Video Quantizer Design With Exponential Q_Index-to-Qstep Mapping
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The quantization step sizes in AV1 have limited resolution, which affects the granularity of bit rates achievable by the codec, particularly for 10-bit and 12-bit internal bitdepth.
Innovation Solution
Implementing an exponential mapping of Q_index to Qstep, which increases the range and granularity of quantization step sizes, allowing for more precise bit rate control and improved video coding efficiency in AV2.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If linear mapping is used from Q_index to Qstep, then the quantization step sizes have limited resolution, but the codec complexity remains low
Solution Approach 1:
The patent changes the mapping function from linear to exponential, transforming the relationship between Q_index and Qstep. This parameter change in the mapping function enables finer granularity in quantization step sizes, particularly for higher bit depths (10-bit and 12-bit), thereby improving measurement precision without significantly increasing system complexity
Solution Approach 2:
The patent introduces dynamic adjustment capabilities by allowing different exponential mapping parameters (alpha and beta) to be configured based on bit depth requirements. This enables the system to adaptively optimize quantization step size resolution for different video quality requirements and bit depth configurations
2Manufacturing precision
If exponential mapping is implemented to increase quantization step size granularity, then bit rate control precision improves, but computational complexity increases
Solution Approach 1:
The patent applies exponential mapping with configurable parameters (alpha and beta) to transform the Q_index to Qstep relationship. This enables precise control over quantization step sizes and consequently bit rate granularity, allowing the encoder to achieve finer bit rate control while managing computational complexity through parameter optimization
Solution Approach 2:
The patent pre-calculates and stores exponential mapping tables for different bit depth configurations. By preparing these mapping relationships in advance, the system reduces real-time computational complexity during encoding while maintaining the precision benefits of exponential mapping for bit rate control
3Adaptability or versatility
If separate look up tables are used for different bit depths, then bit depth-specific optimization is achieved, but memory requirements increase
Solution Approach 1:
The patent creates a unified exponential mapping framework that serves multiple bit depths (8-bit, 10-bit, 12-bit) through a single mathematical model. By using universal mapping parameters (alpha and beta) that can be configured for different bit depths, the system achieves bit depth adaptability without requiring separate look up tables for each bit depth, thereby reducing memory requirements
Solution Approach 2:
The patent enables the same exponential mapping structure to adapt to different bit depths by changing the parameters (alpha and beta) rather than changing the entire data structure. This parameter-based adaptation allows the system to support multiple bit depths with a single look up table or even computed on-the-fly, significantly reducing memory storage requirements
Data Source
AI summary
A method, computer program, and computer system for video coding is provided. Video data including one or more quantized coefficients is received. One or more index values associated with the quantized coefficients are mapped to one or more step values based on an exponential mapping. The video data is decoded based on the one or more step values.