Parallel Video Encoder Context Transfer for Single Slice Deblocking

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Conventional video encoders are limited in their ability to perform whole picture deblocking and compression as a single slice, leading to inefficiencies in bit usage, motion vector prediction, and visible seams across slice boundaries.

Innovation Solution

A parallel video encoding apparatus comprising a segmentor circuit, multiple encoders, and a multiplexer circuit that allows for parallel encoding of picture strips, enabling whole picture deblocking and compression as a single slice, reducing bit usage, and eliminating slice borders by transferring context information between encoders.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If multiple encoder chips are used for parallel encoding, then encoding speed is improved, but slice boundaries are created that prevent whole picture deblocking and context transfer

Engineering Contradiction:

Improveencoding speedVSAvoidslice boundary management

Core Design Contradiction:

ProductivityVSDevice complexity

Solution Approach 1:

The picture is divided into multiple strips that can be processed in parallel by different encoder chips, while maintaining the ability to perform whole-picture operations through context transfer mechanisms between chips

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

Context information is transferred between encoder chips through defined interfaces, enabling whole picture deblocking and compression as a single slice despite physical segmentation across multiple chips

Inventive Principle:

Principle #24Intermediary (Mediator)

2Productivity

If slices are divided across chip boundaries, then parallel processing capability is improved, but motion vector prediction and statistical context between macroblock rows cannot be used across boundaries

Engineering Contradiction:

Improveparallel processing capabilityVSAvoidcontext information loss

Core Design Contradiction:

ProductivityVSLoss of information

Solution Approach 1:

Context information including motion vectors and statistical data is transferred across chip boundaries through standardized interfaces, preventing information loss while maintaining parallel processing benefits

Inventive Principle:

Principle #24Intermediary (Mediator)

Solution Approach 2:

The encoder chip interfaces are designed to handle multiple functions including context transfer, deblocking operations, and compression across slice boundaries, making the system versatile enough to maintain whole-picture processing capabilities

Inventive Principle:

Principle #6Universality (Multi-functionality)

3Manufacturing precision

If deblocking is performed across slice boundaries, then picture quality is improved, but additional processing complexity is required

Engineering Contradiction:

Improvepicture qualityVSAvoiddeblocking processing complexity

Core Design Contradiction:

Manufacturing precisionVSDevice complexity

Solution Approach 1:

The interface between encoder chips provides standardized mechanisms for transferring deblocking context and parameters, enabling cross-slice deblocking without requiring complex custom integration logic

Inventive Principle:

Principle #24Intermediary (Mediator)

Data Source

PatentUS7630565B2Parallel video encoder with whole picture deblocking and/or whole picture compressed as a single slice

Publication Date: 2009.12.08 AVAGO TECHNOLOGIES INTERNATIONAL SALES PTE LTD

AI summary

An apparatus comprising (i) a segmentor circuit, (ii) a plurality of encoders and (iii) a multiplexer circuit. The segmentor circuit may be configured to segment a picture into a plurality of strips. Each of the plurality of encoders may be configured (i) to encode respective strips of the plurality of strips in parallel and (ii) to transfer context information with at least one other encoder of the plurality of encoders. The multiplexer circuit may be configured to combine the encoded strips from the plurality of encoders into an encoded bit stream.