Smart Chunking for Video Encoding Quality

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Conventional video encoding techniques face inefficiencies due to fixed chunk sizes, leading to quality degradation at segment borders, sub-optimal bitrate usage, and increased computational costs, as they struggle to balance quality, bitrate, and computational cost effectively.

Innovation Solution

Smart chunking dynamically determines internal chunk lengths based on video input characteristics and quality requirements, decoupling chunk duration from output segment duration, allowing for variable chunk sizes and bitrate distribution to maintain consistent quality across segments.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If video encoding uses fixed chunk sizes matching output segment duration, then parallel encoding can be simplified and distributed across multiple machines, but quality degrades at segment borders and bitrate usage becomes sub-optimal

Engineering Contradiction:

Improveencoding throughputVSAvoidvideo quality consistency

Core Design Contradiction:

ProductivityVSManufacturing precision

Solution Approach 1:

The patent divides the video encoding process into two independent layers: encoding chunks (for parallel processing) and output segments (for streaming delivery). By decoupling these layers, the system can use larger encoding chunks to maintain quality while still producing appropriately-sized output segments, resolving the contradiction between encoding efficiency and quality consistency.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent introduces dynamic chunk duration that is not fixed but adapts based on the relationship between encoding requirements and output segment duration. The chunk duration becomes a flexible parameter that can be adjusted to optimize both parallel encoding efficiency and quality consistency, rather than being rigidly tied to output segment boundaries.

Inventive Principle:

Principle #15Dynamics

2Manufacturing precision

If variable bitrate mode is used to maintain quality at chunk endings, then quality consistency improves, but computational cost increases

Engineering Contradiction:

Improvevideo quality consistencyVSAvoidcomputational cost

Core Design Contradiction:

Manufacturing precisionVSUse of energy by moving object

Solution Approach 1:

The patent extracts the quality consistency problem from the rate-control algorithm by using larger encoding chunks that span multiple output segments. This removes the need for complex variable bitrate adjustments at every chunk boundary, as the quality can be maintained more efficiently through proper chunk segmentation rather than continuous bitrate modulation.

Inventive Principle:

Principle #2Taking out (Extraction)

3Device complexity

If target bitrate is applied uniformly to all chunks, then encoding process is simplified, but high complexity segments suffer quality loss and low complexity segments waste bitrate

Engineering Contradiction:

Improveencoding process complexityVSAvoidbitrate allocation efficiency

Core Design Contradiction:

Device complexityVSManufacturing precision

Solution Approach 1:

The patent applies different target bitrates to different encoding chunks based on their complexity characteristics. By analyzing video complexity at the chunk level and assigning appropriate bitrates locally, the system optimizes quality for both high and low complexity segments without requiring complex real-time adjustments during encoding.

Inventive Principle:

Principle #3Local quality

4Ease of manufacture

If chunk duration is strictly connected to output segment duration, then segment encoding can be straightforwardly implemented, but overhead from downloading, demuxing, decoding, filtering, and muxing increases for long videos

Engineering Contradiction:

Improvesegment encoding simplicityVSAvoidprocessing overhead time

Core Design Contradiction:

Ease of manufactureVSLoss of time

Solution Approach 1:

The patent merges multiple output segments into a single encoding chunk for processing. This reduces the number of times the video needs to be downloaded, demuxed, decoded, filtered, and muxed, thereby significantly reducing processing overhead for long videos while maintaining the ability to produce appropriately-sized output segments for streaming.

Inventive Principle:

Principle #5Merging (Combining)

Data Source

PatentUS20240291983A1Smart Chunking for Video Encoding

Publication Date: 2024.08.29 BITMOVIN

AI summary

Techniques for video encoding are described herein. A method for video encoding with smart chunking includes receiving, by a distributed video encoding system, a video input and a target bitrate, the video input having segments of a segment duration, determining an internal chunk length that is a multiple of the segment duration, encoding chunks having the internal chunk length, wherein the average bitrate across the chunk is equal to the target bitrate, and segmenting the encoded chunks into encoded segments of the segment duration. The distributed video encoding system may include various video encoders, or encoder instances, able to encode multiple chunks in parallel. The encoded segments may be output to a client, all of the encoded segments being of equal or similar quality.