Local Bit Rate Curve for Streaming Adaptation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing adaptive bit rate (ABR) streaming systems rely on peak and average bit rates to switch between different bit rate tiers, which can lead to suboptimal choices due to brief high-motion scenes or varying bit rate requirements throughout a program.
Innovation Solution
A method and apparatus that use a local bit rate curve, determined from a window of media segments, to select the appropriate bit rate tier for streaming content, allowing for more precise matching of network speed requirements to the bit rate demands of the content.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If peak and average bit rate are used to select bit rate tiers, then the selection process is simple, but the accuracy of bit rate matching to network conditions and content requirements deteriorates
Solution Approach 1:
The patent segments the bit rate analysis into multiple dimensions: peak bit rate, average bit rate, and percentile-based bit rates (e.g., 90th percentile). This segmentation allows the system to capture different aspects of bit rate requirements - peak for maximum quality potential, average for sustained performance, and percentiles for typical experience - thereby improving matching accuracy while maintaining selection simplicity through automated multi-criteria evaluation
Solution Approach 2:
The patent introduces multiple bit rate parameters (peak, average, and various percentiles) to characterize content requirements. By changing from a single parameter (peak or average) to multiple parameters, the system achieves more precise matching between network conditions and content demands, resolving the contradiction between simple selection and accurate matching
2Ease of operation
If only peak bit rate is used for tier selection, then the selection is straightforward, but the system cannot adapt to varying bit rate requirements throughout different scenes
Solution Approach 1:
The patent implements dynamic bit rate tier selection by continuously monitoring multiple bit rate parameters and adapting selections based on current network conditions and content characteristics. Instead of a static peak-based approach, the system dynamically adjusts tier selections using percentile-based metrics that reflect actual usage patterns across different scenes, thereby achieving both simplicity and adaptability
3Reliability
If only average bit rate is used for tier selection, then the system can sustain playback, but it may overlook higher quality tiers that could be sustained during lower bit rate scenes
Solution Approach 1:
The patent applies local quality analysis by examining bit rate requirements at different points in the content (through percentile calculations). Instead of treating the entire program with a single average bit rate, the system identifies local variations in bit rate demands and matches appropriate tiers to these local conditions, enabling higher quality playback during suitable scenes while maintaining sustainability during challenging segments
Data Source
AI summary
A method and apparatus of a device that receives streaming content is described. In an exemplary embodiment, the device receives a media playlist for the content, wherein the media playlist references a plurality of media stream and each of the plurality of media stream are encoded at a different bit rate. The device further determines an initial media stream selected from the plurality of media stream. In addition, the device examines a window of media segments of the initial media stream and determines a local bit rate curve from the window of media segments. The device further selects another one of the plurality of media streams using at least the local bit rate curve.