HEVC Video Decoding Parallelism Negotiation

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

High Efficiency Video Coding (HEVC) faces challenges in optimizing parallel encoding/decoding processes due to conflicting demands between parallelization and Maximum Transmission Unit (MTU) size matching, leading to inefficiencies in bitstream partitioning and inter-processor communication.

Innovation Solution

A method and apparatus are introduced to negotiate and configure HEVC profile and level settings between encoders and decoders, using a parameter (Y) to identify and utilize optimal levels of parallelism, ensuring efficient decoding and encoding by matching encoder and decoder capabilities, thereby improving video quality and processing efficiency.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If regular slices are used for parallelization in HEVC, then parallel decoding can be achieved, but coding overhead increases due to slice headers and lack of prediction across slice boundaries

Engineering Contradiction:
Improveparallel decoding capabilityVSAvoidcoding overhead
Core Design Contradiction:
ProductivityVSLoss of information

Solution Approach 1:

The patent extracts the parallelization functionality from the regular slice structure by introducing wavefront parallel processing (WPP) as a separate, dedicated mechanism. This allows parallel decoding to be achieved without the overhead of multiple regular slice headers, as WPP uses a single slice structure with internal row-based partitioning that enables parallel processing while maintaining prediction efficiency.

Inventive Principle:
Principle #2Taking out (Extraction)

2Adaptability or versatility

If regular slices are used for bitstream partitioning to match MTU size, then transmission requirements are met, but parallelization efficiency is reduced due to conflicting slice layout demands

Engineering Contradiction:
ImproveMTU size matching capabilityVSAvoidparallelization efficiency
Core Design Contradiction:
Adaptability or versatilityVSProductivity

Solution Approach 1:

The patent segments the picture into multiple rows of treeblocks for WPP processing, allowing independent parallel decoding of each row while maintaining a single continuous bitstream. This segmentation enables both MTU size matching through controlled row boundaries and high parallelization efficiency, as each row can be processed independently without the overhead of multiple slice headers.

Inventive Principle:
Principle #1Segmentation

3Productivity

If wavefront parallel processing is used, then parallel decoding efficiency improves, but inter-processor communication requirements increase substantially

Engineering Contradiction:
Improveparallel decoding efficiencyVSAvoidinter-processor communication overhead
Core Design Contradiction:
ProductivityVSDevice complexity

Solution Approach 1:

The patent applies preliminary action by pre-calculating and storing the wavefront offset values in the bitstream during encoding. This allows the decoder to determine the exact data dependencies between rows in advance, enabling efficient inter-processor communication by having each processing element know exactly what data it needs from neighboring rows before starting decoding, thus reducing communication overhead while maintaining high parallelization efficiency.

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentUS11570456B2Apparatus and methods thereof for video processing
Publication Date: 2023.01.31 TELEFONAKTIEBOLAGET LM ERICSSON (PUBL)
  • US11570456B2 patent drawing
  • US11570456B2 patent drawing
  • US11570456B2 patent drawing

AI summary

A method to be performed by a receiving apparatus for decoding an encoded bitstream representing a sequence of pictures of a video stream is provided. In the method, capabilities relating to level of decoding parallelism for the decoder are identified, a parameter indicative of the decoder's capabilities relating to level of decoding parallelism is kept, and for a set of levels of decoding parallelism, information relating to HEVC profile and HEVC level that the decoder is capable of decoding is kept.A method for encoding a bitstream representing a sequence of pictures of a video stream is also provided. In the method, a parameter is received from a transmitting apparatus that should decode the encoded bitstream.