Video Coding Phase Indication for Downsampling Precision
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current video coding standards face challenges in efficiently signaling downsampling information, leading to potential issues in video data compression and decoding processes, particularly in handling various sampling formats and filter operations.
Innovation Solution
The proposed solution involves signaling phase indication information for video data by expressing horizontal and vertical positions of luminance sampling locations relative to a display window using specific mathematical expressions, allowing for accurate downsampling and filtering operations across different video coding standards, including ITU-T H.264, H.265, JEM, and JVET-T2001.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional video coding standards are used to signal downsampling information, then video compression can be achieved, but precision and consistency of downsampling operations deteriorate due to insufficient sampling location information
Solution Approach 1:
The patent introduces new parameters (phase offset values for horizontal and vertical positions) to the video coding standard to precisely specify sampling locations. By changing the parameter set used to describe downsampling operations, the patent enables exact control over sampling phase, thereby improving precision without losing critical location information.
Solution Approach 2:
The patent uses phase indication information as an intermediary element that bridges the gap between encoded video data and accurate downsampling reconstruction. This intermediary carries the necessary sampling location information that was previously lost, enabling precise recovery of luminance and chrominance sample positions during decoding.
2Manufacturing precision
If downsampling information is not explicitly signaled, then bitstream complexity is reduced, but decoding accuracy deteriorates due to ambiguous sampling locations
Solution Approach 1:
The patent applies preliminary action by encoding phase offset information into the bitstream during the encoding phase. This pre-signaling of sampling location data ensures that the decoder has all necessary information upfront to perform accurate downsampling operations, eliminating the need for complex runtime calculations or assumptions about sampling positions.
Solution Approach 2:
By modifying the bitstream to include phase indication parameters (horizontal and vertical offset values), the patent increases bitstream complexity in a controlled manner to achieve significantly higher decoding accuracy. The added parameters provide explicit sampling location data that resolves ambiguity in downsampling operations.
3Adaptability or versatility
If various sampling formats are supported without phase information, then video compression flexibility is improved, but filter operation consistency deteriorates
Solution Approach 1:
The patent creates a universal phase indication mechanism that works across all sampling formats (4:4:4, 4:2:2, 4:2:0, etc.). The same phase offset parameters can be applied regardless of the specific sampling format, providing a multi-functional solution that maintains filter operation consistency across diverse video formats while preserving format flexibility.
Solution Approach 2:
The patent uses parameter changes by introducing phase offset values that can be independently adjusted for different sampling formats. This allows the filtering and downsampling operations to adapt to each format's specific requirements while maintaining consistent mathematical relationships, ensuring that filter operations remain stable and predictable across all supported formats.
Data Source
AI summary
A device may be configured to signal downsampling offset information for coded video in a phase indication information message. A device may be configured to receive the phase indication information message. The phase indication message may apply to a current cropped decoded picture and persist for all subsequent pictures of a current layer in output order with the same cropped picture width and cropped picture height as the current cropped decoded picture until a picture in the current layer with an associated phase indication information message and the same cropped picture width and cropped picture height as the current cropped decoded picture is output and follows the current picture in output order.