Hadamard Transform Coefficient Ordering for Video Encoding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current video coding standards, such as MPEG, H.264, and H.265, are inefficient in encoding residual data due to their serial processing nature and reliance on prediction methods, which limits parallel processing capabilities in modern computing devices.
Innovation Solution
The method involves converting residual data into Hadamard transform coefficients and arranging them using a tile raster order or n-by-n block 'Z'-order for improved data compression, allowing for more efficient encoding and decoding of residual data, compatible with VC-6 and LCEVC standards.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If traditional serial processing methods are used for video encoding, then compatibility with legacy systems is maintained, but processing efficiency and compression performance deteriorate
Solution Approach 1:
The patent segments the Hadamard transform coefficients into multiple surfaces (e.g., 4 surfaces for 4x4 blocks) and processes each surface independently with different scan orders. This segmentation enables parallel processing while maintaining the transform's computational simplicity, resolving the contradiction between processing efficiency and device complexity.
Solution Approach 2:
The patent introduces a new dimension to the encoding process by organizing coefficients into multi-surface structures and applying different scan orders (raster, reverse raster, diagonal) to each surface. This dimensional organization enables parallel processing streams without increasing the fundamental complexity of the Hadamard transform itself.
2Productivity
If prediction methods are used in video coding, then temporal and spatial correlation is exploited, but parallel processing capability is limited
Solution Approach 1:
The patent extracts the prediction step from the core transform process, applying Hadamard transforms directly to prediction residuals rather than to original pixel data. This extraction allows parallel processing of residuals while still benefiting from prediction, as the prediction stage can operate independently and feed into the transform stage.
Solution Approach 2:
The patent performs prediction as a preliminary action before applying the Hadamard transform. By computing prediction residuals first and then transforming those residuals in parallel, the system maintains the benefits of prediction while enabling parallel processing in the transform and encoding stages.
3Loss of information
If standard coefficient ordering is used for Hadamard transform, then encoding simplicity is maintained, but data compression performance deteriorates
Solution Approach 1:
The patent applies different scan orders (raster, reverse raster, diagonal) to different surfaces of the Hadamard transform coefficients based on their local characteristics. Each surface can use the scan order that best compresses its specific data pattern, improving overall compression performance while keeping the complexity localized to the ordering logic rather than the transform itself.
Solution Approach 2:
The patent changes the ordering parameter of the coefficients by applying multiple scan order modes to different surfaces. This parameter change (from single ordering to multi-ordering) improves compression performance by better matching the statistical properties of different coefficient surfaces, while the complexity remains manageable through systematic application of standard scan patterns.
Data Source
AI summary
The present disclosure provides a method for encoding input data in an encoder to generate encoded output data, wherein the method includes: instructing the conversion of the input data into base-layer reconstruction data using a down-sampled version of the input data; and generating enhancement-layer data that provides instructions for enhancing the base-layer reconstruction data, wherein the enhancement-layer data comprises residual data representing a difference between a reconstruction generated from the base-layer reconstruction data and data derived from the input data, wherein the residual data are subjected to at least one transform and at least one data compression method to generate the encoded output data, characterized in that the method includes: arranging for the at least one transform to include a Hadamard transform to convert the residual data to corresponding transform coefficients, wherein different output elements of the Hadamard transform are arranged as different surfaces to be encoded, and selecting an output order for the transform coefficients in the surfaces to provide a sequence of transform coefficients for the at least one data compression method to generate the encoded output data, wherein the output order is selected as one of at least a tile raster order and a n-by-n block “Z”-order.