Unified Multi-Bitrate Media Encoding Parameter Optimization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing media encoding methods face inefficiencies in producing multiple bitrate and spatial resolution streams, particularly due to high computational complexity and suboptimal reuse of information from high-resolution input streams, leading to increased costs and subpar rate-distortion performance.
Innovation Solution
The proposed system employs an intelligent encoding module that identifies optimal combinations of encoding parameters for each coding tree block across multiple output bitrates and spatial resolutions, coupled with normative encoding modules to efficiently encode media streams, reducing computational complexity through cost function evaluation and parameter reuse.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If multiple bitrate streams are encoded independently using traditional methods, then each stream achieves optimal rate-distortion performance, but computational complexity increases linearly with the number of output bitrates
Solution Approach 1:
The patent merges multiple independent encoding processes into a single unified encoding pass. By combining the encoding of multiple bitrate streams into one process that processes all target bitrates simultaneously, the system achieves optimal rate-distortion performance for each stream while avoiding the linear increase in computational complexity that would result from encoding each stream separately.
Solution Approach 2:
The encoding system is designed to perform multiple functions simultaneously - encoding the same input media into multiple different bitrate streams in a single processing operation. This multi-functional approach allows the system to produce multiple output streams with different bitrates while maintaining optimal compression for each, without requiring separate encoding passes for each bitrate.
2Manufacturing precision
If high-resolution input media is used for encoding lower resolution streams, then source quality is preserved, but information from high-resolution input is not effectively utilized and encoding efficiency decreases
Solution Approach 1:
The patent applies local quality by processing different regions of the image at different resolutions simultaneously within a single encoding pass. The system identifies and processes only the necessary portions of the high-resolution input media for each target resolution, effectively utilizing the available information from the high-resolution source while maintaining encoding efficiency by avoiding unnecessary processing of redundant high-resolution data for lower resolution outputs.
3Productivity
If coding tools from newer standards like HEVC are used, then encoding efficiency improves, but computational complexity increases significantly compared to older standards
Solution Approach 1:
The patent changes the fundamental parameter of how encoding is performed by transitioning from sequential independent encoding of multiple bitrates to a simultaneous unified encoding process. This parameter change enables the system to leverage advanced coding tools from newer standards like HEVC to improve overall encoding efficiency, while the unified approach prevents the computational complexity from increasing proportionally to the number of output bitrates.
Data Source
AI summary
In one example, a sequence of pictures may be transformed at a given spatial resolution to a plurality of output spatial resolutions and/or an additional spatial resolution that is lower than the plurality of output spatial resolutions. The sequence of pictures and a picture type may be received for each of the pictures at respective output spatial resolutions. Estimating a set of combinations of encoding parameters for coding tree blocks (CTBs) in each of the pictures at a lowest output spatial resolution or mapping combinations of encoding parameters for each CTB in each of the pictures at an immediate coarser spatial resolution to corresponding blocks at the respective output spatial resolution is performed. A set of combinations of encoding parameters may be determined for each CTB and best combinations of encoding parameters from the determined set of combinations may be identified for encoding the sequence of pictures.