmax, ..., Tmax} where Tmax ≥ 2. This parameter change enables better adaptation to different video picture data characteristics while maintaining encoding efficiency through optimized bit allocation for sign and magnitude.
2Adaptability or versatility
If the ranges of PSAO parameter values are extended to include positive and negative values, then the adaptation to video picture data is improved, but the encoding complexity increases
Solution Approach 1:
The patent segments the encoding of PSAO parameters into two independent parts: sign encoding and magnitude encoding. The sign bit indicates whether the parameter is positive or negative, while the magnitude is encoded separately. This segmentation allows the decoder to efficiently reconstruct the full range parameter values without increasing overall encoding complexity, as each component can be processed independently using standard coding techniques.
Solution Approach 2:
The patent adds a sign dimension to the previously one-dimensional non-negative parameter space. By separating the parameter representation into sign (positive/negative) and magnitude components, the system effectively transforms the parameter space from [0, Dmax] to [-Dmax, Dmax], enabling broader adaptation while maintaining efficient encoding through the added sign dimension.
3Productivity
If the PSAO parameter ranges are extended, then the compression performance is improved, but the bitrate increases
Solution Approach 1:
The patent implements partial extension of parameter ranges by allowing both positive and negative values, which provides the necessary flexibility for improved compression performance. The extension is controlled and bounded by maximum values (Dmax for offset coefficients, Tmax for thresholds), preventing excessive bitrate increase. This partial extension achieves the optimal balance between compression gain and bitrate cost by only extending the range as much as needed for practical video data adaptation.
Data Source
Figure 1~2
Figure 3~4
Figure 5
AI summary
The present application relates to a method of encoding or decoding video picture comprising pre-Sample Adaptive Offset vertical filtering and pre-Sample Adaptive Offset horizontal filtering using PSAO parameter values intended to be applied on video picture data, wherein PSAO parameter values are either positive or negative values.