Macroblock Quantization Rounding for Low-Bitrate Image Encoding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
High-resolution image encoding at low bit rates in MPEG schemes results in significant image degradation due to insufficient code allocation to image information, especially with an increased number of macroblocks, as the proportion of header information increases, making it impossible to allocate sufficient codes to image data.
Innovation Solution
An image encoding apparatus that generates and rounds quantization information based on the amount of codes generated, allowing for the omission of quantization information when differences between blocks are within a set range, thereby reducing the number of quantization values and increasing code allocation to image information.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If high-resolution images are encoded at low bit rates, then the target bit rate is reduced, but image degradation becomes significant due to insufficient code allocation to image information
Solution Approach 1:
The patent extracts and removes quantization information from the encoded stream when consecutive macroblocks have identical quantization values. By taking out redundant header information (Q-scale) that would otherwise consume valuable bits, more code allocation is available for image information, thereby maintaining image quality at low bit rates
Solution Approach 2:
The patent changes the parameter of quantization information representation by using differential encoding and run-length encoding techniques. Instead of transmitting absolute Q-scale values for every macroblock, the system transmits differences from previous values and uses run-length encoding to represent sequences of identical values, significantly reducing the number of bits required for header information
2Area of stationary object
If the number of macroblocks is increased for high-resolution images, then the resolution is improved, but the proportion of header information increases, making it impossible to allocate sufficient codes to image data
Solution Approach 1:
The patent merges consecutive macroblocks that have identical quantization values into a single representation. By combining multiple macroblocks with the same Q-scale into one header entry using run-length encoding, the number of header bits required is dramatically reduced, allowing sufficient code allocation to image data even when the number of macroblocks is large
Solution Approach 2:
Instead of encoding each macroblock's quantization information independently, the patent inverts the approach by encoding the difference from the previous macroblock's value. This differential encoding approach, combined with run-length encoding, reduces the average number of bits per macroblock header, freeing up codes for image information
Data Source
AI summary
An image encoding apparatus includes a quantization control section that supplies a rounded Q-scale to a quantization section for quantizing each macroblock according to the rounded Q-scale. A Q-scale is generated for each macroblock based on the amount of generated codes produced by encoding. The Q-scale is rounded within a set range by a rounding control section. The rounded Q-scale is smaller in amount than the Q-scale. Since the same Q-scale values will continue in succession, these can be omitted from a stream, allowing the reduced amount of quantization information and the increased amount of information allocated to images.