Entropy Coding for Display Stream Compression Throughput
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional video compression techniques, such as DSC, face challenges in achieving high throughput for high-resolution displays, as existing entropy coding methods like Huffman, Arithmetic, and Exponential-Golomb codes have limited throughput, making visually lossless coding economically infeasible for high-resolution applications.
Innovation Solution
The proposed solution involves entropy coding techniques that provide higher throughput, such as combining skip and DSU-VLC coding, predicting group sizes based on previous groups, and coding both positive and negative differences, allowing for efficient encoding and decoding of video data at 4 samples/clock.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If conventional entropy coding methods (Huffman, Arithmetic, Exponential-Golomb) are used for video compression, then compression is achieved, but throughput is limited and economically infeasible for high-resolution displays
Solution Approach 1:
The patent segments the video data processing into distinct stages: prediction, residual calculation, and entropy coding. By dividing the block into 2x2 sub-blocks and processing them independently with parallel entropy coders, the system achieves higher throughput while maintaining compression effectiveness, making high-resolution display compression economically feasible
Solution Approach 2:
The patent changes the entropy coding parameters by using simplified coding tables and adaptive coding modes (skipped, non-skipped, partially-skipped) based on residual characteristics. This parameter adaptation allows the system to achieve high throughput by selecting appropriate coding strategies for different data patterns, improving economic feasibility for high-resolution applications
2Quantity of substance
If image compression is applied to pixel data, then bandwidth requirements are reduced, but implementation becomes difficult and expensive in conventional display devices
Solution Approach 1:
The patent implements self-service by using the display device's own existing hardware resources (pixel pipelines, memory structures) to perform compression operations. The compression is integrated into the display controller, allowing the device to compress its own output without requiring separate complex compression hardware, thereby reducing implementation complexity while lowering bandwidth requirements
Solution Approach 2:
The patent makes the display controller universal by enabling it to perform both traditional display functions and video compression functions using the same hardware resources. The pixel processing pipelines are utilized for both rendering and compression, eliminating the need for dedicated compression hardware and reducing overall device complexity
3Area of moving object
If display link bandwidth is reduced to support high-resolution displays, then high-resolution output is achieved, but compression must be visually lossless which increases processing requirements
Solution Approach 1:
The patent applies preliminary action by performing prediction and residual calculation before entropy coding. By predicting pixel values from neighboring pixels and encoding only the residuals, the system reduces the data volume early in the processing pipeline, allowing visually lossless compression of high-resolution displays to be achieved at manageable throughput levels
Solution Approach 2:
The patent introduces dynamics by using adaptive coding modes that change based on the residual data characteristics. The system dynamically selects between skipped, non-skipped, and partially-skipped coding modes for different blocks, and uses adaptive quantization parameters, allowing the processing throughput to efficiently adapt to varying image content while maintaining visually lossless quality for high-resolution displays
Data Source
AI summary
Entropy coding techniques for display stream compression (DSC) are disclosed. In one aspect, a method of entropy coding video data includes partitioning a block of the video data into a plurality of groups of samples based at least in part on a coding mode of the block. The method further includes entropy coding the block via performing a group-wise skip on at least one of the groups in response to all of the samples in the at least one group being equal to a predetermined value.