Video Transcoding Using Motion-Adaptive Coefficient Discarding

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing video transcoding technologies face challenges in achieving high-quality output with low complexity and bit-rate while maintaining fast adaptation speed, often trading off between quality and speed or complexity, and suffer from issues like distortion and artefacts in high motion areas.

Innovation Solution

A method and apparatus that dynamically adjust the proportion of non-zero transform coefficients in video frames based on local motion activity, using a factor α to control bit-rate, which varies on a per block basis, ensuring even distribution of visual degradation and avoiding heavy blocking and drift artefacts.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If a cascade of decoder and encoder is used for transrating, then flexibility is improved, but computational complexity increases and speed decreases

Engineering Contradiction:
ImproveflexibilityVSAvoidcomputational complexity
Core Design Contradiction:
Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The patent extracts and eliminates the decoder component from the transrating architecture. Instead of decoding and re-encoding, the system directly processes the compressed video stream by selectively discarding transform coefficients, achieving the same bit-rate reduction without the computational overhead of full decoding and re-encoding operations

Inventive Principle:
Principle #2Taking out (Extraction)

Solution Approach 2:

Rather than following the conventional approach of decoding to uncompressed data and then re-compressing, the patent inverts the process by operating directly on the compressed transform coefficients in the frequency domain, eliminating the need for inverse transform and re-compression steps

Inventive Principle:
Principle #13The other way round (Inversion)

2Quantity of substance

If transform coefficients are discarded to reduce bit-rate, then bit-rate reduction is achieved, but visual quality deteriorates with distortion and artefacts

Engineering Contradiction:
Improvebit-rateVSAvoidvisual quality
Core Design Contradiction:
Quantity of substanceVSManufacturing precision

Solution Approach 1:

The patent applies different processing strategies to different regions of the video frame based on local motion activity. Blocks with high motion activity retain more transform coefficients to preserve quality, while blocks with low motion activity can afford greater coefficient discarding for bit-rate reduction, achieving an even distribution of visual degradation

Inventive Principle:
Principle #3Local quality

Solution Approach 2:

The system dynamically adjusts the number of transform coefficients to discard based on local motion activity in each block. The motion activity measure is calculated and used to adaptively determine the threshold for coefficient discarding, allowing the system to optimize the balance between bit-rate reduction and quality preservation in real-time

Inventive Principle:
Principle #15Dynamics

3Quantity of substance

If transrating is applied to all frames including I frames, then bit-rate reduction is maximized, but reference frame quality deteriorates causing drift artefacts

Engineering Contradiction:
Improvebit-rate reductionVSAvoidreference frame quality
Core Design Contradiction:
Quantity of substanceVSReliability

Solution Approach 1:

The patent segments the video frame processing based on frame type. I frames (intra-coded frames) are identified and excluded from transform coefficient discarding, while only P and B frames (inter-coded frames) undergo the transrating process. This segmentation preserves reference frame quality while still achieving bit-rate reduction on non-reference frames

Inventive Principle:
Principle #1Segmentation

4Manufacturing precision

If complex transrating techniques are used to maintain quality, then output quality is improved, but adaptation speed decreases

Engineering Contradiction:
Improveoutput qualityVSAvoidadaptation speed
Core Design Contradiction:
Manufacturing precisionVSSpeed

Solution Approach 1:

The patent replaces complex mechanical-style processing (full decoding, re-quantisation, re-transform) with a simpler computational approach that operates directly on the existing transform coefficients. This substitution eliminates multiple processing stages while maintaining quality through intelligent coefficient selection based on motion activity

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

Data Source

PatentEP1978743B1A method and apparatus for transcoding a video signal
Publication Date: 2020.07.01 VESTEL ELEKTRONIK SANAYI & TICARET ANONIM SIRKETI
  • EP1978743B1 patent drawingFigure 1~3
  • EP1978743B1 patent drawingFigure 4~5
  • EP1978743B1 patent drawingFigure 6~7

AI summary

The invention provides a method of transcoding a video signal made up of an input bit stream representative of frames of a video, each frame being made up of blocks of pixels, there being a corresponding block of data within the input bit stream for each block of pixels, the method comprising: for the bit stream of a frame of the video signal, identifying the type of frame; and for certain types of frame, disregarding a configurable proportion of the data in respect of plural blocks within the frame, thereby taking into account local motion activity within the frame.