Video Coding Noise Level Control via Rate-Distortion Optimization

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Conventional video coding techniques fail to generate coded video data that appears natural, as they do not effectively account for noise levels across different frames, leading to suboptimal perceived quality during decoding.

Innovation Solution

The proposed solution involves estimating noise associated with candidate coding modes and selecting a coding mode based on these estimates to maintain consistent noise levels across frames, using a cost modeling technique that incorporates noise estimation into the rate-distortion optimization process.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Loss of energy

If conventional rate-distortion optimization is used to select coding modes, then bandwidth compression is achieved, but the decoded video does not appear natural due to inconsistent noise levels

Engineering Contradiction:

Improvebandwidth compressionVSAvoidnoise level consistency

Core Design Contradiction:

Loss of energyVSObject-affected harmful factors

Solution Approach 1:

The patent modifies the rate-distortion cost model by introducing noise level estimation as an additional parameter. Instead of using only distortion and rate, the system estimates noise levels for each candidate coding mode and incorporates this into the cost function, allowing the optimizer to select modes that maintain consistent noise characteristics while achieving compression

Inventive Principle:

Principle #35Parameter changes

Solution Approach 2:

The patent replaces the conventional mechanical rate-distortion optimization process with an enhanced version that includes noise estimation. The system substitutes traditional distortion metrics with a composite cost function that factors in noise level predictions, enabling more natural-looking decoded video

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

2Measurement precision

If multiple candidate coding modes are evaluated using conventional cost modeling, then rate-distortion optimization is achieved, but noise levels vary across frames leading to poor perceived quality

Engineering Contradiction:

Improverate-distortion estimationVSAvoidnoise level consistency

Core Design Contradiction:

Measurement precisionVSManufacturing precision

Solution Approach 1:

The patent implements feedback by estimating noise levels for each candidate coding mode and using these estimates to guide the mode selection process. The noise level estimation acts as a feedback mechanism that informs the cost function, allowing the system to adjust its selection criteria to maintain consistent noise characteristics across different frames

Inventive Principle:

Principle #23Feedback

Solution Approach 2:

The patent performs preliminary noise level estimation for each candidate coding mode before final mode selection. By evaluating noise characteristics in advance and incorporating them into the cost function, the system can predict which modes will produce consistent noise levels, enabling better optimization of both rate and distortion

Inventive Principle:

Principle #10Preliminary action

Data Source

PatentUS10574997B2Noise level control in video coding

Publication Date: 2020.02.25 APPLE INC

AI summary

Systems and methods are disclosed for coding pixel blocks of an input frame in which coding costs of a plurality of candidate coding modes are estimated. A coding cost of a candidate coding mode may be estimated based on noise estimate associated with the candidate coding mode. A coding mode for the input pixel block may be selected based on a comparison of the estimated coding costs of the plurality of candidate coding modes. The input pixel block may then be coded according to the selected coding mode.