Encoder Flicker Reduction via Frame-Adaptive Quantization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Flicker in image frames, particularly in background regions, is not effectively addressed by existing encoders, which can lead to objectionable visual disturbances, especially when the image information diverges significantly between successive frames.
Innovation Solution
An encoder system that adapts quantization coefficients using information from both the current and previous frames, employing a transformer to generate frequency domain coefficients, a quantizer to adjust these coefficients, and a coefficient calculator to minimize the absolute difference between inverse-quantized and frequency domain coefficients, thereby reducing flicker.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If quantization coefficients are adjusted using only current frame information, then encoding efficiency is improved, but flicker reduction is insufficient
Solution Approach 1:
The encoder performs preliminary action by calculating adjusting coefficients based on both current and previous frame information before final quantization. The coefficient calculator computes adjusting coefficients using motion compensation and previous frame data, then applies these coefficients to reduce flicker in the quantization process, ensuring stable visual output before encoding completion.
Solution Approach 2:
The system implements feedback by utilizing previous frame information to adjust current frame quantization coefficients. The encoder feeds back motion vector data and previous frame coefficient information to the coefficient calculator, which then modifies adjusting coefficients to compensate for flicker, creating a closed-loop system that improves both encoding efficiency and visual stability.
2Reliability
If adjusting coefficients are calculated using both current and previous frames, then flicker reduction is improved, but device complexity increases
Solution Approach 1:
The encoder applies local quality by calculating adjusting coefficients for each block individually based on its specific motion characteristics and previous frame data. Rather than applying a global adjustment, the system computes block-specific coefficients using motion compensation, allowing precise flicker reduction in each region while maintaining overall encoder efficiency and avoiding unnecessary complexity.
Solution Approach 2:
The system changes parameters dynamically by adjusting quantization coefficients based on motion vector magnitude and direction. The coefficient calculator modifies adjusting coefficients according to motion intensity and frame differences, adapting the quantization process to local motion characteristics without requiring complex structural changes to the encoder architecture.
3Reliability
If motion compensation is applied for adjusting coefficients, then flicker reduction in moving regions is improved, but processing time increases
Solution Approach 1:
The encoder applies partial action by performing motion compensation and adjusting coefficient calculation only for blocks where motion is detected. The system evaluates motion vector magnitude and applies flicker reduction processing selectively to moving regions rather than processing the entire frame, reducing overall processing time while maintaining effective flicker reduction in areas where it is needed.
Data Source
AI summary
In an application processor having an encoder configured to reduce flicker occurring in a display, the encoder includes a transformer configured to generate first frequency domain coefficients for individual block values of a current frame, a quantizer configured to quantize the first frequency domain coefficients to generate quantized coefficients, and a coefficient calculator configured to generate first adjusting coefficients for reducing the flicker, where the quantizer is configured to adjust the quantization coefficients based on the first adjusting coefficients and output adjusted quantization coefficients.


