Video Encoder Segmentation for Limited Cache Memory

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Video encoding devices face challenges in efficiently encoding large digital video data due to the limited size of reference cache memories, which are expensive and power-intensive to manufacture, and cannot match the increasing pixel counts of source images.

Innovation Solution

A method that divides digital source images into smaller parts to fit within a cache memory of limited width, using skip blocks and INTRA/INTER-frame encoding to ensure compatibility with existing standards while optimizing cache usage, allowing for efficient encoding of large video images with smaller cache memories.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If the cache memory width is increased to match the source image width for efficient encoding, then the encoding speed and efficiency improve, but the manufacturing cost and power consumption increase significantly

Engineering Contradiction:

Improveencoding speedVSAvoidmanufacturing cost

Core Design Contradiction:

ProductivityVSEase of manufacture

Solution Approach 1:

The source image is divided into multiple segments (first part and second part) that can be processed separately using a smaller cache memory. Each segment fits within the cache width, allowing the encoding process to proceed efficiently without requiring a large cache memory matching the full source image width.

Inventive Principle:

Principle #1Segmentation

2Productivity

If the cache memory width is increased to match the source image width, then the encoding efficiency improves, but the power consumption increases

Engineering Contradiction:

Improveencoding efficiencyVSAvoidpower consumption

Core Design Contradiction:

ProductivityVSUse of energy by stationary object

Solution Approach 1:

The source image is divided into multiple segments that fit within the smaller cache memory width. This segmentation allows the encoding process to maintain high efficiency by keeping all necessary data for each segment within the cache, avoiding the need for a power-intensive large cache memory.

Inventive Principle:

Principle #1Segmentation

3Productivity

If the cache memory width is increased to match the source image width, then the encoding performance improves, but the device complexity increases

Engineering Contradiction:

Improveencoding performanceVSAvoiddevice complexity

Core Design Contradiction:

ProductivityVSDevice complexity

Solution Approach 1:

The source image is segmented into multiple parts that can be processed sequentially using a smaller cache memory. This approach maintains encoding performance by ensuring each segment fits in cache, while avoiding the complexity of designing and fabricating a large cache memory structure.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The encoding process dynamically switches between processing different segments of the source image. The system adapts its operation to work with the smaller cache memory by processing one segment at a time, thereby achieving high performance without the need for a statically large cache memory.

Inventive Principle:

Principle #15Dynamics

Data Source

PatentUS10165290B2Method for encoding digital video data

Publication Date: 2018.12.25 AXIS

AI summary

The present invention relates to a method for encoding digital video data corresponding to a sequence of digital source images using a cache memory, each of the digital source images having an equal source image width corresponding to a first number of blocks, the cache memory having a cache width corresponding to a second number of blocks, wherein the second number of blocks is smaller than the first number of blocks.