Video Transcoding Using Motion-Adaptive Coefficient Discarding
Find Innovative SolutionsGenerate 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
Engineering 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
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
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
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
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
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
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
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
4Manufacturing precision
If complex transrating techniques are used to maintain quality, then output quality is improved, but adaptation speed decreases
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
Data Source
Figure 1~3
Figure 4~5
Figure 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.