Pixel Block Compression for HDR Textures
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current HDR texture compression schemes are inefficient due to costly hardware requirements for luminance-chrominance transforms and provide low quality at high bit rates, limiting their effectiveness in reducing bandwidth and maintaining image quality.
Innovation Solution
A method for compressing and decompressing pixel blocks with associated property values, involving lossless compression followed by quantization to meet target bit lengths, using a combination of prediction-based algorithms and quantization parameters to adjust quality while maintaining image integrity.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If luminance-chrominance transform is used for HDR texture compression, then compression efficiency is improved, but hardware complexity increases
Solution Approach 1:
The patent extracts and eliminates the luminance-chrominance transform step from the compression pipeline. Instead of converting HDR textures to YCrCb color space, the method operates directly on RGB values, removing the complex transform hardware while maintaining compression effectiveness through alternative techniques.
Solution Approach 2:
The patent changes the color space representation parameter from YCrCb (luminance-chrominance) to RGB, and modifies the compression approach from transform-based to prediction-based with run-length encoding, achieving compression without the costly luminance-chrominance transform.
2Manufacturing precision
If lossless compression is used for HDR textures, then image quality is maintained, but bandwidth reduction is insufficient
Solution Approach 1:
The patent introduces a dynamic compression approach that adapts to the specific characteristics of each texture block. By using prediction-based compression with run-length encoding, the system dynamically adjusts compression levels based on local patterns, achieving variable compression ratios that maintain quality where needed while reducing bandwidth elsewhere.
Solution Approach 2:
The patent changes the compression methodology parameter from traditional lossless schemes to a hybrid approach combining prediction, run-length encoding, and selective quantization, enabling flexible bandwidth reduction while preserving image quality through intelligent parameter selection.
3Quantity of substance
If quantization is applied to reduce bit length, then bandwidth is reduced, but image quality deteriorates
Solution Approach 1:
The patent applies quantization selectively and locally rather than uniformly across the entire texture. By using prediction-based compression first, the method identifies regions where quantization has minimal visual impact and applies it there, while preserving higher precision in important regions, thus reducing bandwidth without significant quality loss.
Solution Approach 2:
The patent implements dynamic quantization where the quantization level is adjusted based on local image characteristics and compression requirements. The system dynamically selects quantization parameters to achieve the desired bandwidth reduction while maintaining acceptable image quality through adaptive parameter selection.
Data Source
AI summary
A pixel block (300) is losslessly compressed into a candidate compressed block. If the bit length of the candidate block exceeds a threshold value, the property values of the pixels (310-317) in the block (300) are quantized and the quantized values are losslessly compressed into a new candidate compressed block. The procedure is repeated until a candidate compressed block having good image quality and a bit length below the threshold value is found. A compressed representation (400) of the block (300) is determined based on the found candidate compressed block (420) and an identifier (410) of the used quantization parameter.