Non-linear Video Encoding Brightness Compensation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Lossy digital video encoding in a non-linear color space causes unnatural shifts in image brightness, particularly at lower frequencies, due to changes in high frequency content during compression, which are easily detectable even at distances further than ideal viewing distances.
Innovation Solution
The proposed solution involves transforming the video signal into the frequency domain, analyzing and correcting for quantization errors in the non-linear domain, and then re-doing the compression to mitigate these artifacts, by estimating and applying corrections to the signal before further quantization and encoding, thereby maintaining the benefits of non-linear compression while preventing brightness shifts.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If lossy compression is applied in a non-linear color space, then encoding efficiency is improved, but brightness stability deteriorates
Solution Approach 1:
The patent applies preliminary action by performing brightness compensation before the final encoding step. The system first encodes the video signal in a non-linear color space to achieve compression efficiency, then calculates the quantization errors, transforms them to the spatial domain, and compensates for the brightness shifts before re-encoding. This preliminary compensation action prevents the brightness instability from manifesting in the final output while preserving the encoding efficiency benefits of non-linear color space compression.
Solution Approach 2:
The patent uses an intermediary approach by introducing a compensation signal as a mediator between the compressed video signal and the final output. The system calculates quantization errors from the non-linear encoding, transforms these errors to the spatial domain, and adds a compensated signal to correct the brightness shifts. This intermediary compensation mechanism allows the system to maintain both the encoding efficiency of non-linear color space and the brightness stability required for quality output.
2Productivity
If compression level is increased, then encoding efficiency is improved, but artifact visibility worsens
Solution Approach 1:
The patent applies the blessing in disguise principle by converting the harmful quantization errors into a beneficial compensation signal. Instead of simply discarding the quantization errors as waste, the system calculates these errors, transforms them to the spatial domain, and uses them to generate a compensation signal that corrects the brightness artifacts. This approach transforms the harmful effect of quantization into a beneficial artifact reduction mechanism, allowing higher compression levels to be used without proportionally increasing artifact visibility.
3Productivity
If high frequency content is compressed, then encoding efficiency is improved, but low frequency brightness stability worsens
Solution Approach 1:
The patent applies segmentation by separating the frequency domain into high frequency and low frequency components. The system performs DCT transformation to divide the video signal into different frequency bands, applies quantization primarily to high frequency components for compression efficiency, then transforms the quantization errors back to the spatial domain to compensate for low frequency brightness shifts. This segmentation allows independent optimization of compression for high frequencies while protecting low frequency brightness stability.
Data Source
AI summary
A method of encoding a non-linear, color space video signal includes converting the non-linear video signal to a linear video signal, transforming the non-linear signal to produce a transformed non-linear signal, transforming the linear signal to product a transformed linear signal, using the transformed linear signal to determine errors, applying the errors to the non-linear, color space video signal to produce a compensated non-linear color space signal, and encoding the compensated signal.