Video Decoder Inverse Transforms with Signaled Bit-Depth Shifts
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The complexity of implementing inverse transforms in video codecs is exacerbated by large bit-widths of multiply operations and memory accesses, particularly when coding video data with higher bit-depths, leading to increased processing challenges.
Innovation Solution
Incorporating syntax elements in the bitstream to signal dequantization and mid-transform shifts, allowing for inverse transform implementation at a lower bit-depth than the signaled coefficients, thereby reducing memory and processing requirements.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If inverse transform is implemented at high bit-depth to maintain precision, then measurement precision is improved, but device complexity increases due to larger bit-width multiply operations and memory accesses
Solution Approach 1:
The inverse transform process is segmented into two distinct stages: a first inverse transform performed at a lower bit-depth (reducing complexity) and a second inverse transform performed at a higher bit-depth (maintaining precision). This segmentation allows the system to balance between processing complexity and measurement precision by performing critical transformations at lower precision levels and only applying high precision where necessary.
Solution Approach 2:
The patent dynamically adjusts the bit-depth used for inverse transform operations based on the specific requirements of different transform coefficients. By making the bit-depth flexible rather than fixed, the system can optimize between complexity and precision for each coefficient, applying higher bit-depth only when the transform coefficients require it, rather than consistently using high bit-depth throughout the entire process.
2Adaptability or versatility
If dynamic shift value determination is implemented without explicit signaling, then adaptability is improved, but device complexity increases due to buffer storage requirements at maximum possible bit-depth
Solution Approach 1:
The syntax element is positioned in the bitstream before the transform coefficients, allowing the shift values to be determined in advance during the parsing process. This preliminary action enables the decoder to prepare the necessary shift values before processing the actual transform coefficients, eliminating the need for complex dynamic determination and large buffers during the transform operation itself.
Solution Approach 2:
Instead of dynamically calculating shift values during the transform process, the patent uses syntax elements that explicitly signal these values in the bitstream. This approach creates a simplified copy of the necessary information (shift values) that can be directly used during decoding, avoiding the need for complex runtime calculations and large buffer storage requirements.
3Manufacturing precision
If multiple steps are implemented to address multiplier bit-width issues, then manufacturing precision is improved, but device complexity increases due to additional processing steps
Solution Approach 1:
The transform process is divided into multiple stages with different precision requirements. The first inverse transform operates at a lower bit-depth with simpler operations, while the second inverse transform operates at a higher bit-depth for enhanced precision. This segmentation allows the system to achieve manufacturing precision without requiring all operations to run at maximum precision simultaneously, thereby reducing overall device complexity.
Solution Approach 2:
The patent changes the bit-depth parameter dynamically between different transform operations. By adjusting the bit-depth parameter from lower to higher levels at appropriate stages, the system achieves the necessary manufacturing precision for final output while avoiding the complexity of maintaining high precision throughout the entire processing pipeline.
Data Source
AI summary
A device for decoding video data receives encoded video data encoded at a first bit-depth; receives in an encoded video bitstream for the video data one or more syntax elements; determines a dequantization shift and a mid-transform shift based on the one or more syntax elements; and performs an inverse transform of the encoded video data at a second bit-depth using the dequantization shift and the mid-transform shift to determine residual values for a block of video data, wherein the second bit-depth is different than the first bit-depth.