Entropy Transcoder Circuit for Video Decoding Throughput
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional video decoding methods require multiple hardware accelerators or general-purpose processors for high-resolution video decoding, such as 4K@P60 or 8K@P60, which can be inefficient due to frame-level parallelization and inter-frame decoding dependencies, limiting throughput and scalability.
Innovation Solution
A circuit and method employing multiple entropy transcoders for frame-level parallelization and multi-layer wavefront parallelization, where entropy transcoders transform bit streams into unified wavefront sub-streams, and main decoding units perform decompression and pixel decoding in parallel, optimizing processing efficiency and reducing bandwidth requirements.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If frame-level parallelization is used with multiple hardware accelerators, then decoding throughput is improved, but device complexity and inter-frame decoding dependencies increase
Solution Approach 1:
The patent segments the video decoding process into two independent stages: entropy transcoding (processed by multiple hardware accelerators in parallel across frames) and main decoding (processed by a single decoder unit using wavefront parallelization within frames). This segmentation allows throughput improvement through frame-level parallelization while managing device complexity by assigning different processing strategies to different stages.
Solution Approach 2:
The patent introduces an entropy transcoder as an intermediary component that processes bitstreams before they reach the main decoding unit. The entropy transcoder performs entropy decoding and reorganization into wavefront sub-streams, acting as a mediator between the bitstream and the main decoding process. This intermediary enables efficient frame-level parallelization in the first stage while the second stage handles dependencies through wavefront parallelization.
2Adaptability or versatility
If multiple general-purpose processors are used for slice or tile parallelization, then decoding flexibility is improved, but processing efficiency and scalability for high resolution decrease
Solution Approach 1:
The patent replaces the mechanical approach of using multiple general-purpose processors (CPU, GPU, DSP) with a specialized hardware accelerator architecture. The hardware accelerator is dedicated to entropy transcoding operations, providing optimized performance for this specific function. This substitution of general-purpose processing with specialized hardware achieves higher processing efficiency and scalability for high-resolution video decoding while maintaining flexibility through the wavefront parallelization approach.
3Reliability
If wavefront parallelization is used within frames, then inter-frame dependencies are reduced, but processing time for each frame increases
Solution Approach 1:
The patent applies preliminary action by performing entropy transcoding and reorganization into wavefront sub-streams before the main decoding process. This preliminary processing stage (handled by hardware accelerators) prepares the data in advance, organizing it in a wavefront-friendly format. The main decoding unit then processes these pre-organized sub-streams using wavefront parallelization, which reduces inter-frame dependencies while the preliminary action ensures efficient processing time utilization.
Data Source
AI summary
A circuit for decoding a video, comprising: at least one entropy transcoder configured to perform at least one entropy transcoding on a received bit stream, to generate transcoded streams, wherein the entropy transcoding includes transcoding syntax for the bit stream; at least one main decoding unit configured to perform at least one transcoded stream decompression and pixel decodings on a frame of the transcoded streams.