Block-Wise Quantization in Encoders for Adaptive Compression
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current data compression methods lack flexibility and efficiency in switching between different quantization strengths and coding methods on a block-by-block or packet-by-packet basis, leading to suboptimal compression ratios and quality, especially when dealing with diverse data types and channels.
Innovation Solution
A method and encoder/decoder system that sets quality parameters and selects encoding methods and quantization strengths on a portion-specific basis for each data block or packet, allowing for different quantization strengths and methods to be used for the same quality parameter, with information about these selections embedded in the encoded data for efficient decoding.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If a single generic encoding method and quantization strength are used for all data blocks, then device complexity is reduced and ease of operation is improved, but compression ratio and quality deteriorate because different methods are better suited for different kinds of data blocks
Solution Approach 1:
The patent applies local quality by selecting different encoding methods and quantization strengths for different data blocks based on their specific characteristics. The system analyzes each data block and applies the most suitable encoding method (e.g., DCT, DST, Hadamard) and quantization strength locally, rather than using a uniform approach across all blocks, thereby optimizing compression ratio and quality for each block's specific content type.
Solution Approach 2:
The patent implements dynamics by making the encoding method and quantization strength adjustable and switchable on a block-by-block basis. The system dynamically selects from multiple available encoding methods and quantization strengths depending on the data block characteristics, allowing the encoding parameters to adapt flexibly to different content types rather than being fixed globally.
2Productivity
If multiple different encoding methods and quantization strengths are selected for different data blocks, then compression ratio and quality are improved, but device complexity increases due to the need to manage and switch between multiple methods
Solution Approach 1:
The patent applies preliminary action by pre-defining multiple encoding methods and quantization strengths in the system before actual encoding operations. The system prepares a library of available encoding methods (DCT, DST, Hadamard, etc.) and quantization strength levels in advance, so that during encoding, the system can quickly select from these pre-prepared options based on data block characteristics, reducing the operational complexity of managing multiple methods.
3Manufacturing precision
If different quantization strengths are used for the same quality parameter across different data blocks, then manufacturing precision of quality control is improved, but device complexity increases due to the need to manage multiple quantization parameters
Solution Approach 1:
The patent applies parameter changes by allowing the quantization strength parameter to vary independently for different data blocks while maintaining the same quality parameter target. The system changes the quantization strength parameter based on data block characteristics (e.g., complexity, content type), enabling precise quality control for each block by adjusting the quantization parameter to match the block's specific requirements rather than using a uniform quantization strength.
Data Source
AI summary
A method of encoding input data to generate corresponding encoded data includes: setting one or more values of a quality parameter for the plurality of data portions in a portion-specific manner; analysing the plurality of data portions and selecting at least one encoding method and at least two different quantization strengths for encoding the plurality of data portions, based upon a given value of the quality parameter set for the given data portion; encoding the plurality of data portions to generate corresponding encoded data portions, wherein at least two data portions are encoded by an encoding method and different quantization strengths, for a same value of the quality parameter for the at least two data portions; aggregating the encoded data portions to generate encoded data; and assembling information indicative of the at least two quantization strengths and the one or more values of the quality parameter into the encoded data.